[Federal Register: January 24, 2007 (Volume 72, Number 15)]
[Proposed Rules]               
[Page 3199-3344]
From the Federal Register Online via GPO Access [wais.access.gpo.gov]
[DOCID:fr24ja07-30]                         
 

[[Page 3199]]

-----------------------------------------------------------------------

Part II





Environmental Protection Agency





-----------------------------------------------------------------------



40 CFR Part 86



Control of Air Pollution From New Motor Vehicles and New Motor Vehicle 
Engines--Heavy-Duty Vehicle and Engine Standards; Onboard Diagnostic 
Requirements; Proposed Rule


[[Page 3200]]


-----------------------------------------------------------------------

ENVIRONMENTAL PROTECTION AGENCY

40 CFR Part 86

[OAR-2005-0047; FRL-8256-9]
RIN 2060-AL92

 
Control of Air Pollution From New Motor Vehicles and New Motor 
Vehicle Engines; Regulations Requiring Onboard Diagnostic Systems on 
2010 and Later Heavy-Duty Engines Used in Highway Applications Over 
14,000 Pounds; Revisions to Onboard Diagnostic Requirements for Diesel 
Highway Heavy-Duty Vehicles Under 14,000 Pounds

AGENCY: Environmental Protection Agency (EPA).

ACTION: Notice of proposed rulemaking.

-----------------------------------------------------------------------

SUMMARY: In 2001, EPA finalized a new, major program for highway heavy-
duty engines. That program, the Clean Diesel Trucks and Buses program, 
will result in the introduction of advanced emissions control systems 
such as catalyzed diesel particulate filters (DPF) and catalysts 
capable of reducing harmful nitrogen oxide (NOX) emissions. 
This proposal would require that these advanced emissions control 
systems be monitored for malfunctions via an onboard diagnostic system 
(OBD), similar to those systems that have been required on passenger 
cars since the mid-1990s. This proposal would require manufacturers to 
install OBD systems that monitor the functioning of emission control 
components and alert the vehicle operator to any detected need for 
emission related repair. This proposal would also require that 
manufacturers make available to the service and repair industry 
information necessary to perform repair and maintenance service on OBD 
systems and other emission related engine components. Lastly, this 
proposal would revise certain existing OBD requirements for diesel 
engines used in heavy-duty vehicles under 14,000 pounds.

DATES: If we do not receive a request for a public hearing, written 
comments are due March 26, 2007. Requests for a public hearing must be 
received by February 8, 2007. If we do receive a request for a public 
hearing, we will publish a notice in the Federal Register and on the 
Web at http://www.epa.gov/obd/regtech/heavy.htm containing details 

regarding the location, date, and time of the public hearing. In that 
case, the public comment period would close 30 days after the public 
hearing. Under the Paperwork Reduction Act, comments on the information 
collection provisions must be received by OMB on or before February 23, 
2007.

ADDRESSES: Submit your comments, identified by Docket ID No. EPA-HQ-
OAR-2005-0047, by one of the following methods:
     http://www.regulations.gov: Follow the on-line 

instructions for submitting comments.
     Mail: Onboard Diagnostic (OBD) Systems on 2010 and Later 
Heavy-Duty Highway Vehicles and Engines, Environmental Protection 
Agency, Mailcode: 6102T, 1200 Pennsylvania Ave., NW., Washington, DC, 
20460, Attention Docket ID No. EPA-HQ-OAR-2005-0047. In addition, 
please mail a copy of your comments on the information collection 
provisions to the Office of Information and Regulatory Affairs, Office 
of Management and Budget (OMB), Attn: Desk Officer for EPA, 725 17th 
St. NW., Washington, DC 20503.
    Instructions: Direct your comments to Docket ID No. EPA-HQ-OAR-
2005-0047. EPA's policy is that all comments received will be included 
in the public docket without change and may be made available online at 
http://www.regulations.gov, including any personal information 

provided, unless the comment includes information claimed to be 
Confidential Business Information (CBI) or other information whose 
disclosure is restricted by statute. Do not submit information that you 
consider to be CBI or otherwise protected through http://www.regulations.gov or e-mail. The http://www.regulations.gov Web site 

is an ``anonymous access'' system, which means EPA will not know your 
identity or contact information unless you provide it in the body of 
your comment. If you send an e-mail comment directly to EPA without 
going through http://www.regulations.gov your e-mail address will be 

automatically captured and included as part of the comment that is 
placed in the public docket and made available on the Internet. If you 
submit an electronic comment, EPA recommends that you include your name 
and other contact information in the body of your comment and with any 
disk or CD-ROM you submit. If EPA cannot read your comment due to 
technical difficulties and cannot contact you for clarification, EPA 
may not be able to consider your comment. Electronic files should avoid 
the use of special characters, any form of encryption, and be free of 
any defects or viruses.
    Docket: All documents in the docket are listed in the http://www.regulations.gov
 index. Although listed in the index, some 

information is not publicly available, e.g., CBI or other information 
whose disclosure is restricted by statute. Certain other material, such 
as copyrighted material, will be publicly available only in hard copy. 
Publicly available docket materials are available either electronically 
in http://www.regulations.gov or in hard copy at the Air Docket, EPA/

DC, EPA West, Room B102, 1301 Constitution Ave., NW., Washington, DC. 
The Public Reading Room is open from 8:30 a.m. to 4:30 p.m., Monday 
through Friday, excluding legal holidays. The telephone number for the 
Public Reading Room is (202) 566-1744, and the telephone number for the 
Air Docket is (202) 566-1742.


    Note: The EPA Docket Center suffered damage due to flooding 
during the last week of June 2006. The Docket Center is continuing 
to operate. However, during the cleanup, there will be temporary 
changes to Docket Center telephone numbers, addresses, and hours of 
operation for people who wish to make hand deliveries or visit the 
Public Reading Room to view documents. Consult EPA's Federal 
Register notice at 71 FR 38147 (July 5, 2006) or the EPA Web site at 
http://www.epa.gov/epahome/dockets.htm for current information on 

docket operations, locations and telephone numbers. The Docket 
Center's mailing address for U.S. mail and the procedure for 
submitting comments to http://www.regulations.gov are not affected by the 

flooding and will remain the same.


FOR FURTHER INFORMATION CONTACT: U.S. EPA, National Vehicle and Fuel 
Emissions Laboratory, Assessment and Standards Division, 2000 
Traverwood Drive, Ann Arbor, MI 48105; telephone (734) 214-4405, fax 
(734) 214-4816, email sherwood.todd@epa.gov.

SUPPLEMENTARY INFORMATION:

Regulated Entities

    This action will affect you if you produce or import new heavy-duty 
engines which are intended for use in highway vehicles such as trucks 
and buses, or produce or import such highway vehicles, or convert 
heavy-duty vehicles or heavy-duty engines used in highway vehicles to 
use alternative fuels.
    The following table gives some examples of entities that may have 
to follow the regulations. But because these are only examples, you 
should carefully examine the regulations in 40 CFR part 86. If you have 
questions, call the person listed in the FOR FURTHER INFORMATION 
CONTACT section of this preamble:

[[Page 3201]]



----------------------------------------------------------------------------------------------------------------
                                                                              Examples of potentially regulated
                  Category                   NAICS Codes\a\   SIC Codes\b\                 entities
----------------------------------------------------------------------------------------------------------------
Industry...................................          336111            3711  Motor Vehicle Manufacturers; Engine
                                                     336112                   and Truck Manufacturers.
                                                     336120
Industry...................................          811112            7533  Commercial Importers of Vehicles
                                                     811198            7549   and Vehicle Components.
                                                     541514            8742
Industry...................................          336111            3592  Alternative fuel vehicle
                                                                              converters.
                                                     336312            3714
                                                     422720            5172
                                                     454312            5984
                                                     811198            7549
                                                     541514            8742
                                                     541690           8931
----------------------------------------------------------------------------------------------------------------
\a\North American Industry Classification Systems (NAICS).
\b\Standard Industrial Classification (SIC) system code.

What Should I Consider as I Prepare My Comments for EPA?

    Submitting CBI. Do not submit this information to EPA through 
http://www.regulations.gov or e-mail. Clearly mark the part or all of the 

information that you claim to be CBI. For CBI information in a disk or 
CD ROM that you mail to EPA, mark the outside of the disk or CD ROM as 
CBI and then identify electronically within the disk or CD ROM the 
specific information that is claimed as CBI). In addition to one 
complete version of the comment that includes information claimed as 
CBI, a copy of the comment that does not contain the information 
claimed as CBI must be submitted for inclusion in the public docket. 
Information so marked will not be disclosed except in accordance with 
procedures set forth in 40 CFR part 2.
    Tips for Preparing Your Comments. When submitting comments, 
remember to:
     Identify the rulemaking by docket number and other 
identifying information (subject heading, Federal Register date and 
page number).
     Follow directions--The agency may ask you to respond to 
specific questions or organize comments by referencing a Code of 
Federal Regulations (CFR) part or section number.
     Explain why you agree or disagree; suggest alternatives 
and substitute language for your requested changes.
     Describe any assumptions and provide any technical 
information and/or data that you used.
     If you estimate potential costs or burdens, explain how 
you arrived at your estimate in sufficient detail to allow for it to be 
reproduced.
     Provide specific examples to illustrate your concerns, and 
suggest alternatives.
     Explain your views as clearly as possible, avoiding the 
use of profanity or personal threats.
     Make sure to submit your comments by the comment period 
deadline identified.

Outline of this Preamble

I. Overview
    A. Background
    B. What Is EPA Proposing?
    1. OBD Requirements for Engines Used in Highway Vehicles Over 
14,000 Pounds GVWR
    2. Requirements That Service Information Be Made Available
    3. OBD Requirements for Diesel Heavy-Duty Vehicles and Engines 
Used in Vehicles Under 14,000 Pounds
    C. Why Is EPA Making This Proposal?
    1. Highway Engines and Vehicles Contribute to Serious Air 
Pollution Problems
    2. Emissions Control of Highway Engines and Vehicles Depends on 
Properly Operating Emissions Control Systems
    3. Basis for Action Under the Clean Air Act
    D. How Has EPA Chosen the Level of the Proposed Emissions 
Thresholds?
    E. World Wide Harmonized OBD (WWH-OBD)
    F. Onboard Diagnostics for Diesel Engines Used in Nonroad Land-
Based Equipment
    1. What Is the Baseline Nonroad OBD System?
    2. What Is The Appropriate Level of OBD Monitoring for Nonroad 
Diesel Engines?
    3. What Should the OBD Standardization Features Be?
    4. What Are the Prospects and/or Desires for International 
Harmonization of Nonroad OBD?
II. What Are the Proposed OBD Requirements and When Would They Be 
Implemented?
    A. General OBD System Requirements
    1. The OBD System
    2. Malfunction Indicator Light (MIL) and Diagnostic Trouble 
Codes (DTC)
    3. Monitoring Conditions
    4. Determining the Proper OBD Malfunction Criteria
    B. Monitoring Requirements and Timelines for Diesel-Fueled/
Compression-Ignition Engines
    1. Fuel System Monitoring
    2. Engine Misfire Monitoring
    3. Exhaust Gas Recirculation (EGR) System Monitoring
    4. Turbo Boost Control System Monitoring
    5. Non-Methane Hydrocarbon (NMHC) Converting Catalyst Monitoring
    6. Selective Catalytic Reduction (SCR) and Lean NOX 
Catalyst Monitoring
    7. NOX Adsorber System Monitoring
    8. Diesel Particulate Filter (DPF) System Monitoring
    9. Exhaust Gas Sensor Monitoring
    C. Monitoring Requirements and Timelines for Gasoline/Spark-
Ignition Engines
    1. Fuel System Monitoring
    2. Engine Misfire Monitoring
    3. Exhaust Gas Recirculation (EGR) Monitoring
    4. Cold Start Emission Reduction Strategy Monitoring
    5. Secondary Air System Monitoring
    6. Catalytic Converter Monitoring
    7. Evaporative Emission Control System Monitoring
    8. Exhaust Gas Sensor Monitoring
    D. Monitoring Requirements and Timelines for Other Diesel and 
Gasoline Systems
    1. Variable Valve Timing and/or Control (VVT) System Monitoring
    2. Engine Cooling System Monitoring
    3. Crankcase Ventilation System Monitoring
    4. Comprehensive Component Monitors
    5. Other Emissions Control System Monitoring
    6. Exceptions to Monitoring Requirements
    E. A Standardized Method To Measure Real World Monitoring 
Performance
    1. Description of Software Counters To Track Real World 
Performance
    2. Proposed Performance Tracking Requirements
    F. Standardization Requirements
    1. Reference Documents
    2. Diagnostic Connector Requirements
    3. Communications to a Scan Tool
    4. Required Emissions Related Functions
    5. In-Use Performance Ratio Tracking Requirements
    6. Exceptions to Standardization Requirements
    G. Implementation Schedule, In-Use Liability, and In-Use 
Enforcement
    1. Implementation Schedule and In-Use Liability Provisions
    2. In-Use Enforcement
    H. Proposed Changes to the Existing 8,500 to 14,000 Pound Diesel 
OBD Requirements

[[Page 3202]]

    1. Selective Catalytic Reduction and Lean NOX 
Catalyst Monitoring
    2. NOX Adsorber System Monitoring
    3. Diesel Particulate Filter System Monitoring
    4. NMHC Converting Catalyst Monitoring
    5. Other Monitors
    6. CARB OBDII Compliance Option and Deficiencies
    I. How Do the Proposed Requirements Compare to California's?
III. Are the Proposed Monitoring Requirements Feasible?
    A. Feasibility of the Monitoring Requirements for Diesel/
Compression-Ignition Engines
    1. Fuel System Monitoring
    2. Engine Misfire Monitoring
    3. Exhaust Gas Recirculation (EGR) Monitoring
    4. Turbo Boost Control System Monitoring
    5. Non-Methane Hydrocarbon (NMHC) Converting Catalyst Monitoring
    6. Selective Catalytic Reduction (SCR) and NOX 
Conversion Catalyst Monitoring
    7. NOX Adsorber Monitoring
    8. Diesel Particulate Filter (DPF) Monitoring
    9. Exhaust Gas Sensor Monitoring
    B. Feasibility of the Monitoring Requirements for Gasoline/
Spark-Ignition Engines
    1. Fuel System Monitoring
    2. Engine Misfire Monitoring
    3. Exhaust Gas Recirculation (EGR) Monitoring
    4. Cold Start Emission Reduction Strategy Monitoring
    5. Secondary Air System Monitoring
    6. Catalytic Converter Monitoring
    7. Evaporative System Monitoring
    8. Exhaust Gas Sensor Monitoring
    C. Feasibility of the Monitoring Requirements for Other Diesel 
and Gasoline Systems
    1. Variable Valve Timing and/or Control (VVT) System Monitoring
    2. Engine Cooling System Monitoring
    3. Crankcase Ventilation System Monitoring
    4. Comprehensive Component Monitoring
IV. What Are the Service Information Availability Requirements?
    A. What Is the Important Background Information for the Proposed 
Service Information Provisions?
    B. How Do the Below 14,000 Pound and Above 14,000 Pounds 
Aftermarket Service Industry Compare?
    C. What Provisions Are Being Proposed for Service Information 
Availability?
    1. What Information Is Proposed To Be Made Available by OEMs?
    2. What Are the Proposed Requirements for Web-Based Delivery of 
the Required Information?
    3. What Provisions Are Being Proposed for Service Information 
for Third Party Information Providers?
    4. What Requirements Are Being Proposed for the Availability of 
Training Information?
    5. What Requirements Are Being Proposed for Reprogramming of 
Vehicles?
    6. What Requirements Are Being Proposed for the Availability of 
Enhanced Information for Scan Tools for Equipment and Tool 
Companies?
    7. What Requirements Are Being Proposed for the Availability of 
OEM--Specific Diagnostic Scan Tools and Other Special Tools?
    8. Which Reference Materials Are Being Proposed for 
Incorporation by Reference?
V. What Are the Emissions Reductions Associated With the Proposed 
OBD Requirements?
VI. What Are the Costs Associated With the Proposed OBD 
Requirements?
    A. Variable Costs for Engines Used in Vehicles Over 14,000 
Pounds
    B. Fixed Costs for Engines Used in Vehicles Over 14,000 Pounds
    C. Total Costs for Engines Used in Vehicles Over 14,000 Pounds
    D. Costs for Diesel Heavy-Duty Vehicles and Engines Used in 
Heavy-Duty Vehicles Under 14,000 Pounds
VII. What are the Updated Annual Costs and Costs per Ton Associated 
With the 2007/2010 Heavy-Duty Highway Program?
    A. Updated 2007 Heavy-Duty Highway Rule Costs Including OBD
    B. Updated 2007 Heavy-Duty Highway Rule Costs Per Ton Including 
OBD
VIII. What Are the Requirements for Engine Manufacturers?
    A. Documentation Requirements
    B. Catalyst Aging Procedures
    C. Demonstration Testing
    1. Selection of Test Engines
    2. Required Testing
    3. Testing Protocol
    4. Evaluation Protocol
    5. Confirmatory Testing
    D. Deficiencies
    E. Production Evaluation Testing
    1. Verification of Standardization Requirements
    2. Verification of Monitoring Requirements
    3. Verification of In-Use Monitoring Performance Ratios
IX. What Are the Issues Concerning Inspection and Maintenance 
Programs?
    A. Current Heavy-Duty I/M Programs
    B. Challenges for Heavy-Duty I/M
    C. Heavy-Duty OBD and I/M
X. Statutory and Executive Order Reviews
    A. Executive Order 12866: Regulatory Planning and Review
    B. Paperwork Reduction Act
    C. Regulatory Flexibility Act (RFA), as Amended by the Small 
Business Regulatory Enforcement Fairness Act of 1996 (SBREFA), 5 
U.S.C. 601 et. seq.
    D. Unfunded Mandates Reform Act
    E. Executive Order 13132: Federalism
    F. Executive Order 13175: Consultation and Coordination With 
Indian Tribal Governments
    G. Executive Order 13045: Protection of Children From 
Environmental Health and Safety Risks
    H. Executive Order 13211: Actions That Significantly Affect 
Energy Supply, Distribution, or Use
    I. National Technology Transfer Advancement Act
XI. Statutory Provisions and Legal Authority

I. Overview

A. Background

    Section 202(m) of the CAA, 42 U.S.C. 7521(m), directs EPA to 
promulgate regulations requiring 1994 and later model year light-duty 
vehicles (LDVs) and light-duty trucks (LDTs) to contain an OBD system 
that monitors emission-related components for malfunctions or 
deterioration ``which could cause or result in failure of the vehicles 
to comply with emission standards established'' for such vehicles. 
Section 202(m) also states that, ``The Administrator may, in the 
Administrator's discretion, promulgate regulations requiring 
manufacturers to install such onboard diagnostic systems on heavy-duty 
vehicles and engines.''
    On February 19, 1993, we published a final rule requiring 
manufacturers of light-duty applications to install such OBD systems on 
their vehicles beginning with the 1994 model year (58 FR 9468). The OBD 
systems must monitor emission control components for any malfunction or 
deterioration that could cause exceedance of certain emission 
thresholds. The regulation also required that the driver be notified of 
any need for repair via a dashboard light, or malfunction indicator 
light (MIL), when the diagnostic system detected a problem. We also 
allowed optional compliance with California's second phase OBD 
requirements, referred to as OBDII (13 CCR 1968.1), for purposes of 
satisfying the EPA OBD requirements. Since publishing the 1993 OBD 
final rule, EPA has made several revisions to the OBD requirements, 
most of which served to align the EPA OBD requirements with revisions 
to the California OBDII requirements (13 CCR 1968.2).
    On August 9, 1995, EPA published a final rulemaking that set forth 
service information regulations for light-duty vehicles and light-duty 
trucks (60 FR 40474). These regulations, in part, required each 
Original Equipment Manufacturer (OEM) to do the following: (1) List all 
of its emission-related service and repair information on a Web site 
called FedWorld (including the cost of each item and where it could be 
purchased); (2) either provide enhanced information to equipment and 
tool companies or make its OEM-specific diagnostic tool available for 
purchase by aftermarket technicians, and (3) make reprogramming 
capability available to independent service and repair professionals if 
its franchised dealerships had such capability. These requirements are 
intended to ensure that aftermarket service and repair facilities

[[Page 3203]]

have access to the same emission-related service information, in the 
same or similar manner, as that provided by OEMs to their franchised 
dealerships. These service information availability requirements have 
been revised since that first final rule in response to changing 
technology among other reasons. (68 FR 38428)
    In October of 2000, we published a final rule requiring OBD systems 
on heavy-duty vehicles and engines up to 14,000 pounds GVWR (65 FR 
59896). In that rule, we expressed our intention of developing OBD 
requirements in a future rule for vehicles and engines used in vehicles 
over 14,000 pounds. We expressed this same intention in our 2007HD 
highway final rule (66 FR 5002) which established new heavy-duty 
highway emissions standards for 2007 and later model year engines. In 
June of 2003, we published a final rule extending service information 
availability requirements to heavy-duty vehicles and engines weighing 
up to 14,000 pounds GVWR. We declined extending these requirements to 
engines above 14,000 pounds GVWR at least until such engines are 
subject to OBD requirements.
    On January 18, 2001, EPA established a comprehensive national 
control program--the Clean Diesel Truck and Bus program--that regulates 
the heavy-duty vehicle and its fuel as a single system. (66 FR 5002) As 
part of this program, new emission standards will begin to take effect 
in model year 2007 and will apply to heavy-duty highway engines and 
vehicles. These standards are based on the use of high-efficiency 
catalytic exhaust emission control devices or comparably effective 
advanced technologies. Because these devices are damaged by sulfur, the 
regulation also requires the level of sulfur in highway diesel fuel be 
reduced by 97 percent.\1\
---------------------------------------------------------------------------

    \1\ Note that the 2007HD highway rule contained new emissions 
standards for gasoline engines as well as diesel engines.
---------------------------------------------------------------------------

    Today's action proposes new OBD requirements for highway engines 
used in vehicles greater than 14,000 pounds. Today's action also 
proposes new availability requirements for emission-related service 
information that will make this information more widely available to 
the industry servicing vehicles over 14,000 pounds.
    In addition to these proposed requirements and changes, we are 
seeking comment on possible future regulations that would require OBD 
systems on heavy-duty diesel engines used in nonroad equipment. Diesel 
engines used in nonroad equipment are, like highway engines, a major 
source of NOX and particulate matter (PM) emissions, and the 
diesel engines used in nonroad equipment are essentially the same as 
those used in heavy-duty highway trucks. Further, new regulations 
applicable to nonroad diesel engines will result in the introduction of 
advanced emissions control systems like those expected for highway 
diesel engines. (69 FR 38958) Therefore, having OBD systems and OBD 
regulations for nonroad engines seems to be a natural progression from 
the proposed requirements for heavy-duty highway engines. We discuss 
this issue in greater detail in section I of this preamble with the 
goal of soliciting public comment regarding how we should proceed with 
respect to nonroad OBD.

B. What Is EPA Proposing?

1. OBD Requirements for Engines Used in Highway Vehicles Over 14,000 
Pounds GVWR
    We believe that OBD requirements should be extended to include over 
14,000 pound heavy-duty vehicles and engines for many reasons. In the 
past, heavy-duty diesel engines have relied primarily on in-cylinder 
modifications to meet emission standards. For example, emission 
standards have been met through changes in fuel timing, piston design, 
combustion chamber design, charge air cooling, use of four valves per 
cylinder rather than two valves, and piston ring pack design and 
location improvements. In contrast, the 2004 and 2007 emission 
standards represent a different sort of technological challenge that 
are being met with the addition of exhaust gas recirculation (EGR) 
systems and the addition of exhaust aftertreatment devices such as 
diesel particulate filters (DPF), sometimes called PM traps, and 
NOX catalysts. Such ``add on'' devices can experience 
deterioration and malfunction that, unlike the engine design elements 
listed earlier, may go unnoticed by the driver. Because deterioration 
and malfunction of these devices can go unnoticed by the driver, and 
because their primary purpose is emissions control, and because the 
level of emission control is on the order of 50 to 99 percent, some 
form of diagnosis and malfunction detection is crucial. We believe that 
such detection can be effectively achieved by employing a well designed 
OBD system.
    The same is true for gasoline heavy-duty vehicles and engines. 
While emission control is managed with both engine design elements and 
aftertreatment devices, the catalytic converter is the primary emission 
control feature accounting for over 95 percent of the emission control. 
We believe that monitoring the emission control system for proper 
operation is critical to ensure that new vehicles and engines certified 
to the very low emission standards set in recent years continue to meet 
those standards throughout their full useful life.
    Further, the industry trend is clearly toward increasing use of 
computer and electronic controls for both engine and powertrain 
management, and for emission control. In fact, the heavy-duty industry 
has already gone a long way, absent any government regulation, to 
standardize computer communication protocols.\2\ Computer and 
electronic control systems, as opposed to mechanical systems, provide 
improvements in many areas including, but not limited to, improved 
precision and control, reduced weight, and lower cost. However, 
electronic and computer controls also create increased difficulty in 
diagnosing and repairing the malfunctions that inevitably occur in any 
engine or powertrain system. Today's proposed OBD requirements would 
build on the efforts already undertaken by the industry to ensure that 
key emissions related components will be monitored in future heavy-duty 
vehicles and engines and that the diagnosis and repair of those 
components will be as efficient and cost effective as possible.
---------------------------------------------------------------------------

    \2\ See ``On-Board Diagnostics, A Heavy-Duty Perspective,'' SAE 
951947; ``Recommended Practice for a Serial Control and 
Communications Vehicle Network,'' SAE J1939 which may be obtained 
from Society of Automotive Engineers International, 400 Commonwealth 
Dr., Warrendale, PA, 15096-0001; and ``Road Vehicles-Diagnostics on 
Controller Area Network (CAN)--Part 4: Requirements for emission-
related systems,'' ISO 15765-4:2001 which may be obtained from the 
International Organization for Standardization, Case Postale 56, CH-
1211 Geneva 20, Switzerland.
---------------------------------------------------------------------------

    Lastly, heavy-duty engines and, in particular, diesel engines tend 
to have very long useful lives. With age comes deterioration and a 
tendency toward increasing emissions. With the OBD systems proposed 
today, we expect that these engines will continue to be properly 
maintained and therefore will continue to emit at low emissions levels 
even after accumulating hundreds of thousands and even a million miles.
    For the reasons laid out above, most manufacturers of vehicles, 
trucks, and engines have incorporated some type of OBD system into 
their products that are capable of identifying when certain types of 
malfunctions occur, and in what systems. In the heavy-duty industry, 
those OBD systems traditionally have been geared toward

[[Page 3204]]

detecting malfunctions causing drivability and/or fuel economy related 
problems. Without specific requirements for manufacturers to include 
OBD mechanisms to detect emission-related problems, those types of 
malfunctions that could result in high emissions without a 
corresponding adverse drivability or fuel economy impact could go 
unnoticed by both the driver and the repair technician. The resulting 
increase in emissions and detrimental impact on air quality could be 
avoided by incorporating an OBD system capable of detecting emission 
control system malfunctions.
2. Requirements That Service Information Be Made Available
    We are proposing that makers of engines that go into vehicles over 
14,000 pounds make available to any person engaged in repair or service 
all information necessary to make use of the OBD systems and for making 
emission-related repairs, including any emissions-related information 
that is provided by the OEM to franchised dealers. This information 
includes, but is not limited to, manuals, technical service bulletins 
(TSBs), a general description of the operation of each OBD monitor, 
etc. We discuss the proposed requirements further in section IV of this 
preamble.
    The proposed requirements are similar to those required currently 
for all 1996 and newer light-duty vehicles and light-duty trucks and 
2005 and newer heavy-duty applications up to 14,000 pounds. While EPA 
understands that there may be some differences between aftermarket 
service for the under 14,000 pound and over 14,000 pound applications, 
we believe that any such differences would not substantially affect the 
implementation of such requirements and that, therefore, it is 
reasonable to use EPA's existing service information regulations as a 
basis for proposing service information requirements for the over 
14,000 pound arena. See section IV for a complete discussion of the 
service information provisions being proposed for the availability of 
over 14,000 pound service information.
    Note that information for making emission-related repairs does not 
include information used to design and manufacture parts, but it may 
include OEM changes to internal calibrations and other indirect 
information, as discussed in section IV.
3. OBD Requirements for Diesel Heavy-Duty Vehicles and Engines Used in 
Vehicles Under 14,000 Pounds
    We are also proposing some changes to the existing diesel OBD 
requirements for heavy-duty applications under 14,000 pounds (i.e., 
8,500 to 14,000 pounds). Some of these changes are being proposed for 
the 2007 and later model years (i.e., for immediate implementation) 
because we believe that some of the requirements that we currently have 
in place for 8,500 to 14,000 pound applications cannot be met by 
diesels without granting widespread deficiencies to industry. Other 
changes are being proposed for the 2010 and later model years since 
they represent an increase in the stringency of our current OBD 
requirements and, therefore, some leadtime is necessary for 
manufacturers to comply. All of the changes being proposed for 8,500 to 
14,000 pound diesel applications would result in OBD emissions 
thresholds identical, for all practical purposes, to the OBD thresholds 
being proposed for over 14,000 pound applications.

C. Why Is EPA Making This Proposal?

1. Highway Engines and Vehicles Contribute to Serious Air Pollution 
Problems
    The pollution emitted by heavy-duty highway engines contributes 
greatly to our nation's continuing air quality problems. Our 2007HD 
highway rule was designed to address these serious air quality 
problems. These problems include premature mortality, aggravation of 
respiratory and cardiovascular disease, aggravation of existing asthma, 
acute respiratory symptoms, chronic bronchitis, and decreased lung 
function. Numerous studies also link diesel exhaust to increased 
incidence of lung cancer. We believe that diesel exhaust is likely to 
be carcinogenic to humans by inhalation and that this cancer hazard 
exists for occupational and environmental levels of exposure.
    Our 2007HD highway rule will regulate the heavy-duty vehicle and 
its fuel as a single system. As part of this program, new emission 
standards will begin to take effect in model year 2007 and phase-in 
through model year 2010, and will apply to heavy-duty highway engines 
and vehicles. These standards are based on the use of high-efficiency 
catalytic exhaust emission control devices or comparably effective 
advanced technologies and a cap on the allowable sulfur content in both 
diesel fuel and gasoline.
    In the 2007HD highway final rule, we estimated that, by 2007, 
heavy-duty trucks and buses would account for about 28 percent of 
nitrogen oxides emissions and 20 percent of particulate matter 
emissions from mobile sources. In some urban areas, the contribution is 
even greater. The 2007HD highway program will reduce particulate matter 
and oxides of nitrogen emissions from heavy-duty engines by 90 percent 
and 95 percent below current standard levels, respectively. In order to 
meet these more stringent standards for diesel engines, the program 
calls for a 97 percent reduction in the sulfur content of diesel fuel. 
As a result, diesel vehicles will achieve gasoline-like exhaust 
emission levels. We have also established more stringent standards for 
heavy-duty gasoline vehicles, based in part on the use of the low 
sulfur gasoline that will be available when the standards go into 
effect.
2. Emissions Control of Highway Engines and Vehicles Depends on 
Properly Operating Emissions Control Systems
    The emissions reductions and resulting health and welfare benefits 
of the 2007HD highway program will be dramatic when fully implemented. 
By 2030, the program will reduce annual emissions of nitrogen oxides, 
nonmethane hydrocarbons, and particulate matter by a projected 2.6 
million, 115,000 and 109,000 tons, respectively. However, to realize 
those large emission reductions and health benefits, the emission 
control systems on heavy-duty highway engines and vehicles must 
continue to provide the 90 to 95 percent emission control effectiveness 
throughout their operating life. Today's proposed OBD requirements will 
help to ensure that emission control systems continue to operate 
properly by detecting when those systems malfunction, by then notifying 
the driver that a problem exists that requires service and, lastly, by 
informing the service technician what the problem is so that it can be 
properly repaired.
3. Basis for Action Under the Clean Air Act
    Section 202(m) of the CAA, 42 U.S.C. 7521(m), directs EPA to 
promulgate regulations requiring 1994 and later model year light-duty 
vehicles (LDVs) and light-duty trucks (LDTs) to contain an OBD system 
that monitors emission-related components for malfunctions or 
deterioration ``which could cause or result in failure of the vehicles 
to comply with emission standards established'' for such vehicles. 
Section

[[Page 3205]]

202(m) also states that, ``The Administrator may, in the 
Administrator's discretion, promulgate regulations requiring 
manufacturers to install such onboard diagnostic systems on heavy-duty 
vehicles and engines.''
    Section 202(m)(5) of the CAA states that the Administrator shall 
require manufacturers to, ``provide promptly to any person engaged in 
the repairing or servicing of motor vehicles or motor vehicle engines * 
* * with any and all information needed to make use of the emission 
control diagnostics system prescribed under this subsection and such 
other information including instructions for making emission related 
diagnosis and repairs.''

D. How Has EPA Chosen the Level of the Proposed Emissions Thresholds?

    The OBD emissions thresholds that we are proposing are summarized 
in Tables II.B-1, II.C-1, II.H-1 and II.H-2. These tables show the 
actual threshold levels and how they relate to current emissions 
standards. Here, we wish to summarize how we chose those proposed 
thresholds. First, it is important to note that OBD is more than 
emissions thresholds. In fact, most OBD monitors are not actually tied 
to an emissions threshold. Instead, they monitor the performance of a 
given component or system and evaluate that performance based on 
electrical information (e.g., voltage within proper range) or 
temperature information (e.g., temperature within range), etc. Such 
monitors often detect malfunctions well before emissions are seriously 
compromised. Nonetheless, emissions thresholds are a critical element 
to OBD requirements since some components and systems, most notably any 
aftertreatment devices, cannot be monitored in simple electrical or 
temperature related terms. Instead, their operating characteristics can 
be measured and correlated to an emissions impact. This way, when those 
operating characteristics are detected, an unacceptable emissions 
increase can be inferred and a malfunction can be noted to the driver.
    Part of the challenge in establishing OBD requirements is 
determining the point--the OBD threshold--at which an unacceptable 
emissions increase has occurred that is detectable by the best 
available OBD technology. Two factors have gone into our determination 
of the emissions thresholds we are proposing: technological 
feasibility; and the costs and emissions reductions associated with 
repairs initiated as a result of malfunctions found by OBD systems. The 
first of these factors is discussed in more detail in section III where 
we present our case for the technological feasibility of the 
thresholds. In summary, we believe that the thresholds we are proposing 
are, while challenging, technologically feasible in the 2010 and later 
timeframe. We have carefully considered monitoring system capability, 
sensor capability, emissions measurement capability, test-to-test 
variability and, perhaps most importantly, the manufacturers' 
engineering and test cell resources and have arrived at thresholds we 
believe can be met on one engine family per manufacturer in the 2010 
model year and on all engine families by the 2013 model year.
    We believe that the proposed thresholds strike the proper balance 
between environmental protection, OBD and various sensor capabilities, 
and avoidance of repairs whose costs could be high compared to their 
emission control results. One must keep in mind that increasingly 
stringent OBD thresholds (i.e., OBD detection at lower emissions 
levels) may lead to more durable emission controls due to a 
manufacturer's desire to avoid the negative impression given their 
product upon an OBD detection. Such an outcome would result in lower 
fleetwide emissions while increasing costs to manufacturers. However, 
increasingly stringent OBD thresholds may also lead to more OBD 
detections and more OBD induced repairs and, perhaps, many OBD induced 
repairs for malfunctions having little impact on emissions. Such an 
outcome would result in lower fleetwide emissions while increasing 
costs to both manufacturers and truck owners.

E. World Wide Harmonized OBD (WWH-OBD)

    Within the United Nations (UN), the World Forum for Harmonization 
of Vehicle Regulations (WP.29) administers the 1958 Geneva Agreement 
(1958 Agreement) to facilitate the adoption of uniform conditions of 
approval and reciprocal recognition of approval for motor vehicle 
equipment and parts. As a result, WP.29 has responsibility for vehicle 
regulations within Europe and, indirectly, many countries outside of 
Europe that have voluntarily adopted the WP.29 regulations. The United 
States was never a party to the 1958 Agreement, but EPA has monitored 
the WP.29 regulations developed under the 1958 Agreement and we have 
benefited from a reciprocal consultative relationship with our European 
counterparts. More recently, WP.29 took on the responsibility of 
administering the 1998 Global Agreement that established a process to 
permit all regions of the world to jointly develop global technical 
regulations without required mutual recognition of approvals or 
designated compliance and enforcement. The United States is a signatory 
of the 1998 Global Agreement (1998 Agreement), and EPA has 
responsibility for representing the U.S. with respect to environmental 
issues within WP.29 as they pertain to the 1998 Agreement.
    During the one-hundred-and-twenty-sixth session of WP.29 of March 
2002, the Executive Committee (AC.3) of the 1998 Global Agreement (1998 
Agreement) adopted a Programme of Work, which includes the development 
of a Global Technical Regulation (GTR) concerning onboard diagnostic 
systems for heavy-duty vehicles and engines. An informal working 
group--the WWH-OBD working group--was established to develop the GTR. 
The working group was instructed that the OBD system should detect 
failures from the engine itself, as well as from the exhaust 
aftertreatment systems fitted downstream of the engine, and from the 
package of information exchanged between the engine electronic control 
unit(s) and the rest of vehicle and/or powertrain. The working group 
was also instructed to base the OBD requirements on the technologies 
expected to be industrially available at the time the GTR would be 
enforced, and to take into account both the expected state of 
electronics in the years 2005-2008 and the expected newest engine and 
aftertreatment technologies.
    In November 2003, AC.3 further directed the working group to 
structure the GTR in such a manner as to enable its future extension to 
other functions of the vehicle. In so doing, AC.3 did not revise the 
scope of the task given to the working group (i.e., the scope remained 
emissions-related heavy-duty OBD). As a result, the GTR is structured 
such that OBD monitoring and communications could be extended to other 
systems such as vehicle safety systems. This has been achieved by 
dividing the GTR into a set of generic OBD requirements to be followed 
by specific OBD requirements concerning any future desired OBD systems. 
The generic OBD requirements contain definitions and other OBD 
regulatory elements that are meant to be applicable throughout the GTR 
and all of its modules, annexes, and appendices. This generic section 
is followed by the first specific OBD section--emission-related OBD--
which contains definitions and OBD regulatory elements specific to 
emissions-related OBD.
    EPA has been active in the WWH-OBD working group for more than 
three

[[Page 3206]]

years. Because that group has been developing a regulation at the same 
time that we have been developing the requirements in this proposal, 
our proposed OBD requirements are consistent, for the most part, with 
the current efforts of the WWH-OBD group.
    The WWH-OBD working group submitted a draft GTR as a formal 
document in March of 2006. During the months immediately following, the 
WWH-OBD working group has made final revisions to the GTR and will 
submit it to WP.29 for consideration. If approved by WP.29 and adopted 
as a formal global technical regulation, we would intend to propose any 
revisions to our OBD regulations that might be necessary to make them 
consistent with WWH-OBD.\3\
---------------------------------------------------------------------------

    \3\ Note that, while the WWH-OBD GTR is consistent with many of 
the specific requirements we are proposing, it is not currently as 
comprehensive as our proposal (e.g., it does not contain the same 
level of detail with respect to certification requirements and 
enforcement provisions). For that reason, at this time, we do not 
believe that the GTR would fully replace what we are proposing 
today.
---------------------------------------------------------------------------

    The latest version of the draft WWH-OBD GTR has been placed in the 
docket for this rule.\4\ While it is not yet a final document, we are 
nonetheless interested in comments regarding the current version. More 
specifically, we are interested in comments regarding any possible 
inconsistencies between the requirements of the draft GTR and the 
requirements being proposed today. We believe that if such 
inconsistencies exist, they are minor. WWH-OBD provides a framework for 
nations to establish a heavy-duty OBD program. It has the potential to 
result in similar OBD systems, but the WWH-OBD GTR must fit into the 
context of any country's existing heavy-duty emissions regulations. For 
example, at this time, the draft GTR does not specify emissions 
threshold levels, implementation dates, or phase-in schedules. As such, 
our proposal today is much more detailed than the draft WWH-OBD GTR, 
but we believe there exist no major inconsistencies between the two 
regulations.
---------------------------------------------------------------------------

    \4\ ``Revised Proposal for New Draft Global Technical Regulation 
(gtr): Technical Requirements for On-Board Diagnostic Systems (OBD) 
for Road Vehicles;'' ECE/TRANS/WP.29/GRPE/2006/8/Rev.1; March 27, 
2006, Docket ID EPA-HQ-OAR-2005-0047-0004.
---------------------------------------------------------------------------

F. Onboard Diagnostics for Diesel Engines Used in Nonroad Land-Based 
Equipment

    We are also considering regulations--although we are not making any 
proposals today--that would require OBD systems on heavy-duty diesel 
engines used in nonroad land-based equipment. The pollution emitted by 
diesel nonroad engines contributes greatly to our nation's continuing 
air quality problems. Our recent Nonroad Tier 4 rulemaking was designed 
to address these serious air quality problems from land-based diesel 
engines. (69 FR 38958) Like with diesel highway emissions, these 
problems include premature mortality, aggravation of respiratory and 
cardiovascular disease, aggravation of existing asthma, acute 
respiratory symptoms, chronic bronchitis, and decreased lung function. 
And, as noted above, we believe that diesel exhaust is likely to be 
carcinogenic to humans by inhalation and that this cancer hazard exists 
for occupational and environmental levels of exposure.
    In our preamble to the Nonroad Tier 4 final rule, we estimated 
that, absent the nonroad Tier 4 standards, by 2020, land based nonroad 
diesel engines would account for as much as 70 percent of the diesel 
mobile source PM inventory. As part of our nonroad Tier 4 program, new 
emission standards will begin to take effect in calendar year 2011 that 
are based on the use of high-efficiency catalytic exhaust emission 
control devices or comparably effective advanced technologies. As with 
our 2007HD highway program, a cap is also included on the allowable 
sulfur content in nonroad diesel fuel.
    The diesel engines used in nonroad land-based equipment are, in 
certain horsepower ranges, often essentially the same as those used in 
heavy-duty highway trucks. In other horsepower ranges--e.g., very large 
nonroad machines with engines having more than 1,500 horsepower--the 
engine is quite different. Such differences can include the addition of 
cylinders and turbo chargers among other things. Notably, the new 
nonroad Tier 4 regulations will result in the introduction of advanced 
emissions control systems on nonroad land-based equipment; those 
advanced emissions control systems will be the same type of systems as 
those expected for highway diesel engines.
    Therefore, having OBD systems and OBD regulations for nonroad 
diesel engines seems to be a natural progression from the proposed 
requirements for heavy-duty highway engines. Nonetheless, we believe 
that there are differences between nonroad equipment and highway 
applications, and differences between the nonroad market and the 
highway market such that proposing the same OBD requirements for 
nonroad as for highway may not be appropriate. Therefore, we are 
providing advance notice to the public with the goal of soliciting 
public comment regarding how we should proceed with respect to nonroad 
OBD. This section presents issues we have identified and solicits 
comment. We also welcome comment with respect to other issues we have 
not addressed here, such as service information availability.
1. What Is the Baseline Nonroad OBD System?
    We know that highway diesel engines already use a sophisticated 
level of OBD system. For nonroad diesel engines in the 200 to 600 
horsepower range--i.e., the typical range of highway engines--are the 
current OBD system identical to their highway counterparts? How would 
the proposed highway OBD requirements change this, if at all? Do diesel 
engines outside the range typical of highway engines use OBD?
2. What Is the Appropriate Level of OBD Monitoring for Nonroad Diesel 
Engines?
    The proposed OBD requirements for highway engines are very 
comprehensive and would result in virtually every element of the 
emissions control system being monitored. Is this appropriate for 
nonroad diesel engines? And to what degree should such monitoring be 
required? The emissions thresholds proposed for highway engines will 
push OBD and sensor technology beyond where it is today because of 
their stringency. Is a similar level of stringency appropriate for 
nonroad engines? Should emissions thresholds analogous to those 
presented in Table II.B-1 of this preamble even be a part of any 
potential nonroad OBD requirements or should nonroad OBD rely more 
heavily on comprehensive component monitoring as discussed in section 
II.D.4 of this preamble? This latter question is particularly 
compelling given the incredibly broad range of operating 
characteristics for nonroad equipment. Similar to the issue of 
emissions thresholds, certain aspects of the proposed highway OBD 
requirements carry with them serious concerns given the range of use 
for heavy-duty highway trucks (line-haul trucks versus garbage trucks 
versus urban delivery trucks, etc.). As discussed in various places in 
section II of this preamble, this broad range of uses makes it 
difficult for manufacturers to design a single approach that would, for 
example, ensure frequent monitoring events on all possible 
applications. This difficulty could be even more pronounced in the 
nonroad industry given the greater number of possible applications.

[[Page 3207]]

    We request comment regarding what any potential nonroad OBD 
monitoring requirements should look like. More specifically, we request 
comment regarding the inclusion of emissions thresholds versus relying 
solely on comprehensive component monitoring. From commenters in favor 
of emissions thresholds, we request details regarding the appropriate 
level of emissions thresholds including data and strong engineering 
analyses for/against the suggested level. We request comment regarding 
the comprehensiveness of monitoring (i.e., the entire emissions control 
system, aftertreatement devices only, feedback control systems only, 
etc.).
3. What Should the OBD Standardization Features Be?
    Should nonroad OBD include a requirement for a dedicated, OBD-only 
malfunction indicator light? Should nonroad OBD require specific 
communication protocols for communication of onboard information to 
offboard devices and scan tools? What should those protocols be? What 
are the needs of the nonroad service industry with respect to 
standardization of onboard to offboard communications?
4. What Are the Prospects and/or Desires for International 
Harmonization of Nonroad OBD?
    Nonroad equipment is perhaps the most international of all mobile 
source equipment. Land based nonroad equipment, while not as much so as 
marine equipment, tends to be designed, produced, marketed, and sold to 
a world market to a greater extent than is highway equipment. Given 
that, is there a sense within the nonroad industry that international 
harmonization is important? Imperative? Is the proper avenue for 
putting into place nonroad OBD regulations the WWH-OBD process 
discussed above? If so, is industry prepared to play a role in 
developing a nonroad OBD element to the WWH-OBD document? Are other 
government representatives prepared to do so?

II. What Are the Proposed OBD Requirements and When Would They Be 
Implemented?

    The following subsections describe our proposed OBD monitoring 
requirements and the timelines for their implementation. The 
requirements are indicative of our goal for the program which is a set 
of OBD monitors that provide robust diagnosis of the emission control 
system. Our intention is to provide industry sufficient time and 
experience with satisfying the demands of the proposed OBD program. 
While their engines already incorporate OBD systems, those systems are 
generally less comprehensive and do not monitor the emission control 
system in the ways we are proposing. Additionally, the proposed OBD 
requirements represent a new set of technological requirements and a 
new set of certification requirements for the industry in addition to 
the 2007HD highway program and its challenging emission standards for 
PM and NOX and other pollutants. As a result, we believe the 
monitoring requirements and timelines outlined in this section 
appropriately weigh the need for OBD monitors on the emission control 
system and the need to gain experience with not only those monitors but 
also the newly or recently added emission control hardware.
    We request comment on all aspects of the requirements laid out in 
this section and throughout this preamble. As discussed in Section IX, 
we are also interested in comments concerning state run HDOBD-based 
inspection and maintenance (I/M) programs, the level of interest in 
such programs, and comments concerning the suitability of today's 
proposed OBD requirements toward facilitating potential HDOBD I/M 
programs in the future.

A. General OBD System Requirements

1. The OBD System
    We are proposing that the OBD system be designed to operate for the 
actual life of the engine in which it is installed. Further, the OBD 
system cannot be programmed or otherwise designed to deactivate based 
on age and/or mileage of the vehicle during the actual life of the 
engine. This requirement is not intended to alter existing law and 
enforcement practice regarding a manufacturer's liability for an engine 
beyond its regulatory useful life, except where an engine has been 
programmed or otherwise designed so that an OBD system deactivates 
based on age and/or mileage of the engine.
    We are also proposing that computer coded engine operating 
parameters not be changeable without the use of specialized tools and 
procedures (e.g. soldered or potted computer components or sealed (or 
soldered) computer enclosures). Upon Administrator approval, certain 
product lines may be exempted from this requirement if those product 
lines can be shown to not need such protections. In making the approval 
decision, the Administrator will consider such things as the current 
availability of performance chips, performance capability of the 
engine, and sales volume.
2. Malfunction Indicator Light (MIL) and Diagnostic Trouble Codes (DTC)
    Upon detecting a malfunction within the emission control system,\5\ 
the OBD system must make some indication to the driver so that the 
driver can take action to get the problem repaired. The proposal would 
require that a dashboard malfunction indicator light (MIL) be 
illuminated to inform the driver that a problem exists that needs 
attention. Upon illumination of the MIL, the proposal would require 
that a diagnostic trouble code (DTC) be stored in the engine's computer 
that identifies the detected malfunction. This DTC would then be read 
by a service technician to assist in making the necessary repair.
---------------------------------------------------------------------------

    \5\ What constitutes a ``malfunction'' for over 14,000 pound 
applications under today's proposal is covered in section II.B for 
diesel engines, section II.C for gasoline engines, and section II.D 
for all engines.
---------------------------------------------------------------------------

    Because the MIL is meant to inform the driver of a detected 
malfunction, we are proposing that the MIL be located on the driver's 
side instrument panel and be of sufficient illumination and location to 
be readily visible under all lighting conditions. We are proposing that 
the MIL be amber (yellow) in color when illuminated because yellow is 
synonymous with the notion of a ``cautionary warning''; the use of red 
for the MIL would be strictly prohibited because red signifies 
``danger'' which is not the proper message for malfunctions detected 
according to today's proposal. Further, we are proposing that, when 
illuminated, the MIL display the International Standards Organization 
(ISO) engine symbol because this symbol has become accepted after 10 
years of light-duty OBD as a communicator of engine and emissions 
system related problems. We are also proposing that there be only one 
MIL used to indicate all malfunctions detected by the OBD system on a 
single vehicle. We believe this is important to avoid confusion over 
multiple lights and, potentially, multiple interpretations of those 
lights. Nonetheless, we seek comment on this limitation to one 
dedicated MIL to communicate emissions-related malfunctions. We also 
seek comment on the requirement that the MIL be amber in color since 
some trucks may use liquid crystal display (LCD) panels to display 
dashboard information and some such panels are monochromic and unable 
to display color.
    We are also interested in comments regarding the malfunction 
indicator light and the symbol displayed to

[[Page 3208]]

communicate that there is an engine and/or emission-related 
malfunction. As noted, we are proposing use of the ISO engine symbol as 
shown in Table II.A-1. The U.S. Department of Transportation has 
proposed use of an alternative ISO symbol to denote, specifically, an 
emission-related malfunction. (68 FR 55217) That symbol is also shown 
in Table II.A-1. While we are not proposing that this alternative 
symbol be used, comments are solicited regarding whether this 
alternative symbol provides a clearer message to the driver.
    Generally, a manufacturer would be allowed sufficient time to be 
certain that a malfunction truly exists before illuminating the MIL. No 
one benefits if the MIL illuminates spuriously when a real malfunction 
does not exist. Thus, for most OBD monitoring strategies, manufacturers 
would not be required to illuminate the MIL until a malfunction clearly 
exists which will be considered to be the case when the same problem 
has occurred on two sequential driving cycles.\6\
---------------------------------------------------------------------------

    \6\ Generally, a ``driving cycle'' or ``drive cycle'' consists 
of engine startup and engine shutoff or consists of four hours of 
continuous engine operation.
[GRAPHIC] [TIFF OMITTED] TP24JA07.000

    To keep this clear in the onboard computer, we are proposing that 
the OBD system make certain distinctions between the problems it has 
detected, and that the system maintain a strict logic for diagnostic 
trouble code (DTC) storage/erasure and for MIL illumination/
extinguishment. Whenever the enable criteria for a given monitor are 
met, we would expect that monitor to run. For continuous monitors, this 
would be during essentially all engine operation.\7\ For non-continuous 
monitors, it would be during only a subset of engine operation.\8\ In 
general, we are proposing that monitors make a diagnostic decision just 
once per drive cycle that contains operation satisfying the enable 
criteria for the given monitor.
---------------------------------------------------------------------------

    \7\ A ``continuous'' monitor--if used in the context of 
monitoring conditions for circuit continuity, lack of circuit 
continuity, circuit faults, and out-of-range values--means sampling 
at a rate no less than two samples per second. If a computer input 
component is sampled less frequently for engine control purposes, 
the signal of the component may instead be evaluated each time 
sampling occurs.
    \8\ A ``non-continuous'' monitor being a monitor that runs only 
when a limited set of operating conditions occurs.
---------------------------------------------------------------------------

    When a problem is first detected, we are proposing that a 
``pending'' DTC be stored. If, during the subsequent drive cycle that 
contains operation satisfying the enable criteria for the given 
monitor, a problem in the components/system is not again detected, the 
OBD system would declare that a malfunction does not exist and would, 
therefore, erase the pending DTC. However, if, during the subsequent 
drive cycle that contains operation satisfying the enable criteria for 
the given monitor, a problem in the component/system is again detected, 
a malfunction has been confirmed and, hence, a ``confirmed'' or ``MIL-
on'' DTC would be stored.\9\ Section II.F presents the requirements for 
standardization of OBD information and communications. Upon storage of 
a MIL-on DTC and, depending on the communication protocol used--ISO 
15765-4 or SAE J1939--the pending DTC would either remain stored or be 
erased, respectively. Today's proposal neither stipulates which 
communication protocol nor which pending DTC logic be used. We are 
proposing to allow the use of either of the existing protocols as is 
discussed in more detail in section II.F. Upon storage of the MIL-on 
DTC, the MIL must be illuminated.\10\ Also at this time, a 
``permanent'' DTC would be stored (see section II.F.4 for more details 
regarding permanent DTCs and our rationale for proposing them).\11\
---------------------------------------------------------------------------

    \9\ Different industry standards organizations--the Society of 
Automotive Engineers (SAE) and the International Standards 
Organization (ISO)--use different terminology to refer to a ``MIL-
on'' DTC. For clarity, we use the term ``MIL-on'' DTC throughout 
this preamble to convey the concept and not any requirement that 
standard making bodies use the term in their standards.
    \10\ Throughout this proposal, we refer to MIL illumination to 
mean a steady, continuous illumination during engine operation 
unless stated otherwise. This contrasts with the MIL illumination 
logic used by many engine manufacturers today by which the MIL would 
illuminate upon detection of a malfunction but would remain 
illuminated only while the malfunction was actually occurring. Under 
this latter logic, an intermittent malfunction or one that occurs 
under only limited operating conditions may result in a MIL that 
illuminates, extinguishes, illuminates, etc., as operating 
conditions change.
    \11\ A permanent DTC must be stored in a manner such that 
electrical disconnections do not result in their erasure (i.e., they 
must be stored in non-volatile random access memory (NVRAM)).
---------------------------------------------------------------------------

    We are also proposing that, after three subsequent drive cycles 
that contain operation satisfying the enable criteria for the given 
monitor without any recurrence of the previously detected malfunction, 
the MIL should be extinguished (unless there are other MIL-on DTCs 
stored for which the MIL must also be illuminated), the permanent DTC 
should be erased, but a ``previous-MIL-on'' DTC should remain 
stored.\12\ We are proposing that the previous MIL-on DTC remain stored 
for 40 engine warmup cycles after which time, provided the identified 
malfunction has not been detected again and the MIL is presently not 
illuminated for that malfunction, the previous-MIL-on DTC can be 
erased.\13\ However, if an illuminated MIL is not extinguished, or if a 
MIL-on DTC is not erased, by the OBD system itself but is instead 
erased via scan tool or battery disconnect (which would erase all non-
permanent, volatile memory), the permanent DTC must remain stored. This 
way, permanent DTCs can only be erased by the OBD system itself and 
cannot be erased through human interaction with the system.
---------------------------------------------------------------------------

    \12\ This general ``three trip'' condition for extinguishing the 
MIL is true for all but two diesel systems/monitors--the misfire 
monitor and the SCR system--and three gasoline systems/monitors--the 
fuel system, the misfire monitor, and the evaporative system--which 
have further conditions on extinguishing the MIL This is discussed 
in more detail in sections II.B and II.C.
    \13\ For simplicity, the discussion here refers to ``previous-
MIL-on'' DTCs only. The ISO 15765 standard and the SAE J1939 
standard use different terms to refer to the concept of a previous-
MIL-on DTC. Our intent is to present the concept of our proposal in 
this preamble and not to specify the terminology used by these 
standard making bodies.
---------------------------------------------------------------------------

    We are proposing that the manufacturer be allowed, upon

[[Page 3209]]

Administrator approval, to use alternative statistical MIL illumination 
and DTC storage protocols to those described above (i.e., alternatives 
to the ``first trip--pending DTC, second strip--MIL-on DTC logic). The 
Administrator would consider whether the manufacturer provided data 
and/or engineering evaluation adequately demonstrates that the 
alternative protocols can evaluate system performance and detect 
malfunctions in a manner that is equally effective and timely. 
Alternative strategies requiring, on average, more than six driving 
cycles for MIL illumination would probably not be accepted.
    Upon storage of either a pending DTC and/or a MIL-on DTC, we are 
proposing that the computer store a set of ``freeze frame'' data. This 
freeze frame data would provide a snap shot of engine operating 
conditions present at the time the malfunction occurred and was 
detected. This information serves the repair technician in diagnosing 
the problem and conducting the proper repair. The freeze frame data 
should be stored upon storage of a pending DTC. If the pending DTC 
matures to a MIL-on DTC, the manufacturer can choose to update the 
freeze frame data or retain the freeze frame stored in conjunction with 
the pending DTC. Likewise, any freeze frame stored in conjunction with 
any pending or MIL-on DTC should be erased upon erasure of the DTC. 
Further information concerning the freeze frame requirement and the 
data required in the freeze frame is presented in section II.F.4, 
below.
    We are also proposing that the OBD system illuminate the MIL and 
store a MIL-on DTC to inform the vehicle operator whenever the engine 
enters a mode of operation that can affect the performance of the OBD 
system. If such a mode of operation is recoverable (i.e., operation 
automatically returns to normal at the beginning of the following 
ignition cycle \14\), then in lieu of illuminating the MIL when the 
mode of operation is entered, the OBD system may wait to illuminate the 
MIL and store the MIL-on DTC if the mode of operation is again entered 
before the end of the next ignition cycle. We are proposing this 
because many operating strategies are designed such that they continue 
automatically through to the next key-off. Regardless, upon the next 
key-on, the engine control would start off in ``normal'' operating mode 
and would return to the ``abnormal'' operating mode only if the 
condition causing the abnormal mode was again encountered. In such 
cases, we are proposing to allow that the MIL be illuminated during the 
second consecutive drive cycle during which such an ``abnormal'' mode 
is engaged.\15\
---------------------------------------------------------------------------

    \14\ ``Ignition Cycle'' means a drive cycle that begins with 
engine start and includes an engine speed that exceeds 50 to 150 
rotations per minute (rpm) below the normal, warmed-up idle speed 
(as determined in the drive position for vehicles equipped with an 
automatic transmission) for at least two seconds plus or minus one 
second.
    \15\ Note that we use the term ``abnormal'' to refer to an 
operating mode that the engine is designed to enter upon determining 
that ``normal'' operation cannot be maintained. Therefore, the term 
``abnormal'' is somewhat of a misnomer since the engine is doing 
what it has been designed to do. Nonetheless, the abnormal operating 
mode is clearly not the operating mode the manufacturer has intended 
for optimal operation. Such operating modes are sometimes referred 
to as ``default'' operating modes or ``limp-home'' operating modes.
---------------------------------------------------------------------------

    Whether or not the ``abnormal'' mode of operation is recoverable, 
in this context, has nothing to do with whether the detected 
malfunction goes away or stays. Instead, it depends solely on whether 
or not the engine, by design, will stay in abnormal operating mode on 
the next key-on. We are proposing this MIL logic because often the 
diagnostic (i.e., monitor) that caused the engine to enter abnormal 
mode cannot run again once the engine is in the abnormal mode. So, if 
the MIL logic associated with abnormal mode activation was always a 
two-trip diagnostic, abnormal mode activation would set a pending DTC 
on the first trip and, since the system would then be stuck in that 
abnormal operating mode and would never be able to run the diagnostic 
again, the pending DTC could never mature to a MIL-on DTC nor 
illuminate the MIL. Hence, the MIL must illuminate upon the first entry 
into such an abnormal operating mode. If such a mode is recoverable, 
the engine will start at the next key-on in ``normal'' mode allowing 
the monitor to run again and, assuming another detection of the 
condition, the system would set a MIL-on DTC and illuminate the MIL.
    The OBD system would not need to store a DTC nor illuminate the MIL 
upon abnormal mode operation if other telltale conditions would result 
in immediate action by the driver. Such telltale conditions would be, 
for example, an overt indication like a red engine shut-down warning 
light. The OBD system also need not store a DTC nor illuminate the MIL 
upon abnormal mode operation if the mode is indeed an auxiliary 
emission control device (AECD) approved by the Administrator.
    There may be malfunctions of the MIL itself that would prevent it 
from illuminating. A repair technician--or possibly an I/M inspector--
would still be able to determine the status of the MIL (i.e., commanded 
``on'' or ``off'') by reading electronic information available through 
a scan tool, but there would be no indication to the driver of an 
emissions-related malfunction should one occur. Unidentified 
malfunctions may cause excess emissions to be emitted from the vehicle 
and may even cause subsequent deterioration or failure of other 
components or systems without the driver's knowledge. In order to 
prevent this, the manufacturer must ensure that the MIL is functioning 
properly. For this reason, we are proposing two requirements to check 
the functionality of the MIL itself. First, the MIL would be required 
to illuminate for a minimum of five seconds when the vehicle is in the 
key-on, engine-off position. This allows an interested party to check 
the MIL's functionality simply by turning the key to the key-on 
position. While the MIL would be physically illuminated during this 
functional check, the data stream value for the MIL command status 
would be required to indicate ``off'' during this check unless, of 
course, the MIL was currently being commanded ``on'' for a detected 
malfunction. This functional check of the MIL would not be required 
during vehicle operation in the key-on, engine-off position subsequent 
to the initial engine cranking of an ignition cycle (e.g., due to an 
engine stall or other non-commanded engine shutoff).
    The second functional check requirement we are proposing requires 
the OBD system to perform a circuit continuity check of the electrical 
circuit that is used to illuminate the MIL to verify that the circuit 
is not shorted or open (e.g., a burned out bulb). While there would not 
be an ability to illuminate the MIL when such a malfunction is 
detected, the electronically readable MIL command status in the onboard 
computer would be changed from commanded ``off'' to ``on''. This would 
allow the truck owner or fleet maintenance staff to quickly determine 
whether an extinguished MIL means ``no malfunctions'' or ``broken 
MIL.'' It would also serve, should it become of interest in the future, 
complete automation of the I/M process by eliminating the need for 
inspectors to input manually the results of their visual inspections. 
Feedback from passenger car I/M programs indicates that the current 
visual bulb check performed by inspectors is subject to error and 
results in numerous vehicles being falsely failed or passed. By 
requiring monitoring of the circuit itself, the entire pass/fail 
criteria of an I/M program could be determined by the electronic 
information available through a scan tool, thus better facilitating 
quick

[[Page 3210]]

and effective inspections and minimizing the chance for manually-
entered errors.
    At the manufacturer's option, the MIL may be used to indicate 
readiness status in a standardized format (see Section II.F) in the 
key-on, engine-off position. Readiness status is a term used in light-
duty OBD that refers to a vehicle's readiness for I/M inspection. For a 
subset of monitors--those that are non-continuous monitors for which an 
emissions threshold exists (see sections II.B and II.C for more on 
emissions thresholds)--a readiness status indicator must be stored in 
memory to indicate whether or not that particular monitor has run 
enough times to make a diagnostic decision. Until the monitor has run 
sufficient times, the readiness status would indicate ``not ready''. 
Upon running sufficient times, the readiness status would indicate 
``ready.'' This serves to protect against drivers disconnecting their 
battery just prior to the I/M inspection so as to erase any MIL-on 
DTCs. Such an action would simultaneously set all readiness status 
indicators to ``not ready'' resulting in a notice to return to the 
inspection site at a future date. Readiness indicators also help repair 
technicians because, after completing a repair, they can operate the 
vehicle until the readiness status indicates ``ready'' and, provided no 
DTCs are stored, know that the repair has been successful. We are 
proposing that HDOBD systems follow this same readiness status logic as 
used for years in light-duty OBD both to assist repair technicians and 
to facilitate potential future HDOBD I/M programs.
    We are also proposing that the manufacturer, upon Administrator 
approval, be allowed to use the MIL to indicate which, if any, DTCs are 
currently stored (e.g., to ``blink'' the stored codes). The 
Administrator would approve the request if the manufacturer can 
demonstrate that the method used to indicate the DTCs will not be 
unintentionally activated during any inspection test or during routine 
driver operation.
3. Monitoring Conditions
a. Background
    Given that the intent of the proposed OBD requirements is to 
monitor the emission control system for proper operation, it is logical 
that the OBD monitors be designed such that they monitor the emission 
control system during typical driving conditions. While many OBD 
monitors would be designed such that they are continuously making 
decisions about the operational status of the engine, many--and 
arguably the most critical--monitors are not so designed. For example, 
an OBD monitor whose function is to monitor the active fuel injection 
system of a NOX adsorber or a DPF cannot be continuously 
monitoring that function since that function occurs on an infrequent 
basis. This OBD monitor presumably would be expected to ``run,'' or 
evaluate the active injection system, during an actual fuel injection 
event.
    For this reason, manufacturers are allowed to determine the most 
appropriate times to run their non-continuous OBD monitors. This way, 
they are able to make an OBD evaluation either at the operating 
condition when an emission control system is active and its operational 
status can best be evaluated, and/or at the operating condition when 
the most accurate evaluation can be made (e.g., highly transient 
conditions or extreme conditions can make evaluation difficult). 
Importantly, manufacturers are prohibited from using a monitoring 
strategy that is so restrictive such that it rarely or never runs. To 
help protect against monitors that rarely run, we are proposing an 
``in-use monitor performance ratio'' requirement which is detailed in 
section II.E.
    The set of operating conditions that must be met so that an OBD 
monitor can run are called the ``enable criteria'' for that given 
monitor. These enable criteria are often different for different 
monitors and may well be different for different types of engines. A 
large diesel engine intended for use in a Class 8 truck would be 
expected to see long periods of relatively steady-state operation while 
a smaller engine intended for use in an urban delivery truck would be 
expected to see a lot of transient operation. Manufacturers will need 
to balance between a rather loose set of enable criteria for their 
engines and vehicles given the very broad range of operation HD highway 
engines see and a tight set of enable criteria given the desire for 
greater monitor accuracy.
b. General Monitoring Conditions
i. Monitoring Conditions for All Engines
    As guidance to manufacturers, we are proposing the following 
criteria to assist manufacturers in developing their OBD enable 
criteria. These criteria would be used by the Agency during our OBD 
certification approval process to ensure that monitors run on a 
frequent basis during real world driving conditions. These criteria 
would be:
     The monitors should run during conditions that are 
technically necessary to ensure robust detection of malfunctions (e.g., 
to avoid false passes and false indications of malfunctions);
     The monitor enable criteria should ensure monitoring will 
occur during normal vehicle operation; and,
     Monitoring should occur during at least one test used by 
EPA for emissions verification `` either the HD Federal Test Procedure 
(FTP) transient cycle, or the Supplementary Emissions Test (SET).\16\
---------------------------------------------------------------------------

    \16\ See 40 CFR part 86, subpart N for details of EPA's test 
procedures.
---------------------------------------------------------------------------

    As discussed in more detail in sections II.B through II.D, we are 
proposing that manufacturers define the monitoring conditions, subject 
to Administrator approval, for detecting the malfunctions required by 
this proposal. The Administrator would determine if the monitoring 
conditions proposed by the manufacturer for each monitor abide by the 
above criteria.
    In general, except as noted in sections II.B through II.D, the 
proposed regulation would require each monitor to run at least once per 
driving cycle in which the applicable monitoring conditions are met. 
The proposal would also require certain monitors to run continuously 
throughout the driving cycle. These include a few threshold monitors 
(e.g., fuel system monitor) and most circuit continuity monitors. While 
a basic definition of a driving cycle (e.g., from ignition key-on and 
engine startup to engine shutoff) has been sufficient for passenger 
cars, the driving habits of many types of vehicles in the heavy-duty 
industry dictate an alternate definition. Specifically, many heavy-duty 
operators will start the engine and leave it running for an entire day 
or, in some cases, even longer. As such, we are proposing that any 
period of continuous engine-on operation of four hours be considered a 
complete driving cycle. A new driving cycle would begin following such 
a four hour period, regardless of whether or not the engine had been 
shut down. Thus, the ``clock'' for monitors that are required to run 
once per driving cycle would be reset to run again (in the same key-on 
engine start or trip) once the engine has been operated beyond four 
hours continuously. This would avoid an unnecessary delay in detection 
of malfunctions simply because the heavy-duty vehicle operator has 
elected to leave the vehicle running continuously for an entire day or 
days at a time.
    Manufacturers may request Administrator approval to define 
monitoring conditions that are not encountered during the FTP cycle. In 
evaluating the manufacturer's request, the Administrator will consider 
the degree to which the requirement to run

[[Page 3211]]

during the FTP cycle restricts in-use monitoring, the technical 
necessity for defining monitoring conditions that are not encountered 
during the FTP cycle, data and/or an engineering evaluation submitted 
by the manufacturer which demonstrate that the component/system does 
not normally function, or monitoring is otherwise not feasible, during 
the FTP cycle, and, where applicable, the ability of the manufacturer 
to demonstrate that the monitoring conditions will satisfy the minimum 
acceptable in-use monitor performance ratio requirement as defined 
below.
ii. In-Use Performance Tracking Monitoring Conditions
    In addition to the general monitoring conditions above, we are 
proposing that manufacturers be required to implement software 
algorithms in the OBD system to individually track and report in-use 
performance of the following monitors in the standardized format 
specified in section II.E:
 Diesel NMHC converting catalyst(s)
 Diesel NOX converting catalyst(s)
 Gasoline catalyst(s)
 Exhaust gas sensor(s)
 Gasoline evaporative system
 Exhaust gas recirculation (EGR) system
 Variable valve timing (VVT) system
 Gasoline secondary air system
 Diesel particulate filter system
 Diesel boost pressure control system
 Diesel NOX adsorber(s)
    The OBD system is not required to track and report in-use 
performance for monitors other than those specifically identified 
above.
iii. In-Use Performance Ratio Requirement
    We are also proposing that, for all 2013 and subsequent model year 
engines, manufacturers be required to define monitoring conditions 
that, in addition to meeting the general monitoring conditions, ensure 
that certain monitors yield an in-use performance ratio (which monitors 
and the details that define the performance ratio are defined in 
section II.E) that meets or exceeds the minimum acceptable in-use 
monitor performance ratio for in-use vehicles. We are proposing a 
minimum acceptable in-use monitor performance ratio of 0.100 for all 
monitors specifically required to track in-use performance. This means 
that the monitors listed in section II.A.3.ii above must run and make 
valid diagnostic decisions during 10 percent of the vehicle's trips. We 
intend to work with industry during the initial years of implementation 
to gather data on in-use performance ratios and may revise this ratio 
lower as appropriate depending on what we learn.
    Note that manufacturers may not use the calculated ratio (or any 
element thereof), or any other indication of monitor frequency, as a 
monitoring condition for a monitor. For example, the manufacturer would 
not be allowed to use a low ratio to enable more frequent monitoring 
through diagnostic executive priority or modification of other 
monitoring conditions, or to use a high ratio to enable less frequent 
monitoring.
4. Determining the Proper OBD Malfunction Criteria
    For determining the malfunction criteria for diesel engine monitors 
associated with an emissions threshold (see sections II.B and II.C for 
more on emissions thresholds), we are proposing that manufacturers be 
required to determine the appropriate emissions test cycle such that 
the most stringent monitor would result. In general, we believe that 
manufacturers can make this determination based on engineering 
judgement, but there may be situations where testing would be required 
to make the determination. We do not necessarily anticipate challenging 
a manufacturer's determination of which test cycle to use. Nonetheless, 
the manufacturer should be prepared, perhaps with test data, to justify 
their determination.
    We are also proposing that, for engines equipped with emission 
controls that experience infrequent regeneration events (e.g., a DPF 
and/or a NOX adsorber), a manufacturer must adjust the 
emission test results for monitors that are required to indicate a 
malfunction before emissions exceed a certain emission threshold.\17\ 
For each such monitor, the manufacturer would have to adjust the 
emission result as done in accordance with the provisions of section 
86.004-28(i) with the component for which the malfunction criteria are 
being established having been deteriorated to the malfunction 
threshold. As proposed, the adjusted emission value must be used for 
purposes of determining whether or not the applicable emission 
threshold is exceeded.
---------------------------------------------------------------------------

    \17\ See proposed Sec.  86.010-18(f).
---------------------------------------------------------------------------

    While we believe that this adjustment process for monitors of 
systems that experience infrequent regeneration events makes sense and 
would result in robust monitors, we also believe that it could prove to 
be overly burdensome for manufacturers. For example, a NOX 
adsorber threshold being evaluated by running an FTP using a 
``threshold'' part (i.e., a NOX adsorber deteriorated such 
that tailpipe emissions are at the applicable thresholds) may be 
considered acceptable provided the NOX adsorber does not 
regenerate during the test, but it may be considered unacceptable if 
the NOX adsorber does happen to regenerate during the test. 
This could happen because emissions would be expected to increase 
slightly during the regeneration event thereby causing emissions to be 
slightly above the applicable threshold. This would require the 
manufacturer to recalibrate the NOX adsorber monitor to 
detect at a lower level of deterioration to ensure that a regeneration 
event would not cause an exceedance of the threshold during an 
emissions test. After such a recalibration, the emissions occurring 
during the regeneration event would be lower than before because the 
new ``threshold'' NOX adsorber would have a slightly higher 
conversion efficiency. We are concerned that manufacturers may find 
themselves in a difficult iterative process calibrating such monitors 
that, in the end, will not be correspondingly more effective.
    For this reason, we request comment regarding the burden associated 
with the need to consider regeneration events in determining compliance 
with emissions thresholds. We also request comment on how to address 
any environmental concern versus the burden. Would it perhaps be best 
to simply use the emissions adjustments that are determined in 
accordance with section 86.004-28(i)? Is it necessary to even consider 
regeneration emissions when determining emission threshold compliance 
or is it perhaps best to ignore regeneration events in determining 
threshold calibrations?

B. Monitoring Requirements and Timelines for Diesel-Fueled/Compression-
Ignition Engines

    Table II.B-1 summarizes the proposed diesel fueled compression 
ignition emissions thresholds at which point a component or system has 
failed to the point of requiring an illuminated MIL and a stored DTC. 
More detail regarding the specific monitoring requirements, 
implementation schedules, and liabilities can be found in the sections 
that follow.

[[Page 3212]]



          Table II.B-1.--Proposed Emissions Thresholds for Diesel Fueled CI Engines over 14,000 Pounds
----------------------------------------------------------------------------------------------------------------
                     Component/monitor                           MY         NMHC      CO      NOX         PM
----------------------------------------------------------------------------------------------------------------
NMHC catalyst system......................................     2010-2012     2.5x  .......  .......  ...........
                                                                   2013+       2x  .......  .......  ...........
NOX catalyst system.......................................         2010+  .......  .......     +0.3  ...........
DPF system................................................     2010-2012     2.5x  .......  .......   0.05/+0.04
                                                                   2013+       2x  .......  .......   0.05/+0.04
Air-fuel ratio sensors upstream...........................     2010-2012     2.5x     2.5x     +0.3   0.03/+0.02
                                                                   2013+       2x       2x     +0.3   0.03/+0.02
Air-fuel ratio sensors downstream.........................     2010-2012     2.5x  .......     +0.3   0.05/+0.04
                                                                   2013+       2x  .......     +0.3   0.05/+0.04
NOX sensors...............................................         2010+  .......  .......     +0.3   0.05/+0.04
``Other monitors'' with emissions thresholds (see section      2010-2012     2.5x     2.5x     +0.3   0.03/+0.02
 II.B)....................................................
                                                                   2013+       2x       2x     +0.3   0.03/+0.02
----------------------------------------------------------------------------------------------------------------
Notes: MY=Model Year; 2.5x means a multiple of 2.5 times the applicable emissions standard or family emissions
  limit (FEL); +0.3 means the standard or FEL plus 0.3; 0.05/+0.04 means an absolute level of 0.05 or an
  additive level of the standard or FEL plus 0.04, whichever level is higher; not all proposed monitors have
  emissions thresholds but instead rely on functionality and rationality checks as described in section II.D.4.

    There are exceptions to the emissions thresholds shown in Table 
II.B-1 whereby a manufacturer can demonstrate that emissions do not 
exceed the threshold even when the component or system is non-
functional at which point a functional check would be allowed.
    Note that, in general, the monitoring strategies designed to meet 
the requirements discussed below should not involve the alteration of 
the engine control system or the emissions control system such that 
tailpipe emissions would increase. We do not want emissions to 
increase, even for short durations, for the sole purpose of monitoring 
the systems intended to control emissions. The Administrator would 
consider such monitoring strategies on a case-by-case basis taking into 
consideration the emissions impact and duration of the monitoring 
event. However, much effort has been expended in recent years to 
minimize engine operation that results in increased emissions and we 
encourage manufacturers to develop monitoring strategies that do not 
require alteration of the basic control system.
1. Fuel System Monitoring
a. Background
    The fuel system of a diesel engine is an essential component of the 
engine's emissions control system. Proper delivery of fuel--quantity, 
pressure, and timing--can play a crucial role in maintaining low 
engine-out emissions. The performance of the fuel system is also 
critical for aftertreatment device control strategies. As such, 
thorough monitoring of the fuel system is an essential element in an 
OBD system. The fuel system is primarily comprised of a fuel pump, fuel 
pressure control device, and fuel injectors. Additionally, the fuel 
system generally has sophisticated control strategies that utilize one 
or more feedback sensors to ensure the proper amount of fuel is being 
delivered to the cylinders. While gasoline engines have undergone 
relatively minor hardware changes (but substantial fine-tuning in the 
control strategy and feedback inputs), diesel engines have more 
recently undergone substantial changes to the fuel system hardware and 
now incorporate more refined control strategies and feedback inputs.
    For diesel engines, a substantial change has occurred in recent 
years as manufacturers have transitioned to new high-pressure fuel 
systems. One of the most widely used is a high-pressure common-rail 
fuel injection system, which is generally comprised of a high-pressure 
fuel pump, a fuel rail pressure sensor, a common fuel rail that feeds 
all injectors, individual fuel injectors that directly control fuel 
injection quantity and timing for each cylinder, and a closed-loop 
feedback system that uses the fuel rail pressure sensor to achieve the 
commanded fuel rail pressure. Unlike older style fuel systems where 
fuel pressure was mechanically linked to engine speed (and thus, varied 
from low to high as engine speed increased), common-rail systems are 
capable of controlling fuel pressure independent of engine speed. This 
increase in fuel pressure control allows greater flexibility in 
optimizing the performance and emission characteristics of the engine. 
The ability of the system to generate high pressure independent of 
engine speed also improves fuel delivery at low engine speeds.
    Precise control of the fuel injection timing is crucial for optimal 
engine and emission performance. As injection timing is advanced (i.e., 
fuel injection occurs earlier), hydrocarbon (HC) emissions and fuel 
consumption are decreased but oxides of nitrogen (NOX) 
emissions are increased. As injection timing is retarded (i.e., fuel 
injection occurs later), NOX emissions can be reduced but HC 
emissions, particulate matter (PM) emissions, and fuel consumption 
increase. Most modern diesel fuel systems even provide engine 
manufacturers with the ability to separate a single fuel injection 
event into discrete events such as pilot (or pre) injection, main 
injection, and post injection.
    Given the important role that modern diesel fuel systems play in 
emissions control, malfunctions or deterioration that would affect the 
fuel pressure control, injection timing, pilot/main/post injection 
timing or quantity, or ability to accurately perform rate-shaping could 
lead to substantial increases in emissions (primarily NOX or 
PM), often times with an associated change in fuel consumption.
b. Fuel System Monitoring Requirements
    We are proposing that the OBD system monitor the fuel delivery 
system to verify that it is functioning properly. The fuel system 
monitor would be required to monitor for malfunctions in the injection 
pressure control, injection quantity, injection timing, and feedback 
control (if equipped). The individual electronic components (e.g., 
actuators, valves, sensors, pumps) that are used in the fuel system and 
not specifically addressed in this section shall be monitored in 
accordance with the comprehensive component requirements in section 
II.D.4.
i. Fuel System Pressure Control
    We are proposing that the OBD system continuously monitor the fuel 
system's ability to control to the desired fuel pressure. The OBD 
system would have to detect a malfunction of the fuel system's pressure 
control system when

[[Page 3213]]

the pressure control system is unable to maintain an engine's emissions 
at or below the emissions thresholds for ``other monitors'' as shown in 
Table II.B-1. For engines in which no failure or deterioration of the 
fuel system pressure control could result in an engine's emissions 
exceeding the applicable emissions thresholds, the OBD system would be 
required to detect a malfunction when the system has reached its 
control limits such that the commanded fuel system pressure cannot be 
delivered.
ii. Fuel System Injection Quantity
    We are proposing that the OBD system detect a malfunction of the 
fuel injection system when the system is unable to deliver the 
commanded quantity of fuel necessary to maintain an engine's emissions 
at or below the emissions thresholds for ``other monitors'' as shown in 
Table II.B-1. For engines in which no failure or deterioration of the 
fuel injection quantity could result in an engine's emissions exceeding 
the applicable emissions thresholds, the OBD system would be required 
to detect a malfunction when the system has reached its control limits 
such that the commanded fuel quantity cannot be delivered.
iii. Fuel System Injection Timing
    We are proposing that the OBD system detect a malfunction of the 
fuel injection system when the system is unable to deliver fuel at the 
proper crank angle/timing (e.g., injection timing too advanced or too 
retarded) necessary to maintain an engine's emissions at or below the 
emissions thresholds for ``other monitors'' as shown in Table II.B-1. 
For engines in which no failure or deterioration of the fuel injection 
timing could result in an engine's emissions exceeding the applicable 
emissions thresholds, the OBD system would be required to detect a 
malfunction when the system has reached its control limits such that 
the commanded fuel injection timing cannot be achieved.
iv. Fuel System Feedback Control
    If the engine is equipped with feedback control of the fuel system 
(e.g., feedback control of pressure or pilot injection quantity), we 
are proposing that the OBD system detect a malfunction when and if:
     The system fails to begin feedback control within a 
manufacturer specified time interval;
     A failure or deterioration causes open loop or default 
operation; or
     Feedback control has used up all of the adjustment allowed 
by the manufacturer.
    A manufacturer may temporarily disable monitoring for malfunctions 
where the feedback control has used up all of the adjustment allowed by 
the manufacturer during conditions that the monitor cannot distinguish 
robustly between a malfunctioning system and a properly operating 
system. To do so, the manufacturer would be required to submit data 
and/or engineering analyses demonstrating that the control system, when 
operating as designed on an engine with all emission controls working 
properly, routinely operates during these conditions with all of the 
adjustment allowed by the manufacturer used up. In lieu of detecting, 
with a fuel system specific monitor, when the system fails to begin 
feedback control within a manufacturer specified time interval and/or 
when a failure or deterioration causes open loop or default operation, 
the OBD system may monitor the individual parameters or components that 
are used as inputs for fuel system feedback control provided that the 
monitors detect all malfunctions related to feedback control.
c. Fuel System Monitoring Conditions
    The OBD system would be required to monitor continuously for 
malfunctions of the fuel pressure control and feedback control. 
Manufacturers would be required to define the monitoring conditions for 
malfunctions of the injection quantity and injection timing such that 
the minimum performance ratio requirements discussed in section II.E 
would be met.
d. Fuel System MIL Illumination and DTC Storage
    We are proposing the general MIL illumination and DTC storage 
requirements as discussed in section II.A.2.
2. Engine Misfire Monitoring
a. Background
    Misfire, the lack of combustion in the cylinder, causes increased 
engine-out hydrocarbon emissions. On gasoline engines, misfire results 
from the absence of spark, poor fuel metering, and poor compression. 
Further, misfire can be intermittent on gasoline engines (e.g., the 
misfire only occurs under certain engine speeds or loads). 
Consequently, our existing under 14,000 pound OBD regulation requires 
continuous monitoring for misfire malfunctions on gasoline engines.
    In contrast, manufacturers have historically maintained that a 
diesel engine with traditional diesel technology misfires only due to 
poor compression (e.g., worn valves or piston rings, improper injector 
or glow plug seating). They have also maintained that, when poor 
compression results in a misfiring cylinder, the cylinder will misfire 
under all operating conditions rather than only some operating 
conditions. For that reason, our existing under 14,000 pound OBD 
regulation has not required continuous monitoring for misfire 
malfunctions on diesel engines.
    However, with the increased use of EGR and its use to varying 
degrees at different speeds and load, and with emerging technologies 
such as homogeneous charge compression ignition (HCCI), we believe that 
the conventional wisdom regarding diesel engines and misfires no longer 
holds true. These newer technologies may indeed result in misfires that 
are intermittent, spread out among various cylinders, and that only 
happen at certain speeds and loads.
b. Misfire Monitoring Requirements
    We are proposing that the OBD system monitor the engine for misfire 
causing excess emissions. The OBD system must be capable of detecting 
misfire occurring in one or more cylinders. To the extent possible 
without adding hardware for this specific purpose, the OBD system must 
also identify the specific misfiring cylinder. If more than one 
cylinder is continuously misfiring, a separate DTC must be stored 
indicating that multiple cylinders are misfiring. When identifying 
multiple cylinder misfire, the OBD system is not required to also 
identify each of the continuously misfiring cylinders individually 
through separate DTCs.
    For 2013 and subsequent model year engines, we are proposing a more 
stringent requirement that the OBD system detect a misfire malfunction 
causing emissions to exceed the emissions thresholds for ``other 
monitors'' as shown in Table II.B-1. This requirement to detect engine 
misfire prior to exceeding an emissions threshold would apply only to 
those engines equipped with sensors capable of detecting combustion or 
combustion quality (e.g., cylinder pressure sensors used in homogeneous 
charge compression ignition (HCCI) control systems). Engines without 
such sensors would have to detect only when one or more cylinders are 
continually misfiring.
    To determine what level of misfire would cause emissions to exceed 
the applicable emissions thresholds, we are proposing that 
manufacturers determine

[[Page 3214]]

the percentage of misfire evaluated in 1000 revolution increments that 
would cause emissions from an emission durability demonstration engine 
to exceed the emissions thresholds if the percentage of misfire were 
present from the beginning of the test. To establish this percentage of 
misfire, the manufacturer would utilize misfire events occurring at 
equally spaced, complete engine cycle intervals, across randomly 
selected cylinders throughout each 1000-revolution increment. If this 
percentage of misfire is determined to be lower than one percent, the 
manufacturer may set the malfunction criteria at one percent. Any 
malfunction should be detected if the percentage of misfire established 
via this testing is exceeded regardless of the pattern of misfire 
events (e.g., random, equally spaced, continuous).
    The manufacturer may employ other revolution increments besides the 
1000 revolution increment being proposed. To do so, the manufacturer 
would need to demonstrate that the strategy would be equally effective 
and timely in detecting misfire.
c. Engine Misfire Monitoring Conditions
    For engines without combustion sensors, we are proposing that the 
OBD system monitor for misfire during engine idle conditions at least 
once per drive cycle in which the monitoring conditions for misfire are 
met. The manufacturer would be required to define monitoring 
conditions, supported by manufacturer-submitted data and/or engineering 
analyses, that demonstrate that the monitoring conditions: are 
technically necessary to ensure robust detection of malfunctions (e.g., 
avoid false passes and false detection of malfunctions); require no 
more than 1000 cumulative engine revolutions; and, do not require any 
single continuous idle operation of more than 15 seconds to make a 
determination that a malfunction is present (e.g., a decision can be 
made with data gathered during several idle operations of 15 seconds or 
less).
    For 2013 and subsequent model year engines with combustion sensors, 
we are proposing that the OBD system continuously monitor for misfire 
under all positive torque engine speeds and load conditions. If a 
monitoring system cannot detect all misfire patterns under all positive 
torque engine speeds and load conditions, the manufacturer may request 
that the Administrator approve the monitoring system nonetheless. In 
evaluating the manufacturer's request, the Administrator would consider 
the following factors: the magnitude of the region(s) in which misfire 
detection is limited; the degree to which misfire detection is limited 
in the region(s) (i.e., the probability of detection of misfire 
events); the frequency with which said region(s) are expected to be 
encountered in-use; the type of misfire patterns for which misfire 
detection is troublesome; and demonstration that the monitoring 
technology employed is not inherently incapable of detecting misfire 
under required conditions (i.e., compliance can be achieved on other 
engines). The evaluation would be based on the following misfire 
patterns: equally spaced misfire occurring on randomly selected 
cylinders; single cylinder continuous misfire; and, paired cylinder 
(cylinders firing at the same crank angle) continuous misfire.
d. Engine Misfire MIL Illumination and DTC Storage
    For engines without combustion sensors, we are proposing the 
general MIL illumination and DTC storage requirements as discussed in 
section II.A.2.
    For 2013 and subsequent model year engines with combustion sensors, 
we are proposing that, after four detections of the percentage of 
misfire that would cause emissions to exceed the applicable emissions 
thresholds during a single driving cycle, a pending DTC would be 
stored. If a pending DTC is stored, the OBD system would be required to 
illuminate the MIL and store a MIL--on DTC if the percentage of misfire 
is again exceeded four times during either: the driving cycle 
immediately following the storage of the pending DTC, regardless of the 
conditions encountered during the driving cycle; or, the next driving 
cycle in which similar conditions are encountered to the engine 
conditions that occurred when the pending DTC was stored.\18\ For 
erasure of the pending DTC, we are proposing if, by the end of the next 
driving cycle in which similar conditions have been encountered to the 
engine conditions that occurred when the pending DTC was stored without 
an exceedance of the specified percentage of misfire, the pending DTC 
may be erased. The pending DTC may also be erased if similar conditions 
are not encountered during the next 80 driving cycles immediately 
following initial detection of the malfunction.
---------------------------------------------------------------------------

    \18\ ``Similar conditions,'' as used in conjunction with misfire 
and fuel system monitoring, means engine conditions having an engine 
speed within 375 rpm, load conditions within 20 percent, and the 
same warm up status (i.e., cold or hot) as existing during the 
applicable previous problem detection. The Administrator may approve 
other definitions of similar conditions based on comparable 
timeliness and reliability in detecting similar engine operation.
---------------------------------------------------------------------------

    We are proposing some specific items with respect to freeze frame 
storage associated with engine misfire. The OBD system shall store and 
erase freeze frame conditions either in conjunction with storing and 
erasing a pending DTC or in conjunction with storing a MIL--on DTC and 
erasing a MIL--on DTC. In addition to those proposed requirements 
discussed in section II.A.2, we are proposing that, if freeze frame 
conditions are stored for a malfunction other than a misfire 
malfunction when a DTC is stored, the previously stored freeze frame 
information shall be replaced with freeze frame information regarding 
the misfire malfunction (i.e., the misfire's freeze frame information 
should take precedence over freeze frames for other malfunctions). 
Further, we are proposing that, upon detection of misfire, the OBD 
system store the following engine conditions: engine speed, load, and 
warm up status of the first misfire event that resulted in the storage 
of the pending DTC.
    Lastly, we are proposing that the MIL may be extinguished after 
three sequential driving cycles in which similar conditions have been 
encountered without an exceedance of the specified percentage of 
misfire.
3. Exhaust Gas Recirculation (EGR) System Monitoring
a. Background
    Exhaust gas recirculation (EGR) systems are currently being used by 
many heavy-duty engine manufacturers to meet the 2.5 g/bhp-hr 
NOX+NMHC standard for 2004 and later model year engines. (65 
FR 59896) EGR reduces NOX emissions in several ways. First, 
the recirculated exhaust gases dilute the intake air--i.e., oxygen in 
the fresh air is displaced with relatively non-reactive exhaust gases--
which, in turn, results in less oxygen to form NOX. Second, 
EGR absorbs heat from the combustion process which reduces combustion 
chamber temperatures which, in turn, reduces NOX formation. 
The amount of heat absorbed from the combustion process is a function 
of EGR flow rate and recirculated gas temperature, both of which are 
controlled to minimize NOX emissions. An EGR cooler can be 
added to the EGR system to lower the recirculated gas temperature which 
further enhances NOX control. We fully expect that 2007 and 
later model year engines will continue to make use of cooled EGR 
systems.
    While in theory the EGR system simply routes some exhaust gas back 
to the intake, production systems can be complex and involve many 
components to ensure accurate control of EGR flow

[[Page 3215]]

to maintain acceptable PM and NOX emissions while minimizing 
effects on fuel economy. To control EGR flow rates, EGR systems 
normally use the following components: an EGR valve, valve position 
sensor, boost pressure sensor, intake temperature sensor, intake 
(fresh) airflow sensor, and tubing or piping to connect the various 
components of the system. EGR temperature sensors and exhaust 
backpressure sensors can also be used. Additionally, some systems use a 
variable geometry turbocharger to provide the backpressure necessary to 
drive the EGR flow. Therefore, EGR is not a stand alone emission 
control device. Rather, it is carefully integrated with the air 
handling system (turbocharging and intake cooling) to control 
NOX while not adversely affecting PM emissions and fuel 
economy.
b. EGR System Monitoring Requirements
    We are proposing that the OBD system monitor the EGR system on 
engines so equipped for low EGR flow rate, high EGR flow rate, and slow 
EGR flow response malfunctions. For engines so equipped, we are 
proposing that the EGR feedback control be monitored. Also, for engines 
equipped with EGR coolers (e.g., heat exchangers), the OBD system would 
have to monitor the cooler for malfunctions associated with 
insufficient EGR cooling. The individual electronic components (e.g., 
actuators, valves, sensors) that are used in the EGR system would be 
monitored in accordance with the comprehensive component requirements 
presented in section II.D.4.
i. EGR Low Flow Malfunctions
    We are proposing that the OBD system detect a malfunction prior to 
a decrease from the manufacturer's specified EGR flow rate that would 
cause an engine's emissions to exceed the emissions thresholds for 
``other monitors'' as shown in Table II.B-1. For engines in which no 
failure or deterioration of the EGR system that causes a decrease in 
flow could result in an engine's emissions exceeding the applicable 
emissions thresholds, the OBD system would have to detect a malfunction 
when the system has reached its control limits such that it cannot 
increase EGR flow to achieve the commanded flow rate.
ii. EGR High Flow Malfunctions
    We are proposing that the OBD system detect a malfunction of the 
EGR system, including a leaking EGR valve--i.e., exhaust gas flowing 
through the valve when the valve is commanded closed--prior to an 
increase from the manufacturer's specified EGR flow rate that would 
cause an engine's emissions to exceed the emissions thresholds for 
``other monitors'' as shown in Table II.B-1. For engines in which no 
failure or deterioration of the EGR system that causes an increase in 
flow could result in an engine's emissions exceeding the applicable 
emissions thresholds, the OBD system would have to detect a malfunction 
when the system has reached its control limits such that it cannot 
reduce EGR flow to achieve the commanded flow rate.
iii. EGR Slow Response Malfunctions
    We are proposing that the OBD system detect a malfunction of the 
EGR system prior to any failure or deterioration in the capability of 
the EGR system to achieve the commanded flow rate within a 
manufacturer-specified time that would cause an engine's emissions to 
exceed the emissions thresholds for ``other monitors'' as shown in 
Table II.B-1. The OBD system would have to monitor both the capability 
of the EGR system to respond to a commanded increase in flow and the 
capability of the EGR system to respond to a commanded decrease in 
flow.
iv. EGR Feedback Control
    We are proposing that the OBD system on any engine equipped with 
feedback control of the EGR system (e.g., feedback control of flow, 
valve position, pressure differential across the valve via intake 
throttle or exhaust backpressure), detect a malfunction when and if:
     The system fails to begin feedback control within a 
manufacturer specified time interval;
     A failure or deterioration causes open loop or default 
operation; or
     Feedback control has used up all of the adjustment allowed 
by the manufacturer.
v. EGR Cooler Performance
    We are proposing that the OBD system detect a malfunction of the 
EGR cooler prior to a reduction from the manufacturer's specified 
cooling performance that would cause an engine's emissions to exceed 
the emissions thresholds for ``other monitors'' as shown in Table II.B-
1. For engines in which no failure or deterioration of the EGR cooler 
could result in an engine's emissions exceeding the applicable 
emissions thresholds, the OBD system would have to detect a malfunction 
when the system has no detectable amount of EGR cooling.
c. EGR System Monitoring Conditions
    We are proposing that the OBD system monitor continuously for low 
EGR flow, high EGR flow, and feedback control malfunctions. 
Manufacturers would be required to define the monitoring conditions for 
EGR slow response malfunctions such that the minimum performance ratio 
requirements discussed in section II.E would be met with the exception 
that monitoring must occur every time the monitoring conditions are met 
during the driving cycle in lieu of once per driving cycle as required 
for most monitors. For purposes of tracking and reporting as required 
in section II.E, all monitors used to detect EGR slow response 
malfunctions must be tracked separately but reported as a single set of 
values as specified in section II.E.\19\
---------------------------------------------------------------------------

    \19\ For specific components or systems that have multiple 
monitors that are required to be reported (e.g., exhaust gas sensor 
bank 1 may have multiple monitors for sensor response or other 
sensor characteristics), the OBD system must separately track 
numerators and denominators for each of the specific monitors and 
report only the corresponding numerator and denominator for the 
specific monitor that has the lowest numerical ratio. If two or more 
specific monitors have identical ratios, the corresponding numerator 
and denominator for the specific monitor that has the highest 
denominator shall be reported for the specific component.
---------------------------------------------------------------------------

    Manufacturersmay temporarily disable the EGR system check under 
specific conditions (e.g., when freezing may affect performance of the 
system). To do so, the manufacturer would be required to submit data 
and/or engineering analyses that demonstrate that a reliable check 
cannot be made when these specific conditions exist.
d. EGR System MIL Illumination and DTC Storage
    We are proposing the general requirements for MIL illumination and 
DTC storage as discussed in section II.A.2.
4. Turbo Boost Control System Monitoring
a. Background
    Turbochargers are used on internal combustion engines to enhance 
performance by increasing the density of the intake air. Some of the 
benefits of turbocharging include increased horsepower, improved fuel 
economy, and decreased exhaust smoke. Most modern diesel engines take 
advantage of these benefits and are equipped with turbocharging 
systems. Moreover, smaller turbocharged diesel engines can be used in 
place of larger non-turbocharged engines to achieve the desired engine 
performance characteristics.

[[Page 3216]]

    Exhaust gases passing through the turbine cause it to spin which, 
in turn, causes an adjacent centrifugal pump on the same rotating shaft 
to spin. The spinning pump serves to compress the intake air thereby 
increasing its density. Typically, a boost pressure sensor is located 
in the intake manifold to provide a feedback signal of the current 
intake manifold pressure. As turbo speed (boost) increases, the 
pressure in the intake manifold also increases.
    Proper boost control is essential to optimize emission levels. Even 
short periods of over-or under-boost can result in undesired air-fuel 
ratio excursions and corresponding emission increases. Additionally, 
the boost control system directly affects exhaust and intake manifold 
pressures. Another critical emission control system, EGR, is very 
dependent on these two pressures and generally uses the differential 
between them to force exhaust gas into the intake manifold. If the 
boost control system is not operating correctly, the exhaust or intake 
pressures may not be as expected and the EGR system may not function as 
designed. In high-pressure EGR systems, higher exhaust pressures will 
generate more EGR flow and, conversely, lower pressures will reduce EGR 
flow. A malfunction that causes excessive exhaust pressures (e.g., 
wastegate stuck closed at high engine speed) can produce higher EGR 
flowrates at high load conditions and have a negative impact on 
emissions.
    Manufacturers commonly use charge air coolers to maximize the 
benefits of turbocharging and to control NOX emissions. As 
the turbocharger compresses the intake air, the temperature of that 
intake air increases. This increasing air temperature causes the air to 
expand, which conflicts with one of the goals of turbocharging which is 
to increase charge air density. Charge air coolers are used to exchange 
heat between the compressed air and ambient air (or coolant) and cool 
the compressed air. Accordingly, a decrease in charge air cooler 
performance can affect emissions by causing higher intake air 
temperatures that can lead to higher combustion temperatures and higher 
NOX emissions.
    One drawback of turbocharging is known as turbo lag. Turbo lag 
occurs when the driver attempts to accelerate quickly from a low engine 
speed. Since the turbocharger is a mechanical device, a delay exists 
from the driver demand for more boost until the exhaust flow can 
physically speed up the turbocharger enough to deliver that boost. In 
addition to a negative effect on driveability and performance, improper 
fueling (e.g., over-fueling) during this lag can cause emission 
increases (typically PM).
    To decrease the effects of turbo lag, manufacturers design turbos 
that spool up quickly at low engine speeds and low exhaust flowrates. 
However, designing a turbo that will accelerate quickly from a low 
engine speed but will not result in an over-speed/over-boost condition 
at higher engine speeds is challenging. That is, as the engine speed 
and exhaust flowrates near their maximum, the turbo speed increases to 
levels that cause excessive boost pressures and heat that could lead to 
engine or turbo damage. To prevent excessive turbine speeds and boost 
pressures at higher engine speeds, a wastegate is often used to bypass 
part of the exhaust stream around the turbocharger. The wastegate valve 
is typically closed at lower engine speeds so that all exhaust is 
directed through the turbocharger, thus providing quick response from 
the turbocharger when the driver accelerates quickly from low engine 
speeds. The wastegate is then opened at higher engine speeds to prevent 
engine or turbo damage from an over-speed condition.
    An alternative to a wastegate is the variable geometry 
turborcharger (VGT). To prevent over-boost conditions and to decrease 
turbo lag, VGTs are designed such that the geometry of the turbocharger 
changes with engine speed. While various physical mechanisms are used 
to achieve the variable geometry, the overall result is essentially the 
same. At low engine speeds, the exhaust gas into the turbo is 
restricted in a manner that maximizes the use of the available energy 
to spin the turbo. This allows the turbo to spool up quickly and 
provide good acceleration response. At higher engine speeds, the turbo 
geometry changes such that exhaust gas flow into the turbo is not as 
restricted. In this configuration, more exhaust can flow through the 
turbocharger without causing an over-speed condition. The advantage 
that VGTs offer compared to a waste-gated turbocharger is that all 
exhaust flow is directed through the turbocharger under all operating 
conditions. This can be viewed as maximizing the use of the available 
exhaust energy.
b. Turbo Boost Control System Monitoring Requirements
    We are proposing that the OBD system monitor the boost pressure 
control system on engines so equipped for under and over boost 
malfunctions. For engines equipped with variable geometry turbochargers 
(VGT), the OBD system would have to monitor the VGT system for slow 
response malfunctions. For engines equipped with charge air cooler 
systems, the OBD system would have to monitor the charge air cooler 
system for cooling system performance malfunctions. The individual 
electronic components (e.g., actuators, valves, sensors) that are used 
in the boost pressure control system shall be monitored in accordance 
with the comprehensive component requirements in section II.D.4.
i. Turbo Underboost Malfunctions
    We are proposing that the OBD system detect a malfunction of the 
boost pressure control system prior to a decrease from the 
manufacturer's commanded boost pressure that would cause an engine's 
emissions to exceed the emissions thresholds for ``other monitors'' as 
shown in Table II.B-1. For engines in which no failure or deterioration 
of the boost pressure control system that causes a decrease in boost 
could result in an engine's emissions exceeding the applicable 
emissions thresholds, the OBD system must detect a malfunction when the 
system has reached its control limits such that it cannot increase 
boost to achieve the commanded boost pressure.
ii. Turbo Overboost Malfunctions
    We are proposing that the OBD system detect a malfunction of the 
boost pressure control system prior to an increase from the 
manufacturer's commanded boost pressure that would cause an engine's 
emissions to exceed the emissions thresholds for ``other monitors'' as 
shown in Table II.B-1. For engines in which no failure or deterioration 
of the boost pressure control system that causes an increase in boost 
could result in an engine's emissions exceeding the applicable 
emissions thresholds, the OBD system must detect a malfunction when the 
system has reached its control limits such that it cannot decrease 
boost to achieve the commanded boost pressure.
iii. VGT Slow Response Malfunctions
    We are proposing that the OBD system detect a malfunction prior to 
any failure or deterioration in the capability of the VGT system to 
achieve the commanded turbocharger geometry within a manufacturer-
specified time that would cause an engine's emissions to exceed the 
emissions thresholds for ``other monitors'' as shown in Table II.B-1. 
For engines in which no failure or deterioration of the VGT system 
response could result in an engine's emissions exceeding the applicable 
emissions thresholds, the OBD system must detect a malfunction of the 
VGT system when proper functional response

[[Page 3217]]

of the system to computer commands does not occur.
iv. Turbo Boost Feedback Control Malfunctions
    We are proposing that, for engines equipped with feedback control 
of the boost pressure system--e.g., control of VGT position, turbine 
speed, manifold pressure--the OBD system shall detect a malfunction 
when and if:
     The system fails to begin feedback control within a 
manufacturer specified time interval;
     A failure or deterioration causes open loop or default 
operation; or
     Feedback control has used up all of the adjustment allowed 
by the manufacturer.
v. Charge Air Undercooling Malfunctions
    We are proposing that the OBD system detect a malfunction of the 
charge air cooling system prior to a decrease from the manufacturer's 
specified cooling rate that would cause an engine's emissions to exceed 
the emissions thresholds for ``other monitors'' as shown in Table II.B-
1. For engines in which no failure or deterioration of the charge air 
cooling system that causes a decrease in cooling performance could 
result in an engine's emissions exceeding the applicable emissions 
thresholds, the OBD system must detect a malfunction when the system 
has no detectable amount of charge air cooling.
c. Turbo Boost Control System Monitoring Conditions
    We are proposing that the OBD system monitor continuously for 
underboost and overboost malfunctions and for boost feedback control 
malfunctions. Manufacturers would be required to define the monitoring 
conditions for VGT slow response malfunctions such that the minimum 
performance ratio requirements discussed in section II.E would be met 
with the exception that monitoring must occur every time the monitoring 
conditions are met during the driving cycle in lieu of once per driving 
cycle as required for most monitors. For purposes of tracking and 
reporting as required in section II.E, all monitors used to detect VGT 
slow response malfunctions malfunctions must be tracked separately but 
reported as a single set of values as discussed in section II.E.\20\
---------------------------------------------------------------------------

    \20\ For specific components or systems that have multiple 
monitors that are required to be reported (e.g., exhaust gas sensor 
bank 1 may have multiple monitors for sensor response or other 
sensor characteristics), the OBD system must separately track 
numerators and denominators for each of the specific monitors and 
report only the corresponding numerator and denominator for the 
specific monitor that has the lowest numerical ratio. If two or more 
specific monitors have identical ratios, the corresponding numerator 
and denominator for the specific monitor that has the highest 
denominator shall be reported for the specific component.
---------------------------------------------------------------------------

d. Turbo Boost MIL Illumination and DTC Storage
    We are proposing the general MIL illumination and DTC storage 
requirements as discussed in section II.A.2.
5. Non-Methane Hydrocarbon (NMHC) Converting Catalyst Monitoring
a. Background
    Diesel oxidation catalysts (DOCs) have been used on some nonroad 
diesel engines since the 1960s and on some diesel trucks and buses in 
the U.S. since the early 1990s. DOCs are generally used for converting 
HC and carbon monoxide (CO) emissions to water and CO2 via 
an oxidation process. Current DOCs can also be used to convert PM 
emissions. DOCs may also be used in conjunction with other 
aftertreatment emission controls--such as NOX adsorber 
systems, selective catalytic reduction (SCR) systems, and PM filters--
to improve their performance and/or clean up certain reducing agents 
that might slip through the system (e.g., the urea used in urea SCR 
systems).
b. NMHC Converting Catalyst Monitoring Requirements
    We are proposing that the OBD system monitor the NMHC converting 
catalyst(s) for proper NMHC conversion capability. We are also 
proposing that each catalyst that converts NMHC be monitored either 
individually or in combination with others. For engines equipped with 
catalyzed diesel particulate filters (CDPFs) that convert NMHC 
emissions, the catalyst function of the CDPF must be monitored in 
accordance with the CDPF monitoring requirements in section II.B.8.
i. NMHC Converting Catalyst Conversion Efficiency
    We are proposing that the OBD system detect an NMHC catalyst 
malfunction when the catalyst conversion capability decreases to the 
point that NMHC emissions exceed the emissions thresholds for ``NMHC 
catalysts'' as shown in Table II.B-1. If no failure or deterioration of 
the catalyst NMHC conversion capability could result in an engine's 
NMHC emissions exceeding the applicable emissions thresholds, the OBD 
system would have to detect a malfunction when the catalyst has no 
detectable amount of NMHC conversion capability.
ii. Other Aftertreatment Assistance Functions
    For catalysts used to generate an exotherm to assist CDPF 
regeneration, we are proposing that the OBD system detect a malfunction 
when the catalyst is unable to generate a sufficient exotherm to 
achieve that regeneration. For catalysts used to generate a feedgas 
constituency to assist SCR systems (e.g., to increase NO2 
concentration upstream of an SCR system), the OBD system would have to 
detect a malfunction when the catalyst is unable to generate the 
necessary feedgas constituents for proper SCR system operation. For 
catalysts located downstream of a CDPF and used to convert NMHC 
emissions during a CDPF regeneration event, the OBD system would be 
required to detect a malfunction when the catalyst has no detectable 
amount of NMHC conversion capability.
c. NMHC Converting Catalyst Monitoring Conditions
    Manufacturers would be required to define the monitoring conditions 
for NMHC converting catalyst malfunctions such that the minimum 
performance ratio requirements discussed in section II.E would be met. 
For purposes of tracking and reporting as discussed in section II.E, 
all monitors used to detect NMHC converting catalyst malfunctions must 
be tracked separately but reported as a single set of values as 
discussed in section II.E.\21\
---------------------------------------------------------------------------

    \21\ For specific components or systems that have multiple 
monitors that are required to be reported (e.g., exhaust gas sensor 
bank 1 may have multiple monitors for sensor response or other 
sensor characteristics), the OBD system must separately track 
numerators and denominators for each of the specific monitors and 
report only the corresponding numerator and denominator for the 
specific monitor that has the lowest numerical ratio. If two or more 
specific monitors have identical ratios, the corresponding numerator 
and denominator for the specific monitor that has the highest 
denominator shall be reported for the specific component.
---------------------------------------------------------------------------

d. NMHC Converting Catalyst MIL Illumination and DTC Storage
    We are proposing the general requirements for MIL illumination and 
DTC storage discussed in section II.A.2. Note that the monitoring 
method for the catalyst(s) must be capable of detecting all instances, 
except diagnostic self-clearing, when a catalyst DTC has been cleared 
but the catalyst has not been replaced (e.g., catalyst over temperature 
histogram approaches are not acceptable).\22\
---------------------------------------------------------------------------

    \22\ For gasoline catalyst monitoring, manufacturers generally 
use what is called an exponentially weighted moving average (EWMA) 
approach to making decisions about the catalyst's pass/fail status. 
This approach monitors the catalyst and ``saves'' that information. 
The next time it monitors the catalyst, it saves that information 
along with the previous information, placing a higher weighting on 
the most recent information. This is done every time the OBD system 
monitors the catalyst and the EWMA saves six or seven monitoring 
events before making a decision. Importantly, once there exists six 
or seven pieces of information, every monitoring event can result in 
a decision because the EWMA is always using the previous six or 
seven events. Unfortunately, if a service technician clears the data 
with a scan tool, it is going to take six or seven monitoring events 
before the catalyst monitor can make a decision on the pass/fail 
status of the catalyst. So, we want to be sure that, in addition to 
the EWMA aspect of the catalyst monitor, there exists a way of 
determining quickly that someone has cleared the data but perhaps 
did not actually repair the catalyst. This is required to help 
prevent against DTC clearing without fixing a failed catalyst as a 
means of passing an inspection & maintenance test.

---------------------------------------------------------------------------

[[Page 3218]]

6. Selective Catalytic Reduction (SCR) and Lean NOX Catalyst 
Monitoring
a. Background
    Selective Catalytic Reduction (SCR) catalysts that use ammonia as a 
NOX reductant have been used for stationary source 
NOX control for a number of years. Frequently, urea is used 
as the source of ammonia for SCR catalysts, and such systems are 
commonly referred to as Urea SCR systems. In recent years, considerable 
effort has been invested in developing urea SCR systems that could be 
applied to heavy-duty diesel vehicles with low sulfur diesel fuel. We 
now expect that urea SCR systems will be introduced in Europe to comply 
with the EURO IV heavy-duty diesel emission standards. Such systems 
have been introduced in the past year by some heavy-duty diesel engine 
manufacturers both in Europe and in Japan.
    SCR catalyst systems require an accurate urea control system to 
inject precise amounts of reductant. An injection rate that is too low 
may result in lower NOX conversions while an injection that 
is too high may release unwanted ammonia emissions--referred to as 
ammonia slip--to the atmosphere. In general, ammonia to NOX 
ratios of around 1:1 are used to provide the highest NOX 
conversion rates with minimal ammonia slip. Therefore, injecting just 
the right amount of ammonia appropriate for the amount of 
NOX in the exhaust is very important. This can be 
challenging in a highway application because on-road diesel engines 
operate over a variety of speeds and loads. This makes the use of 
closed-loop feedback systems for reductant metering very attractive. 
This can be achieved, for example, with a dedicated NOX 
sensor in the exhaust so that the NOX concentration can be 
accurately known. With an accurate fast response NOX sensor, 
closed-loop control of the ammonia injection can be used to achieve and 
maintain the desired ammonia/NOX ratios in the SCR catalyst 
for the high NOX conversion efficiencies necessary to 
achieve the 2010 emission standards under various engine operating 
conditions.
    Some have estimated that achieving the 2010 NOX emission 
standards with SCR systems will require NOX sensors that can 
measure NOX levels accurately in the 20 to 40 ppm range with 
little cross sensitivity to ammonia. Some in industry have even stated 
a desire for accuracy in the two to three ppm range. Suppliers have 
been developing NOX sensors capable of measuring 
NOX in the 0 to 100 ppm range with +/-5 ppm accuracy which 
we believe will be available by 2010.\23\ Regarding cross-sensitivity 
to ammonia, work has been done that indicates ammonia and 
NOX measurements can be independently measured by 
conditioning the output signal.\24\ This signal conditioning method 
resulted in a linear output for both ammonia and NOX from 
the NOX sensor downstream of the catalyst.
---------------------------------------------------------------------------

    \23\ Draft Technical Support Document, HDOBD NPRM, EPA420-D-06-
006, Docket ID EPA-HQ-OAR-2005-0047-0008.
    \24\ Schaer, C. M., Onder, C. H., Geering, H. P., and Elsener, 
M., ``Control of a Urea SCR Catalytic Converter System for a Mobile 
Heavy-Duty Diesel Engine,'' SAE Paper 2003-01-0776 which may be 
obtained from Society of Automotive Engineers International, 400 
Commonwealth Dr., Warrendale, PA, 15096-0001.
---------------------------------------------------------------------------

    For SCR systems, closed-loop control of the reductant injection may 
require the use of two NOX sensors. The first NOX 
sensor would be located upstream of the catalyst and the reductant 
injection point would be used for measuring the engine-out 
NOX emissions and determining the amount of reductant 
injection needed to reduce emissions. The second NOX sensor 
located downstream of the catalyst would be used for measuring the 
amount of ammonia and NOX emissions exiting the catalyst and 
providing feedback to the reductant injection control system. If the 
downstream NOX sensor detects too much NOX 
emissions exiting the catalyst, the control system can inject higher 
quantities of reductant. Conversely, if the downstream NOX 
sensor detects too much ammonia slip exiting the catalyst, the control 
system can decrease the amount of reductant injection.
    In addition to exhaust NOX levels, another important 
parameter for achieving high NOX conversion rates with 
minimum ammonia slip is catalyst temperature. SCR catalysts have a 
defined temperature range where they are most effective. For example, 
platinum catalysts are effective between 175 and 250 degrees Celsius, 
vanadium catalysts are effective between 300 and 450 degrees Celsius, 
and zeolite catalysts are most effective between 350 and 600 degrees 
Celsius. To determine exhaust catalyst temperature for reductant 
control purposes, manufacturers are likely to use temperature sensors 
placed in the exhaust system. We project that only one temperature 
sensor positioned just downstream of the SCR system will be utilized 
for reductant injection control purposes.
    Production SCR catalyst systems may also contain auxiliary 
catalysts to improve the overall emissions control capability of the 
system. An oxidation catalyst is often positioned downstream of the SCR 
catalyst to help control ammonia slip on systems without closed-loop 
control of ammonia injection. The use of a ``guard'' catalyst could 
allow higher ammonia injection levels, thereby increasing the 
NOX conversion efficiency without releasing un-reacted 
ammonia into the exhaust. The guard catalyst can also reduce HC and CO 
emission levels and diesel odors. However, increased N2O emissions may 
occur and NOX emission levels may actually increase if too 
much ammonia is oxidized in the catalyst. Some SCR systems may also 
include an oxidation catalyst upstream of the SCR catalyst and urea 
injection point to generate NO2 for lowering the effective operating 
temperature and/or volume of the SCR catalyst. Studies have indicated 
that increasing the NO2 content in the exhaust stream can reduce the 
SCR temperature requirements by about 100 degrees Celsius.\25\ This 
``pre-oxidation'' catalyst also has the added benefit of reducing HC 
emissions.
---------------------------------------------------------------------------

    \25\ Walker, A. P., Chandler, G. R., Cooper, B. J., et al., ``An 
Integrated SCR and Continuously Regenerating Trap System to Meet 
Future NOX and PM Legislation,'' SAE Paper 2000-01-0188 
which may be obtained from Society of Automotive Engineers 
International, 400 Commonwealth Dr., Warrendale, PA, 15096-0001.
---------------------------------------------------------------------------

b. SCR and Lean NOX Catalyst Monitoring Requirements
    We are proposing that the OBD system monitor SCR catalysts and lean 
NOX catalysts for proper conversion capability. We are also 
proposing that each catalyst that converts NOX be monitored 
either individually or in combination with others. For engines equipped 
with SCR systems or other catalyst systems that utilize an active/
intrusive reductant injection (e.g., active lean NOX 
catalysts utilizing diesel fuel

[[Page 3219]]

injection), the OBD system would be required to monitor the active/
intrusive reductant injection system for proper performance. The 
individual electronic components (e.g., actuators, valves, sensors, 
heaters, pumps) in the active/intrusive reductant injection system must 
be monitored in accordance with the comprehensive component 
requirements in section II.D.4.
i. Catalyst Conversion Efficiency Malfunctions
    We are proposing that the OBD system detect a catalyst malfunction 
when the catalyst conversion capability decreases to the point that 
would cause an engine's NOX emissions to exceed any of the 
applicable emissions thresholds for ``NOX Catalyst Systems'' 
as shown in Table II.B-1. If no failure or deterioration of the 
catalyst NOX conversion capability could result in an 
engine's NOX emissions exceeding any of the applicable 
emissions thresholds, the OBD system would have to detect a malfunction 
when the catalyst has no detectable amount of NOX conversion 
capability.
ii. Active/Intrusive Reductant Injection System Malfunctions
    Specific to SCR and other active/intrusive reductant injection 
system performance, we are proposing that the OBD system detect a 
malfunction prior to any failure or deterioration of the system to 
regulate reductant delivery properly (e.g., urea injection, separate 
injector fuel injection, post injection of fuel, air assisted 
injection/mixing) that would cause an engine's NOX emissions 
to exceed any of the applicable emissions thresholds for 
``NOX Catalyst Systems'' as shown in Table II.B-1. As above, 
if no failure or deterioration of the reductant delivery system could 
result in an engine's NOX emissions exceeding the applicable 
emissions thresholds, the OBD system would have to detect a malfunction 
when the system has reached its control limits such that it is no 
longer able to deliver the desired quantity of reductant.
    If the system uses a reductant other than the fuel used for the 
engine or uses a reservoir/tank for the reductant that is separate from 
the fuel tank used for the engine, the OBD system must detect a 
malfunction when there is no longer sufficient reductant available 
(e.g., the reductant tank is empty). If the system uses a reservoir/
tank for the reductant that is separate from the fuel tank used for the 
engine, the OBD system must detect a malfunction when an improper 
reductant is used in the reductant reservoir/tank (e.g., the reductant 
tank is filled with something other than the proper reductant).
iii. SCR and Lean NOX Catalyst Feedback Control System 
Malfunctions
    If the engine is equipped with feedback control of the reductant 
injection, we are proposing that the OBD system detect a malfunction 
when and if:
     The system fails to begin feedback control within a 
manufacturer specified time interval;
     A failure or deterioration causes open loop or default 
operation; or
     Feedback control has used up all of the adjustment allowed 
by the manufacturer.
c. SCR and Lean NOX Catalyst Monitoring Conditions
    Manufacturers would be required to define the monitoring conditions 
for catalyst conversion efficiency malfunctions such that the minimum 
performance ratio requirements discussed in section II.E would be met. 
For purposes of tracking and reporting as required in section II.E, all 
monitors used to detect catalyst conversion efficiency malfunctions 
must be tracked separately but reported as a single set of values as 
specified in section II.E.\26\ We are also proposing that the OBD 
system monitor continuously for active/intrusive reductant injection 
system malfunctions. Manufacturers would be required to monitor 
continuously the active/intrusive reductant delivery system.
---------------------------------------------------------------------------

    \26\ For specific components or systems that have multiple 
monitors that are required to be reported (e.g., exhaust gas sensor 
bank 1 may have multiple monitors for sensor response or other 
sensor characteristics), the OBD system must separately track 
numerators and denominators for each of the specific monitors and 
report only the corresponding numerator and denominator for the 
specific monitor that has the lowest numerical ratio. If two or more 
specific monitors have identical ratios, the corresponding numerator 
and denominator for the specific monitor that has the highest 
denominator shall be reported for the specific component.
---------------------------------------------------------------------------

d. SCR and Lean NOX Catalyst MIL Illumination and DTC 
Storage
    We are proposing the general MIL illumination and DTC storage 
requirements presented in section II.A.2 with the exception of active/
intrusive reductant injection related malfunctions. If the OBD system 
is capable of discerning that a system malfunction is being caused by 
an empty reductant tank, the manufacturer may delay illumination of the 
MIL if the vehicle is equipped with an alternative indicator for 
notifying the vehicle operator of the malfunction. The manufacturer 
would be required to demonstrate that: The alternative indicator is of 
sufficient illumination and location to be readily visible to the 
operator under all lighting conditions; and the alternative indicator 
provides equivalent assurance that a vehicle operator will be promptly 
notified; and, that corrective action would be undertaken. If the 
vehicle is not equipped with an alternative indicator and the MIL 
illuminates, the MIL may be immediately extinguished and the 
corresponding DTC erased once the OBD system has verified that the 
reductant tank has been properly refilled and the MIL has not been 
illuminated for any other type of malfunction. The Administrator may 
approve other strategies that provide equivalent assurance that a 
vehicle operator will be promptly notified and that corrective action 
will be undertaken.
    The monitoring method for the catalyst(s) would have to be capable 
of detecting all instances, except diagnostic self-clearing, when a 
catalyst DTC has been cleared but the catalyst has not been replaced 
(e.g., catalyst over temperature histogram approaches are not 
acceptable).
7. NOX Adsorber System Monitoring
a. Background
    NOX adsorbers, or lean NOX traps (LNT), work 
to control NOX emissions by storing NOX on the 
surface of the catalyst during the lean engine operation typical of 
diesel engines and then by undergoing subsequent brief rich 
regeneration events where the NOX is released and reduced 
across a precious metal catalyst.
    NOX adsorber systems generally consist of a conventional 
three-way catalyst function (e.g., platinum) with NOX 
storage components (i.e., adsorbents) incorporated into the washcoat. 
Three-way catalysts convert NOX emissions as well as HC and 
CO emissions (hence the name three-way) by promoting oxidation of HC 
and CO to H2O and CO2 using the oxidation 
potential of the NOX pollutant and, in the process, reducing 
the NOX emissions to nitrogen, N2. Said another 
way, three-way catalysts work with exhaust conditions where the net 
oxidizing and reducing chemistry of the exhaust is approximately equal, 
allowing the catalyst to promote complete oxidation/reduction reactions 
to the desired exhaust components of CO2, H2O, 
and N2. The oxidizing potential in the exhaust comes from 
NOX emissions and any feedgas oxygen (O2) not 
consumed during combustion. The reducing potential in the exhaust

[[Page 3220]]

comes from HC and CO emissions, which represent products of incomplete 
combustion. Operation of the engine to ensure that the oxidizing and 
reducing potential of the combustion and exhaust conditions is 
precisely balanced is referred to as stoichiometric engine operation.
    Because diesel engines run lean of stoichiometric operation, the 
NOX emissions are stored, or absorbed--via chemical reaction 
with alkaline earth metals such as barium nitrate in the washcoat--and 
then released during rich operation for conversion to N2. 
This NOX release during rich operation is referred to as a 
regeneration event. The rich operating conditions required for 
NOX regeneration, which generally last for several seconds, 
are typically achieved using a combination of intake air throttling (to 
reduce the amount of intake air), exhaust gas recirculation, and post-
combustion fuel injection.
    NOX adsorber systems have demonstrated NOX 
reduction efficiencies from 50 percent to in excess of 90 percent. This 
efficiency has been found to be highly dependent on the fuel sulfur 
content because NOX adsorbers are extremely sensitive to 
sulfur. The NOX adsorption material has an even greater 
affinity for sulfur compounds than NOX. Thus, sulfur 
compounds can saturate the adsorber and limit the number of active 
sites for NOX adsorption, thereby lowering the 
NOX reduction efficiency. Accordingly, low sulfur fuel is 
required to achieve the greatest NOX reduction efficiencies. 
Although new adsorber washcoat materials are being developed with a 
higher resistance to sulfur poisoning and ultra-low sulfur fuel will be 
the norm by 2010, NOX adsorber systems will still need to 
purge the stored sulfur from the storage bed by a process referred to 
as desulfation. Because the desulfation process takes longer (e.g., 
several minutes) and requires more fuel and heat than the 
NOX regeneration step, permanent thermal degradation of the 
NOX adsorber and fuel economy penalties may result from 
desulfation events happening with excessive frequency. However, if 
desulfation is not done frequently enough, NOX storage 
capacity would be compromised and fuel economy penalties would be 
incurred from excessive attempts at NOX regeneration.
    In order to achieve and maintain high NOX conversion 
efficiencies while limiting negative impacts on fuel economy and 
driveability, vehicles with NOX adsorber systems will 
require precise air/fuel control in the engine and in the exhaust 
stream. Diesel manufacturers are expected to utilize NOX 
sensors and temperature sensors to provide the most precise closed-loop 
control for the NOX adsorber system. If NOX 
sensors are not used to control the NOX adsorber system, 
manufacturers could use wide-range air-fuel (A/F) sensors located 
upstream and downstream of the adsorber as a substitute. However, A/F 
sensors cannot provide an instantaneous indication of tailpipe 
NOX levels, which would allow the control system to 
precisely determine when the adsorber system is filled to capacity and 
regeneration should be initiated. If A/F sensors are used in lieu of 
NOX sensors, an estimation of engine-out NOX 
emissions and their subsequent storage in the NOX adsorber 
can be achieved indirectly through modeling.
b. NOX Adsorber System Monitoring Requirements
    We are proposing that the OBD system monitor the NOX 
adsorber on engines so equipped for proper performance. For engines 
equipped with active/intrusive injection (e.g., in-exhaust fuel and/or 
air injection) to achieve NOX regeneration, the OBD system 
would have to monitor the active/intrusive injection system for proper 
performance. The individual electronic components (e.g., injectors, 
valves, sensors) that are used in the active/intrusive injection system 
would have to be monitored in accordance with the comprehensive 
component requirements in section II.D.4.
i. NOX Adsorber Capability Malfunctions
    We are proposing that the OBD system detect a NOX 
adsorber malfunction when its capability--i.e., its combined adsorption 
and conversion capability--decreases to the point that would cause an 
engine's NOX emissions to exceed the applicable emissions 
thresholds for ``NOX Catalyst Systems'' as shown in Table 
II.B-1. If no failure or deterioration of the NOX adsorber 
capability could result in an engine's NOX emissions 
exceeding the applicable emissions thresholds, the OBD system would 
have to detect a malfunction when the system has no detectable amount 
of NOX adsorber capability.
ii. Active/Intrusive Reductant Injection System Malfunctions
    For NOX adsorber systems that use active/intrusive 
injection (e.g., in-cylinder post fuel injection, in-exhaust air-
assisted fuel injection) to achieve desorption of the NOX 
adsorber, the OBD system would have to detect a malfunction if any 
failure or deterioration of the injection system's ability to properly 
regulate injection causes the system to be unable to achieve desorption 
of the NOX adsorber.
iii. NOX Adsorber Feedback Control System Malfunctions
    If the engine is equipped with feedback control of the reductant 
injection (e.g., feedback control of injection quantity, time), we are 
proposing that the OBD system detect a malfunction when and if:
     The system fails to begin feedback control within a 
manufacturer specified time interval;
     A failure or deterioration causes open loop or default 
operation; or
     Feedback control has used up all of the adjustment allowed 
by the manufacturer.
c. NOX Adsorber System Monitoring Conditions
    We are proposing that manufacturers define the monitoring 
conditions for NOX adsorber capability malfunctions such 
that the minimum performance ratio requirements discussed in section 
II.E would be met. For purposes of tracking and reporting as required 
in section II.E, all monitors used to detect NOX adsorber 
capability malfunctions must be tracked separately but reported as a 
single set of values as specified in section II.E.\27\ We are also 
proposing that the OBD system monitor continuously for active/intrusive 
reductant injection and feedback control system malfunctions.
---------------------------------------------------------------------------

    \27\ For specific components or systems that have multiple 
monitors that are required to be reported (e.g., exhaust gas sensor 
bank 1 may have multiple monitors for sensor response or other 
sensor characteristics), the OBD system must separately track 
numerators and denominators for each of the specific monitors and 
report only the corresponding numerator and denominator for the 
specific monitor that has the lowest numerical ratio. If two or more 
specific monitors have identical ratios, the corresponding numerator 
and denominator for the specific monitor that has the highest 
denominator shall be reported for the specific component.
---------------------------------------------------------------------------

d. NOX Adsorber System MIL Illumination and DTC Storage
    We are proposing the general requirements for MIL illumination and 
DTC storage discussed in section II.A.2.
8. Diesel Particulate Filter (DPF) System Monitoring
a. Background
    Diesel particulate filters control diesel PM by capturing the soot 
(solid carbon) portion of PM in a filter media, typically a ceramic 
wall flow substrate, and then by oxidizing (burning) it in the oxygen-
rich atmosphere of diesel exhaust.\28\ In aggregate over a driving 
cycle, the PM must be burned at a rate equal to or

[[Page 3221]]

greater than its accumulation rate, or the DPF will clog. Given low 
sulfur diesel fuel (diesel fuel with a sulfur content of 15 ppm or 
lower), highly active catalytic metals (e.g., platinum) can be used to 
promote soot oxidation. This method of PM filter regeneration, called 
passive regeneration, is the primary means of soot oxidation that we 
project industry will use in 2007/2010.
---------------------------------------------------------------------------

    \28\ See ``Regulatory Impact Analysis: Heavy-Duty Engine and 
Vehicle Standards and Highway Diesel Fuel Sulfur Control 
Requirements;'' EPA420-R-00-026; December 2000 at Chapter III for a 
more complete description of DPFs.
---------------------------------------------------------------------------

    The DPF technology has proven itself in tens of thousands of 
retrofit applications where low sulfur diesel fuel is already 
available. More than a million light-duty passenger cars in Europe now 
have diesel particulate filters. DPFs are considered the most effective 
control technology for the reduction of particulate emissions and can 
typically achieve PM reductions in excess of 90 percent.
    In order to maintain the performance of the DPF and the engine, the 
trapped PM must be periodically removed before too much particulate is 
accumulated and exhaust backpressure reaches unacceptable levels. The 
process of periodically removing accumulated PM from the DPF is known 
as ``regeneration'' and is very important for maintaining low PM 
emission levels. DPF regeneration can be passive (i.e., occur 
continuously during regular operation of the filter), active (i.e., 
occur on a controlled, periodic basis after a predetermined quantity of 
particulates have been accumulated), or a combination of the two. With 
passive regeneration, the oxidizing catalyst material on the DPF 
substrate serves to lower the temperature for oxidizing PM. This allows 
the DPF to continuously oxidize trapped PM material during normal 
driving. In contrast, active systems utilize an external heat source--
such as an electric heater or fuel burner--to facilitate DPF 
regeneration. We are projecting that virtually all DPF systems will 
have some sort of active regeneration mechanism as a backup mechanism 
should operating conditions not be conducive for passive regeneration.
    One of the key considerations for a DPF regeneration control system 
is the amount of soot quantity that is stored in the DPF (often called 
soot loading). If too much soot is stored when regeneration is 
activated, the soot can burn uncontrollably and DPF substrate could be 
damaged via melting or cracking. Conversely, activating regeneration 
when there is too little trapped soot will not ensure good combustion 
propagation which would effectively waste the energy (fuel) used to 
initiate the regeneration. Another important consideration in the 
control system design is the fuel economy penalty involved with DPF 
regeneration. Prolonged operation with high backpressures in the 
exhaust and regenerations occurring too frequently are both detrimental 
to fuel economy and DPF durability. Therefore, DPF system designers 
will need to carefully balance the regeneration frequency with various 
conflicting factors. To optimize the trap regeneration for these design 
factors, the DPF regeneration control system is projected to 
incorporate both pressure sensors and temperature sensors to model soot 
loading and other phenomena.\29\ Through the information provided by 
these sensors, designers can optimize the DPF for high effectiveness 
and maximum durability while minimizing fuel economy and performance 
penalties.
---------------------------------------------------------------------------

    \29\ Salvat, O., Marez, P., and Belot, G., ``Passenger Car 
Serial Application of a Particulate Filter System on a Common Rail 
Direct Injection Diesel Engine,'' SAE Paper 2000-01-0473 which may 
be obtained from Society of Automotive Engineers International, 400 
Commonwealth Dr., Warrendale, PA, 15096-0001.
---------------------------------------------------------------------------

b. DPF System Monitoring Requirements
    We are proposing that the OBD system monitor the DPF on engines so-
equipped for proper performance.\30\ For engines equipped with active 
regeneration systems that utilize an active/intrusive injection (e.g., 
in-exhaust fuel injection, in-exhaust fuel/air burner), the OBD system 
would have to monitor the active/intrusive injection system for proper 
performance. The individual electronic components (e.g., injectors, 
valves, sensors) that are used in the active/intrusive injection system 
must be monitored in accordance with the comprehensive component 
requirements in section II.D.4.
---------------------------------------------------------------------------

    \30\ Note that these requirements would also apply to a 
catalyzed diesel particulate filter (CDPF). We use the more common 
term DPF throughout this discussion.
---------------------------------------------------------------------------

i. PM Filtering Performance
    We are proposing that the OBD system detect a malfunction prior to 
a decrease in the filtering capability of the DPF (e.g., cracking, 
melting, etc.) that would cause an engine's PM emissions to exceed the 
applicable emissions thresholds for ``DPF Systems'' as shown in Table 
II.B-1. If no failure or deterioration of the PM filtering performance 
could result in an engine's PM emissions exceeding the applicable 
emissions thresholds, the OBD system would have to detect a malfunction 
when no detectable amount of PM filtering occurs.
ii. DPF Regeneration Frequency Malfunctions--Too Frequent
    We are proposing that the OBD system detect a malfunction when the 
DPF regeneration frequency increases from--i.e., occurs more often 
than--the manufacturer's specified regeneration frequency to a level 
such that it would cause an engine's NMHC emissions to exceed the 
applicable emissions threshold for ``DPF Systems'' as shown in Table 
II.B-1. If no such regeneration frequency exists that could cause NMHC 
emissions to exceed the applicable emission threshold, the OBD system 
would have to detect a malfunction when the PM filter regeneration 
frequency exceeds the manufacturer's specified design limits for 
allowable regeneration frequency.
iii. DPF Incomplete Regeneration Malfunctions
    We are proposing that the OBD system detect a regeneration 
malfunction when the DPF does not properly regenerate under 
manufacturer-defined conditions where regeneration is designed to 
occur.
iv. DPF NMHC Conversion Efficiency Malfunctions
    We are proposing that, for any DPF that serves to convert NMHC 
emissions, the OBD system must monitor the NMHC converting function of 
the DPF and detect a malfunction when the NMHC conversion capability 
decreases to the point that NMHC emissions exceed the NMHC threshold 
for ``DPF Systems'' as shown in Table II.B-1. If no failure or 
deterioration of the NMHC conversion capability could result in NMHC 
emissions exceeding the applicable NMHC threshold, the OBD system would 
have to detect a malfunction when the system has no detectable amount 
of NMHC conversion capability.
v. DPF Missing Substrate Malfunctions
    We are proposing that the OBD system detect a malfunction if either 
the DPF substrate is completely destroyed, removed, or missing, or if 
the DPF assembly has been replaced with a muffler or straight pipe.
vi. DPF Active/Intrusive Injection System Malfunctions
    We are proposing that, for systems that utilize active/intrusive 
injection (e.g., in-cylinder post fuel injection, in-exhaust air-
assisted fuel injection) to achieve DPF regeneration, the OBD system 
detect a malfunction if any

[[Page 3222]]

failure or deterioration of the injection system's ability to properly 
regulate injection causes the system to be unable to achieve DPF 
regeneration.
vii. DPF Regeneration Feedback Control System Malfunctions
    We are proposing that, if the engine is equipped with feedback 
control of the DPF regeneration (e.g., feedback control of oxidation 
catalyst inlet temperature, PM filter inlet or outlet temperature, in-
cylinder or in-exhaust fuel injection), the OBD system must detect a 
malfunction when and if:
     The system fails to begin feedback control within a 
manufacturer specified time interval;
     A failure or deterioration causes open loop or default 
operation; or
     Feedback control has used up all of the adjustment allowed 
by the manufacturer.
c. DPF System Monitoring Conditions
    We are proposing that manufacturers define the monitoring 
conditions for all DPF related malfunctions such that the minimum 
performance ratio requirements discussed in section II.E would be met 
with the exception that monitoring must occur every time the monitoring 
conditions are met during the driving cycle rather than once per 
driving cycle as required for most monitors. For purposes of tracking 
and reporting as required in section II.E, all monitors used to detect 
all DPF related malfunctions would have to be tracked separately but 
reported as a single set of values as specified in section II.E.\31\
---------------------------------------------------------------------------

    \31\ For specific components or systems that have multiple 
monitors that are required to be reported (e.g., exhaust gas sensor 
bank 1 may have multiple monitors for sensor response or other 
sensor characteristics), the OBD system must separately track 
numerators and denominators for each of the specific monitors and 
report only the corresponding numerator and denominator for the 
specific monitor that has the lowest numerical ratio. If two or more 
specific monitors have identical ratios, the corresponding numerator 
and denominator for the specific monitor that has the highest 
denominator shall be reported for the specific component.
---------------------------------------------------------------------------

d. DPF System MIL Illumination and DTC Storage
    We are proposing the general requirements for MIL illumination and 
DTC storage as discussed in section II.A.2.
9. Exhaust Gas Sensor Monitoring
a. Background
    Exhaust gas sensors (e.g., oxygen sensors, wide-range air-fuel (A/
F) sensors, NOX sensors) are important to the emission 
control system of vehicles. These sensors are used for enhancing the 
performance of several emission control technologies (e.g., catalysts, 
EGR systems). We expect that both oxygen sensors and wide range A/F 
sensors may be used by heavy-duty manufacturers to optimize their 
emission control technologies. We would expect that, in addition to 
their emissions control functions, these sensors will also be used to 
satisfy many of the proposed HDOBD monitoring requirements, such as 
fuel system monitoring, catalyst monitoring, and EGR system monitoring. 
NOX sensors may also be used for optimization of several 
diesel emission control technologies, such as NOX adsorbers 
and selective catalytic reduction (SCR) systems. Since an exhaust gas 
sensor can be a critical component of a vehicle's fuel and emission 
control system, the proper performance of this component needs to be 
assured to maintain low emissions. The reliance on these sensors for 
emissions control and OBD monitoring makes it important that any 
malfunction that adversely affects the performance of any of these 
sensors be detected by the OBD system.
b. Exhaust Gas Sensor Monitoring Requirements
    We are proposing that the OBD system monitor all exhaust gas 
sensors (e.g., oxygen, air-fuel ratio, NOX) used either for 
emission control system feedback (e.g., EGR control/feedback, SCR 
control/feedback, NOX adsorber control/feedback), or as a 
monitoring device, for proper output signal, activity, response rate, 
and any other parameter that can affect emissions. For engines equipped 
with heated exhaust gas sensors, the OBD system would have to monitor 
the heater for proper performance.
i. Air/Fuel Ratio Sensor Malfunctions
    For all air/fuel ratio sensors, we are proposing the following:
     Circuit malfunctions: The OBD system must detect 
malfunctions of the sensor caused by either a lack of circuit 
continuity or out-of-range values.
     Feedback malfunctions: The OBD system must detect a 
malfunction of the sensor when a sensor failure or deterioration causes 
an emissions control system--e.g., the EGR, SCR, or NOX 
adsorber systems--to stop using that sensor as a feedback input (e.g., 
causes default or open-loop operation).
     Monitoring capability: To the extent feasible, the OBD 
system must detect a malfunction of the sensor when the sensor output 
voltage, resistance, impedance, current, amplitude, activity, offset, 
or other characteristics are no longer sufficient for use as an OBD 
system monitoring device (e.g., for catalyst, EGR, SCR, or 
NOX adsorber monitoring).
    Specifically for sensors located upstream of an aftertreatment 
device, we are proposing the following:
     Sensor performance malfunctions: The OBD system must 
detect a malfunction prior to any failure or deterioration of the 
sensor voltage, resistance, impedance, current, response rate, 
amplitude, offset, or other characteristic(s) that would cause an 
engine's emissions to exceed the applicable emissions thresholds for 
``Other Monitors'' as shown in Table II.B-1.
    Specifically for sensors located downstream of an aftertreatment 
device, we are proposing the following:
     Sensor performance malfunctions: The OBD system must 
detect a malfunction prior to any failure or deterioration of the 
sensor voltage, resistance, impedance, current, response rate, 
amplitude, offset, or other characteristic(s) that would cause an 
engine's emissions to exceed the applicable emissions thresholds for 
``Air-fuel ratio sensors downstream of aftertreatment devices'' as 
shown in Table II.B-1.
ii. NOX Sensor Malfunctions
    For NOX sensors, we are proposing the following:
     Sensor performance malfunctions: The OBD system must 
detect a malfunction prior to any failure or deterioration of the 
sensor voltage, resistance, impedance, current, response rate, 
amplitude, offset, or other characteristic(s) that would cause an 
engine's emissions to exceed the applicable emissions thresholds for 
``NOX sensors'' as shown in Table II.B-1.
     Circuit malfunctions: The OBD system must detect 
malfunctions of the sensor caused by either a lack of circuit 
continuity or out-of-range values.
     Feedback malfunctions: The OBD system shall detect a 
malfunction of the sensor when a sensor failure or deterioration causes 
an emission control--e.g., the EGR, SCR, or NOX adsorber 
systems--to stop using that sensor as a feedback input (e.g., causes 
default or open-loop operation).
     Monitoring capability: To the extent feasible, the OBD 
system must detect a malfunction of the sensor when the sensor output 
voltage, resistance, impedance, current, amplitude, activity, offset, 
or other characteristics are no longer sufficient for use as an OBD 
system monitoring device (e.g., for catalyst, EGR, SCR, or 
NOX adsorber monitoring).

[[Page 3223]]

iii. Other Exhaust Gas Sensor Malfunctions
    For other exhaust gas sensors, we are proposing that the 
manufacturer submit a monitoring plan to the Administrator for 
approval. The Administrator would approve the request upon determining 
that the manufacturer has submitted data and an engineering evaluation 
that demonstrate that the monitoring plan is as reliable and effective 
as the monitoring plan required for air/fuel ratio sensors and 
NOX sensors.
iv. Exhaust Gas Sensor Heater Malfunctions
    We are proposing that the OBD system detect a malfunction of the 
heater performance when the current or voltage drop in the heater 
circuit is no longer within the manufacturer's specified limits for 
normal operation (i.e., within the criteria required to be met by the 
component vendor for heater circuit performance at high mileage). The 
manufacturer may use other malfunction criteria for heater performance 
malfunctions. To do so, the manufacturer would be required to submit 
data and/or engineering analyses that demonstrate that the monitoring 
reliability and timeliness would be equivalent to the criteria stated 
here. Further, the OBD system would be required to detect malfunctions 
of the heater circuit including open or short circuits that conflict 
with the commanded state of the heater (e.g., shorted to 12 Volts when 
commanded to 0 Volts (ground)).
c. Exhaust Gas Sensor Monitoring Conditions
    For exhaust gas sensor performance malfunctions, we are proposing 
that manufacturers define the monitoring conditions such that the 
minimum performance ratio requirements discussed in section II.E would 
be met. For purposes of tracking and reporting as required in section 
II.E, all monitors used to detect sensor performance malfunctions would 
have to be tracked separately but reported as a single set of values as 
specified in section II.E.\32\
---------------------------------------------------------------------------

    \32\ For specific components or systems that have multiple 
monitors that are required to be reported (e.g., exhaust gas sensor 
bank 1 may have multiple monitors for sensor response or other 
sensor characteristics), the OBD system must separately track 
numerators and denominators for each of the specific monitors and 
report only the corresponding numerator and denominator for the 
specific monitor that has the lowest numerical ratio. If two or more 
specific monitors have identical ratios, the corresponding numerator 
and denominator for the specific monitor that has the highest 
denominator shall be reported for the specific component.
---------------------------------------------------------------------------

    For exhaust gas sensor monitoring capability malfunctions, 
manufacturers would have to define the monitoring conditions such that 
the minimum performance ratio requirements discussed in section II.E 
would be met with the exception that monitoring must occur every time 
the monitoring conditions are met during the driving cycle rather than 
once per driving cycle as required for most monitors.
    For exhaust gas sensor circuit malfunctions and feedback 
malfunctions, monitoring must be conducted continuously.
    The manufacturer may disable continuous exhaust gas sensor 
monitoring when an exhaust gas sensor malfunction cannot be 
distinguished from other effects (e.g., disable ``out-of-range low'' 
monitoring during fuel cut conditions). To do so, the manufacturer 
would be required to submit test data and/or engineering analyses that 
demonstrate that a properly functioning sensor cannot be distinguished 
from a malfunctioning sensor and that the disablement interval is 
limited only to that necessary for avoiding a false detection.
    For exhaust gas sensor heater malfunctions, manufacturers must 
define monitoring conditions such that the minimum performance ratio 
requirements discussed in section II.E would be met. Monitoring for 
sensor heater circuit malfunctions must be conducted continuously.
d. Exhaust Gas Sensor MIL Illumination and DTC Storage
    We are proposing the general requirements for MIL illumination and 
DTC storage as discussed in section II.A.2.

C. Monitoring Requirements and Timelines for Gasoline/Spark-Ignition 
Engines

    Table II.C-1 summarizes the proposed gasoline fueled spark ignition 
emissions thresholds at which point a component or system has failed to 
the point of requiring an illuminated MIL and a stored DTC. Table II.C-
2 summarizes the proposed implementation schedule for these 
thresholds--i.e., the proposed certification requirements and in-use 
liabilities. More detail regarding the specific monitoring 
requirements, implementation schedules, and liabilities can be found in 
the sections that follow.

         Table II.C-1.--Proposed Emissions Thresholds for Gasoline Fueled SI Engines Over 14,000 Pounds
----------------------------------------------------------------------------------------------------------------
        Component/Monitor                  MY                 NMHC                 CO                 NOX
----------------------------------------------------------------------------------------------------------------
Catalytic converter system.......  2010+.............  1.75x.............  ..................  1.75x
``Other monitors'' with emissions  2010+.............  1.5x..............  1.5x..............  1.5x
 thresholds (see section II.C).
Evaporative emissions control      2010+.............  0.150 inch leak ..
 system.
----------------------------------------------------------------------------------------------------------------
Notes: MY=Model Year; 1.75x means a multiple of 1.75 times the applicable emissions standard; not all proposed
  monitors have emissions thresholds but instead rely on functionality and rationality checks as described in
  section II.D.4. The evaporative emissions control system threshold is not, technically, an emissions threshold
  but rather a leak size that must be detected; nonetheless, for ease we refer to this as the threshold.

    There are exceptions to the emissions thresholds shown in Table 
II.C-1 whereby a manufacturer can demonstrate that emissions do not 
exceed the threshold even when the component or system is non-
functional at which point a functional check would be allowed.
    The monitoring requirements described below for gasoline engines 
mirror those that are already in place for gasoline engines used in 
vehicles under 14,000 pounds. The HD gasoline industry--General Motors 
and Ford, as of today \33\--have told us that their preference is to 
use essentially the same OBD system on their engines used in both under 
and over 14,000 pound vehicles.\34\ In general, we agree with the

[[Page 3224]]

HD gasoline industry on this issue for three reasons:
---------------------------------------------------------------------------

    \33\ This is true according to our certification database for 
both he 2004 and 2005 model years. Other manufacturers certify 
engines that use the Otto cycle, but those engines do not burn 
gasoline and instead burn various alternative fuels.
    \34\ ``EMA Comments on Proposed HDOBD Requirements for HDGE,'' 
bullet items 3 and 4; April 28, 2005, Docket ID EPA-HQ-OAR-
2005-0047-0003.
---------------------------------------------------------------------------

     The engines used in vehicles above and below 14,000 pounds 
are the same which makes it easy for industry to use the same OBD 
monitors.
     The existing OBD requirements for engines used in vehicles 
below 14,000 pounds have proven effective; and,
     The industry members have more than 10 years experience 
complying with the OBD requirements for engines used in vehicles below 
14,000 pounds.
    As a result, we are proposing requirements that should allow for 
OBD system consistency in vehicles under and over 14,000 pounds rather 
than proposing requirements that mirror the proposed HD diesel 
requirements discussed in section II.B. Nonetheless, the requirements 
proposed below are for engine-based OBD monitors only rather than 
monitors for the entire powertrain (which would include the 
transmission). We are doing this for the same reasons as done for the 
proposed diesel OBD requirements in that certification of gasoline 
applications over 14,000 pounds, like their diesel counterparts, is 
done on an engine basis and not a vehicle basis.
1. Fuel System Monitoring
a. Background
    As with diesel engines, the fuel system of a gasoline engine is an 
essential component of the engine's emissions control system. Proper 
delivery of fuel is essential to maintain stoichiometric operation and 
minimize engine out emissions. Proper stoichiometric control is also 
critical to maximize catalyst conversion efficiency and reach low 
tailpipe emission levels. As such, thorough monitoring of the fuel 
system is an essential element in an OBD system.
    For gasoline engines, the fuel system generally includes a fuel 
pump, fuel pressure regulator, fuel rail, individual injectors for each 
cylinder, and a closed-loop feedback control system using oxygen 
sensor(s) or air-fuel ratio (A/F) sensor(s). The feedback sensors are 
located in the exhaust system and are used to regulate the fuel 
injection quantity to achieve a stoichiometric mixture in the exhaust. 
If the sensor indicates a rich (or lean) mixture, the system reduces 
(or increases) the amount of fuel being injected by applying a short 
term correction to the fuel injection quantity calculated for the 
current engine operating condition. To account for aging or 
deterioration in the system such as reduced injector flow, more 
permanent long term corrections are also learned and applied to the 
fuel injection quantity for more precise fueling.
    For gasoline engines, fuel system monitoring has been implemented 
on light-duty vehicles since the 1996 model year and on heavy-duty 
vehicles less than 14,000 pounds and the engines used in those vehicles 
since the 2004/2005 model year. For heavy-duty gasoline engines used in 
vehicles over 14,000 pounds (many of which are the same engine as is 
used in vehicles less than 14,000 pounds), the system components and 
control strategies are identical to those used in the light-duty and 
under 14,000 pound categories. As such, the monitoring requirements 
established for engines used in vehicles less than 14,000 pounds can be 
directly applied to engines used in vehicles over 14,000 pounds.
b. Fuel System Monitoring Requirements
    We are proposing that the fuel system be continuously monitored for 
its ability to maintain engine emissions below the applicable emissions 
thresholds. Manufacturers would also be required to verify that the 
fuel system is in closed-loop operation--e.g., that it is using the 
oxygen sensor for feedback control. The individual components of the 
fuel system would also be covered by separate monitoring requirements 
for oxygen sensors, misfire (for the fuel injectors), and comprehensive 
components (in systems such as those with electronically-controlled 
variable speed fuel pumps or electronically-controlled fuel pressure 
regulators).
i. Fuel System Performance
    We are proposing that the OBD system be required to detect a 
malfunction of the fuel delivery system (including feedback control 
based on a secondary oxygen sensor) when the fuel delivery system is 
unable to maintain the engine's emissions at or below the applicable 
emissions thresholds for ``Other monitors'' as shown in Table II.C-1.
ii. Fuel System Feedback Control
    If the engine is equipped with adaptive feedback control, we are 
proposing that the OBD system be required to detect a malfunction when 
the adaptive feedback control has used up all of the adjustment allowed 
by the manufacturer. However, if the engine is equipped with feedback 
control that is based on a secondary oxygen (or equivalent) sensor, the 
OBD system would not be required to detect a malfunction of the fuel 
system solely when the feedback control based on that secondary oxygen 
sensor has used up all of the adjustment allowed by the manufacturer. 
For such systems, the OBD system would be required to meet the fuel 
system performance requirements presented above.
    Additionally, we are proposing that the OBD system be required to 
detect a malfunction whenever the fuel control system fails to enter 
closed loop operation within a time interval after engine startup. The 
manufacturer would be required to submit data and/or engineering 
analyses that support their chosen time interval.
    Lastly, manufacturers would be allowed to adjust the malfunction 
criteria and/or monitoring conditions to compensate for changes in 
altitude, temporary introduction of large amounts of purge vapor, or 
for other similar identifiable operating conditions when they occur.
c. Fuel System Monitoring Conditions
    We are proposing that the OBD system monitor continuously for 
malfunctions of the fuel system.
d. Fuel System MIL Illumination and DTC Storage
    We are proposing that a pending DTC be stored immediately upon 
detecting a malfunction according to the fuel system monitoring 
requirements presented in section II.C.1.b (i.e., rather than waiting 
until the end of the drive cycle to store the pending DTC). Once a 
pending DTC is stored, the OBD system would be required to illuminate 
the MIL immediately and store a MIL-on DTC if a malfunction is again 
detected during either of the following two events: (1) The drive cycle 
immediately following the drive cycle during which the pending DTC was 
stored, regardless of the conditions encountered during the drive 
cycle; or, (2) on the next drive cycle during which similar conditions 
are encountered to those that occurred when the pending DTC was 
stored.\35\
---------------------------------------------------------------------------

    \35\ ``Similar conditions,'' as used in conjunction with misfire 
and fuel system monitoring, means engine conditions having an engine 
speed within 375 rpm, load conditions within 20 percent, and the 
same warm up status (i.e., cold or hot) as existing during the 
applicable previous problem detection. The Administrator may approve 
other definitions of similar conditions based on comparable 
timeliness and reliability in detecting similar engine operation.
---------------------------------------------------------------------------

    We are also proposing that the pending DTC may be erased at the end 
of the next drive cycle in which similar conditions have been 
encountered without detecting a malfunction according to the fuel 
system monitoring requirements. The pending DTC may also be erased if 
similar conditions are not encountered during the 80 drive cycles 
immediately after the initial

[[Page 3225]]

detection of a malfunction for which the pending DTC was set.
    We are proposing some specific requirements with respect to storage 
of freeze frame information associated with fuel system malfunctions. 
First, the OBD system must store and erase freeze frame information 
either in conjunction with storing and erasing a pending DTC or in 
conjunction with storing and erasing a MIL-on DTC. Second, if freeze 
frame information is already stored for a malfunction other than an 
engine misfire or fuel system malfunction at the time that a fuel 
system DTC is stored, the preexisting freeze frame information must be 
replaced with freeze frame information regarding the fuel system 
malfunction.
    The OBD system would also be required to store the engine speed, 
load, and warm up status present when the first fuel system malfunction 
is detected that resulted in the storage of the pending DTC. The MIL 
may be extinguished after three sequential drive cycles in which 
similar conditions have been encountered without detecting a 
malfunction of the fuel system.
2. Engine Misfire Monitoring
a. Background
    Detecting engine misfire on a gasoline spark ignition engine is 
important for two reasons: Its impact on the emissions performance of 
the engine and its impact on the durability of the catalytic converter. 
Engine misfire has two primary causes: Lack of spark and poor fuel 
metering (delivery). When misfire occurs, unburned fuel and air are 
pumped out of the engine and into the exhaust system and into the 
catalyst. This can increase dramatically the operating temperature of 
the catalyst where temperatures can soar to above 900 degrees Celsius. 
This problem is usually most severe under high load/high speed engine 
operating conditions and can cause irreversible damage to the catalyst. 
Though the durability of catalysts has been improving, most are unable 
to sustain continuous operation at such high temperatures. Engine 
misfire also contributes to poor emissions performance, especially when 
the misfire occurs during engine warm-up and the catalyst itself has 
not yet reached its operating temperature.
b. Engine Misfire Monitoring Requirements
    We are proposing that the OBD system detect both engine misfire 
capable of causing catalyst damage and engine misfire capable of 
causing poor emissions performance. Additionally, the OBD system would 
be required to identify the specific cylinder in which misfire is 
occurring and/or if there exists a condition in which more than one 
cylinder is misfiring; when identifying a multiple cylinder misfire 
condition, the OBD system would not be required to identify 
individually each of the misfiring cylinders. We are proposing an 
exception to this whereby if more than 90 percent of the detected 
misfires are occurring in a single cylinder, the manufacturer may elect 
to consider it a single cylinder misfire condition rather than a 
multiple cylinder misfire condition. However, we are proposing that, if 
two or more cylinders individually have more than 10 percent of the 
total number of detected misfires, the manufacturer must consider it a 
multiple cylinder misfire condition.
i. Engine Misfire Capable of Causing Catalyst Damage
    We are proposing that the manufacturer be required to detect the 
percentage of misfire--evaluated in 200 revolution increments--for each 
engine speed and load condition that would result in a temperature 
capable of damaging the catalyst. For every engine speed and load 
condition at which this percentage is determined to be less than five 
percent, the manufacturer may set the malfunction criteria at five 
percent. The manufacturer may use a longer interval than a 200 
revolution increment but only for determining, on a given drive cycle, 
the first misfire exceedance; upon detecting the first such exceedance, 
the 200 revolution increment must be used. The manufacturer may use a 
longer initial interval by submitting data and/or engineering analyses 
that demonstrate that catalyst damage would not occur due to 
unacceptably high catalyst temperatures before the interval has 
elapsed.
    Further, we are proposing that, for the purpose of establishing the 
temperature at which catalyst damage would occur, manufacturers not be 
allowed to define the catalyst damaging temperature at a temperature 
more severe than what the catalyst system could be operated at for 10 
consecutive hours and still meet the applicable standards.
ii. Engine Misfire Causing Poor Emissions Performance
    We are proposing that the manufacturer be required to detect the 
percentage of misfire--evaluated in 1000 revolution increments--that 
would cause emissions to exceed the emissions thresholds for ``Other 
monitors'' as shown in Table II.C-1 if that percentage of misfire were 
present from the beginning of the test procedure. To establish this 
percentage of misfire, the manufacturer would be required to use 
misfire events occurring at equally spaced, complete engine cycle 
intervals, across randomly selected cylinders throughout each 1000 
revolution increment. If this percentage of misfire is determined to be 
lower than one percent, the manufacturer may set the malfunction 
criteria at one percent. The manufacturer may use a different interval 
than a 1000 revolution increment. To do so, the manufacturer would be 
required to submit data and/or engineering analyses demonstrating that 
the strategy would be equally effective and timely at detecting 
misfire. A malfunction must be detected if the percentage of misfire is 
exceeded regardless of the pattern of misfire events (e.g., random, 
equally spaced, continuous).
c. Engine Misfire Monitoring Conditions
    We are proposing that the OBD system monitor continuously to detect 
engine misfire under all of the following conditions:
     From no later than the end of the second crankshaft 
revolution after engine start;
     During the rise time and settling time as the engine 
reaches the desired idle speed immediately following engine start-up 
(i.e., ``flare-up'' and ``flare-down''); and,
     Under all positive torque conditions except within the 
engine operating region bound by lines connecting the following three 
points: An engine speed of 3000 rpm with the engine load at the 
positive torque line (i.e., engine load with the transmission in 
neutral), an engine speed at the redline rpm with the engine load at 
the positive torque line, and an engine speed at the redline rpm with 
an engine load at which intake manifold vacuum is four inches of 
mercury lower than that at the positive torque line (this would be an 
engine load somewhat greater than the engine load at the positive 
torque line).\36\
---------------------------------------------------------------------------

    \36\ ``Redline engine speed'' is actually defined by the 
manufacturer as either the recommended maximum engine speed as 
normally displayed on instrument panel tachometers or the engine 
speed at which fuel shutoff occurs.
---------------------------------------------------------------------------

    If a monitoring system cannot detect all misfire patterns under the 
required engine speed and load conditions, the manufacturer may request 
approval of the system nonetheless. In evaluating the manufacturer's 
request, the Administrator would consider:
     The magnitude of the region(s) in which misfire detection 
is limited;
     The degree to which misfire detection is limited in those 
region(s)

[[Page 3226]]

(i.e., the probability of detection of misfire events);
     The frequency with which said region(s) are expected to be 
encountered in-use;
     The type of misfire patterns for which misfire detection 
is troublesome; and,
     Demonstration that the monitoring technology being used is 
not inherently incapable of detecting misfire under the required 
conditions (i.e., compliance can be achieved by other manufacturers on 
their engines).
    The Administrator's evaluation would be based on the following 
misfire patterns:
     Equally spaced misfire occurring on randomly selected 
cylinders;
     Single cylinder continuous misfire; and,
     Paired cylinder (cylinders firing at the same crank angle) 
continuous misfire.
    Further, a manufacturer may use a monitoring system that has 
reduced misfire detection capability during the portion of the first 
1000 revolutions after engine start during which a cold start emission 
reduction strategy is active that reduces engine torque (e.g., spark 
retard strategies). To do so, the manufacturer would be required to 
submit data and/or engineering analyses demonstrating that the 
probability of detection is greater than or equal to 75 percent during 
the worst case condition (i.e., lowest generated torque) for a vehicle 
operated continuously at idle (park/neutral idle) on a cold start 
between 50 and 86 degrees Fahrenheit and that the technology cannot 
reliably detect a higher percentage of the misfire events during these 
conditions.
    A manufacturer may disable misfire monitoring or use an alternative 
malfunction criterion when misfire cannot be distinguished from other 
effects. To do so, the manufacturer would be required to submit data 
and/or engineering analyses demonstrating that the disablement interval 
or period of use of an alternative malfunction criterion is limited 
only to that necessary for avoiding a false detection (errors of 
commission). Such disablements would be allowed for conditions 
involving:
     Rough road;
     Fuel cut;
     Gear changes for manual transmission vehicles;
     Traction control or other vehicle stability control 
activation such as anti-lock braking or other engine torque 
modifications to enhance vehicle stability;
     Off-board control or intrusive activation of vehicle 
components or diagnostics during service or assembly plant testing;
     Portions of intrusive evaporative system or EGR 
diagnostics that can significantly affect engine stability (i.e., while 
the purge valve is open during the vacuum pull-down of a evaporative 
system leak check but not while the purge valve is closed and the 
evaporative system is sealed or while an EGR diagnostic causes the EGR 
valve to be intrusively cycled on and off during positive torque 
conditions); or,
     Engine speed, load, or torque transients due to throttle 
movements more rapid than occurs over the FTP cycle for the worst case 
engine within each engine family.
    Additionally, the manufacturer may disable misfire monitoring when 
the fuel level is 15 percent or less of the nominal capacity of the 
fuel tank, when PTO units are active, or while engine coolant 
temperature is below 20 degrees Fahrenheit. For the latter case, the 
manufacturer may continue the misfire monitoring disablement until 
engine coolant temperature exceeds 70 degrees Fahrenheit provided the 
manufacturer can demonstrate that it is necessary.
    In general, the Administrator would not approve misfire monitoring 
disablement for conditions involving normal air conditioning compressor 
cycling from on-to-off or off-to-on, automatic transmission gear shifts 
(except for shifts occurring during wide open throttle operation), 
transitions from idle to off-idle, normal engine speed or load changes 
that occur during the engine speed rise time and settling time (i.e., 
``flare-up'' and ``flare-down'') immediately after engine starting 
without any vehicle operator-induced actions (e.g., throttle stabs), or 
excess acceleration (except for acceleration rates that exceed the 
maximum acceleration rate obtainable at wide open throttle while the 
vehicle is in gear due to abnormal conditions such as slipping of a 
clutch).
    Further, the manufacturer may request approval of other misfire 
monitoring disablements or use of alternative malfunction criteria for 
any other condition. The Administrator would consider such requests on 
a case by case basis and will consider whether or not the manufacturer 
has demonstrated that the request is based on an unusual or unforeseen 
circumstance and that it is applying the best available computer and 
monitoring technology.
    For engines with more than eight cylinders that cannot meet the 
continuous monitoring and detection requirements listed above, a 
manufacturer may use alternative misfire monitoring conditions. Any 
manufacturer wishing to use alternative misfire monitoring conditions 
must submit data and/or an engineering evaluation that demonstrate that 
misfire detection throughout the required operating region cannot be 
achieved when using proven monitoring technology (i.e., a technology 
that provides for compliance with these requirements on other engines) 
and provided misfire is detected to the fullest extent permitted by the 
technology. However, the misfire detection system would still be 
required to monitor during all positive torque operating conditions 
encountered during an FTP transient cycle.
d. Engine Misfire MIL Illumination and DTC Storage
    Manufacturers may store a general misfire DTC instead of a cylinder 
specific DTC under certain operating conditions. Do so shall depend on 
the manufacturer submitting data and/or an engineering evaluation that 
demonstrate that the specific misfiring cylinder cannot be reliably 
identified when the certain operating conditions occur.
i. Engine Misfire Capable of Causing Catalyst Damage
    We are proposing that a pending DTC shall be stored immediately if, 
during a single drive cycle, the percentage of misfire determined by 
the manufacturer as being capable of causing catalyst damage is 
exceeded three times when operating in the positive torque region 
encountered during an FTP transient cycle or is exceeded on a single 
occasion when operating at any other engine speed and load condition in 
the positive torque region defined above. Immediately after a pending 
DTC is stored, the MIL shall blink once per second at all times while 
misfire is occurring during the drive cycle (i.e., the MIL may be 
extinguished during those times when misfire is not occurring during 
the drive cycle). If, at the time such a catalyst damaging engine 
misfire is occurring, the MIL is already illuminated for a malfunction 
other than engine misfire, the MIL shall blink similarly while the 
engine misfire is occurring and, if the misfire ceases, the MIL shall 
stop blinking but shall remain illuminated as commanded by the other 
malfunction.
    If a pending DTC is stored as described above, the OBD system shall 
immediately store a MIL-on DTC if the percentage of misfire determined 
by the manufacturer as being capable of causing catalyst damage is 
again exceeded one or more times during either: (a) the drive cycle 
immediately

[[Page 3227]]

following the storage of the pending DTC, regardless of the conditions 
encountered during the drive cycle; or, (b) on the next drive cycle in 
which similar conditions are encountered to those that existed when the 
pending DTC was stored.
    If, during a previous drive cycle, a pending DTC has been stored 
associated with detection of an engine misfire capable of causing poor 
emissions performance, the OBD system shall immediately store a MIL-on 
DTC if the percentage of misfire determined by the manufacturer as 
capable of causing catalyst damage is exceeded, regardless of the 
conditions encountered.
    Upon storage of a MIL-on DTC associated with engine misfire capable 
of causing catalyst damage, the MIL shall blink as described above 
while the engine misfire is occurring and then shall remain 
continuously illuminated if the engine misfire ceases. This MIL 
illumination logic shall continue until the requirements for 
extinguishing the MIL are met, as described below.
    If the engine misfire is not again detected by the end of the next 
drive cycle in which similar conditions are encountered to those that 
existed when the pending DTC was stored then the pending DTC shall be 
erased. The pending DTC may also be erased if similar conditions are 
not encountered during the 80 drive cycles subsequent to the initial 
malfunction detection.
    We are also proposing that engines with fuel shutoff and default 
fuel control--that are used to prevent catalyst damage should engine 
misfire capable of causing catalyst damage be detected--shall have some 
exemptions from these MIL illumination requirements. Most notably, the 
MIL is not required to blink while the catalyst damaging misfire is 
occurring. Instead, the MIL may simply illuminate in a steady fashion 
while the misfire is occurring provided that the fuel shutoff and 
default fuel control are activated as soon as the misfire is detected. 
Fuel shutoff and default fuel control may be deactivated only to permit 
fueling outside of the misfire range. Manufacturers may also 
periodically, but not more than once every 30 seconds, deactivate fuel 
shutoff and default fuel control to determine if the catalyst damaging 
misfire is still occurring. Normal fueling and fuel control may be 
resumed if the catalyst damaging misfire is no longer being detected.
    Manufacturers may also use a MIL illumination strategy that 
continuously illuminates the MIL in lieu of blinking the MIL during 
extreme misfire conditions capable of causing catalyst damage (i.e., 
misfire capable of causing catalyst damage that is occurring at all 
engine speeds and loads). Manufacturers would be allowed to use such a 
strategy only when catalyst damaging misfire levels cannot be avoided 
during reasonable driving conditions and the manufacturer can 
demonstrate that the strategy will encourage operation of the vehicle 
in conditions that will minimize catalyst damage (e.g., at low engine 
speeds and loads).
ii. Engine Misfire Causing Poor Emissions Performance
    We are proposing that, for a misfire detected within the first 1000 
revolutions after engine start during which misfire detection is 
active, a pending DTC shall be stored after the first exceedance of the 
percentage of misfire determined by the manufacturer as capable of 
causing poor emissions performance. If a pending DTC is stored, the OBD 
system shall illuminate the MIL and store a MIL-on DTC within 10 
seconds if an exceedance of the percentage of misfire is again detected 
in the first 1000 revolutions during any subsequent drive cycle, 
regardless of the conditions encountered during the driving cycle. The 
pending DTC shall be erased at the end of the next drive cycle in which 
similar conditions are encountered to those that existed when the 
pending DTC was stored provided the specified percentage of misfire is 
not again detected. The pending DTC may also be erased if similar 
conditions are not encountered during the 80 drive cycles subsequent to 
the initial malfunction detection.
    For a misfire detected after the first 1000 revolutions following 
engine start, a pending DTC shall be stored no later than after the 
fourth exceedance--during a single drive cycle--of the percentage of 
misfire determined by the manufacturer as being capable of causing poor 
emissions performance. If a pending DTC is stored, the OBD system shall 
illuminate the MIL and store a MIL-on DTC within 10 seconds if an 
exceedance of the percentage of misfire is again detected four times 
during: (a) the drive cycle immediately following the storage of the 
pending DTC, regardless of the conditions encountered during the drive 
cycle; or, (b) on the next drive cycle in which similar conditions are 
encountered to those that existed when the pending DTC was stored. The 
pending DTC shall be erased at the end of the next drive cycle in which 
similar conditions are encountered to those that existed when the 
pending DTC was stored provided the specified percentage of misfire is 
not again detected. The pending DTC may also be erased if similar 
conditions are not encountered during the 80 drive cycles subsequent to 
the initial malfunction detection.
    We are proposing some specific items with respect to freeze frame 
storage associated with engine misfire. The OBD system shall store and 
erase freeze frame conditions either in conjunction with storing and 
erasing a pending DTC or in conjunction with storing a MIL-on DTC and 
erasing a MIL-on DTC. In addition to those proposed requirements 
discussed in section II.A.2, we are proposing that, if freeze frame 
conditions are stored for a malfunction other than a misfire 
malfunction when a DTC is stored, the previously stored freeze frame 
information shall be replaced with freeze frame information regarding 
the misfire malfunction (i.e., the misfire's freeze frame information 
should take precedence over freeze frames for other malfunctions). 
Further, we are proposing that, upon detection of misfire, the OBD 
system store the following engine conditions: engine speed, load, and 
warm up status of the first misfire event that resulted in the storage 
of the pending DTC.
    Lastly, we are proposing that the MIL may be extinguished after 
three sequential driving cycles in which similar conditions have been 
encountered without an exceedance of the specified percentage of 
misfire.
3. Exhaust Gas Recirculation (EGR) Monitoring
a. Background
    EGR works to reduce NOX emissions the same way in 
gasoline engines as described earlier for diesel engines. First, the 
recirculated exhaust gases dilute the intake air--i.e., oxygen in the 
fresh air is displaced with relatively non-reactive exhaust gases--
which, in turn, results in less oxygen to form NOX. Second, 
EGR absorbs heat from the combustion process which reduces combustion 
chamber temperatures which, in turn, reduces NOX formation. 
The amount of heat absorbed from the combustion process is a function 
of EGR flow rate and recirculated gas temperature, both of which are 
controlled to minimize NOX emissions. EGR systems can 
involve many components to ensure accurate control of EGR flow, 
including valves, valve position sensors, and actuators.
b. EGR System Monitoring Requirements
    We are proposing that the OBD system monitor the EGR system on 
engines so equipped for low and high

[[Page 3228]]

flow rate malfunctions. The individual electronic components (e.g., 
actuators, valves, sensors) that are used in the EGR system must be 
monitored in accordance with the comprehensive component requirements 
in section II.D.4.
i. EGR Low Flow Malfunctions
    We are proposing that the OBD system detect a malfunction prior to 
a decrease from the manufacturer's specified EGR flow rate that would 
cause an engine's emissions to exceed the emissions thresholds for 
``other monitors'' as shown in Table II.C-1. For engines in which no 
failure or deterioration of the EGR system that causes a decrease in 
flow could result in an engine's emissions exceeding the applicable 
emissions thresholds, the OBD system would have to detect a malfunction 
when the system has reached its control limits such that it cannot 
increase EGR flow to achieve the commanded flow rate.
ii. EGR High Flow Malfunctions
    We are proposing that the OBD system detect a malfunction of the 
EGR system, including a leaking EGR valve--i.e., exhaust gas flowing 
through the valve when the valve is commanded closed--prior to an 
increase from the manufacturer's specified EGR flow rate that would 
cause an engine's emissions to exceed the emissions thresholds for 
``other monitors'' as shown in Table II.C-1. For engines in which no 
failure or deterioration of the EGR system that causes an increase in 
flow could result in an engine's emissions exceeding the applicable 
emissions thresholds, the OBD system would have to detect a malfunction 
when the system has reached its control limits such that it cannot 
reduce EGR flow to achieve the commanded flow rate.
c. EGR System Monitoring Conditions
    We are proposing that manufacturers define the monitoring 
conditions for EGR system malfunctions such that the minimum 
performance ratio requirements discussed in section II.E would be met. 
For purposes of tracking and reporting as required in section II.E, all 
monitors used to detect EGR low flow and high flow malfunctions must be 
tracked separately but reported as a single set of values as specified 
in section II.E.\37\
---------------------------------------------------------------------------

    \37\ For specific components or systems that have multiple 
monitors that are required to be reported (e.g., exhaust gas sensor 
bank 1 may have multiple monitors for sensor response or other 
sensor characteristics), the OBD system must separately track 
numerators and denominators for each of the specific monitors and 
report only the corresponding numerator and denominator for the 
specific monitor that has the lowest numerical ratio. If two or more 
specific monitors have identical ratios, the corresponding numerator 
and denominator for the specific monitor that has the highest 
denominator shall be reported for the specific component.
---------------------------------------------------------------------------

    Manufacturers may temporarily disable the EGR system monitor under 
conditions when monitoring may not be reliable (e.g., when freezing may 
affect performance of the system). Such temporary disablement would be 
allowed provided the manufacturer has submitted data and/or an 
engineering evaluation that demonstrate that the EGR monitor cannot be 
done reliably when these specific conditions exist.
d. EGR System MIL Illumination and DTC Storage
    We are proposing the general requirements for MIL illumination and 
DTC storage as discussed in section II.A.2.
4. Cold Start Emission Reduction Strategy Monitoring
a. Background
    The largest portion of exhaust emissions from gasoline engines is 
generated during the brief period following startup before the engine 
and catalyst have warmed up to their normal operating temperatures. To 
meet increasingly stringent emissions standards, manufacturers are 
developing hardware and associated control strategies to reduce these 
``cold start'' emissions. Most efforts center on reducing catalyst 
warm-up time.
    A cold catalyst is heated mainly by two mechanisms: heat 
transferred from the exhaust gases to the catalyst; and, heat generated 
in the catalyst as a result of the exothermic catalytic reactions. Most 
manufacturers use substantial spark retard and/or increased idle speed 
following a cold engine start, both of which maximize the heat 
available in the exhaust gases which, in turn, increases the heat 
transfer to the catalyst. Vehicle drivability and engine idle quality 
concerns tend to limit the amount of spark retard and/or increased idle 
speed that a manufacturer can use to accelerate catalyst warm up. These 
strategies or, more correctly, the systems used to employ these 
strategies--the ignition system for spark retard and the idle control 
system for control of engine speed--are normally monitored only after 
engine warm-up. Therefore, any malfunctions that might occur during the 
cold start event may not be detected by the OBD system. This could have 
significant emissions consequences due to the unknown loss of emissions 
control during the time following engine startup.
    This concern is exacerbated by the high cost of precious metals--
the platinum group metals (PGM) platinum, palladium, and rhodium--which 
motivates industry to minimize their use in catalysts. To compensate 
for the resultant reduction in overall catalyst performance, 
manufacturers will likely use increasingly more aggressive cold start 
emission reduction strategies in an attempt to further reduce cold 
start emissions. These strategies must be successful--and be properly 
monitored--to meet the more stringent 2008 emissions standards and to 
maintain low emissions in-use.
b. Cold Start Emission Reduction Strategy Monitoring Requirements
    We are proposing that, if an engine incorporates an engine control 
strategy specifically to reduce cold start emissions, the OBD system 
must monitor the key components (e.g., idle air control valve), other 
than the secondary air system, while the control strategy is active to 
ensure that the control strategy is operating properly. Secondary air 
systems would have to be monitored separately as discussed in section 
II.C.5.
    The OBD system would be required to detect a malfunction prior to 
any failure or deterioration of the individual components associated 
with the cold start emissions reduction control strategy that would 
cause an engine's emissions to exceed the emissions thresholds for 
``other monitors'' as shown in Table II.C-1. For components where no 
failure or deterioration of the component used by the cold start 
emission reduction strategy could result in an engine's emissions 
exceeding the applicable emissions thresholds, the individual 
components would have to be monitored for proper functional response as 
described in section II.D.4 while the control strategy is active.
    Manufacturers would be required to establish the appropriate 
malfunction criteria based on data from one or more representative 
engine(s). Further, manufacturers would be required to provide an 
engineering evaluation for establishing the malfunction criteria for 
the remainder of the manufacturer's product line. An annual evaluation 
of these criteria by the Administrator may not be necessary provided 
the manufacturer can demonstrate that any technological changes from 
one year to the next do not affect the previously approved malfunction 
criteria.
c. Cold Start Emission Reduction Strategy Monitoring Conditions
    We are proposing that manufacturers define the monitoring 
conditions for

[[Page 3229]]

malfunctions of the cold start emissions reduction strategy such that 
the minimum performance ratio requirements discussed in section II.E 
would be met.
d. Cold Start Emission Reduction Strategy MIL Illumination and DTC 
Storage
    We are proposing the general requirements for MIL illumination and 
DTC storage as discussed in section II.A.2.
5. Secondary Air System Monitoring
a. Background
    Secondary air systems--expected to be used on gasoline engines 
only--are used to reduce cold start emissions of hydrocarbons and 
carbon monoxide. Many of today's engines operate near stoichiometry 
after a cold engine start. However, the future more stringent emission 
standards may require the addition of a secondary air system in 
combination with a richer than stoichiometric cold start mixture. Such 
an approach could quickly warm up the catalyst for improved cold start 
emissions performance.
    Secondary air systems typically consist of an electric air pump, 
various hoses, and check valves to deliver outside air to the exhaust 
system upstream of the catalytic converter(s). This system usually 
operates only after a cold engine start and usually for only a brief 
period of time. When the electric air pump is operating, fresh air is 
delivered into the exhaust where it mixes with and ignites any unburned 
fuel. This serves to warm up the catalyst far more rapidly than would 
otherwise occur. Any problems that might occur in the field--corroded 
check valves, damaged tubing and hoses, malfunctioning air switching 
valves--could cause cold start emissions performance to suffer. 
Therefore, monitoring is needed given the importance of a properly 
functioning secondary air system to emissions performance.
b. Secondary Air System Monitoring Requirements
    We are proposing that the OBD system on engines equipped with any 
form of secondary air delivery system be required to monitor the proper 
functioning of the secondary air delivery system, including all air 
switching valve(s). The individual electronic components (e.g., 
actuators, valves, sensors) in the secondary air system would have to 
be monitored in accordance with the comprehensive component 
requirements discussed in section II.D.4.
i. Secondary Air System Low Flow Malfunctions
    We are proposing that the OBD system detect a secondary air system 
malfunction prior to a decrease from the manufacturer's specified air 
flow during normal operation that would cause an engine's emissions to 
exceed the emissions thresholds for ``other monitors'' as shown in 
Table II.C-1.\38\ For engines in which no deterioration or failure of 
the secondary air system would result in an engine's emissions 
exceeding any of the applicable emissions thresholds, the OBD system 
would have to detect a malfunction when no detectable amount of air 
flow is delivered during normal operation of the secondary air system.
---------------------------------------------------------------------------

    \38\ For purposes of secondary air system malfunctions, ``air 
flow'' is defined as the air flow delivered by the secondary air 
system to the exhaust system. For engines using secondary air 
systems with multiple air flow paths/distribution points, the air 
flow to each bank (i.e., a group of cylinders that share a common 
exhaust manifold, catalyst, and control sensor) must be monitored in 
accordance with these malfunction criteria. Also, ``normal 
operation'' is defined as the condition where the secondary air 
system is activated during catalyst and/or engine warm-up following 
engine start. ``Normal operation'' does not include the condition 
where the secondary air system is intrusively turned on solely for 
the purpose of monitoring.
---------------------------------------------------------------------------

ii. Secondary Air System High Flow Malfunctions
    We are proposing that the OBD system detect a secondary air system 
malfunction prior to an increase from the manufacturer's specified air 
flow during normal operation that would cause an engine's emissions to 
exceed the emissions thresholds for ``other monitors'' as shown in 
Table II.C-1.\39\ For engines in which no deterioration or failure of 
the secondary air system would result in an engine's emissions 
exceeding any of the applicable emissions thresholds, the OBD system 
would have to detect a malfunction when no detectable amount of air 
flow is delivered during normal operation of the secondary air system.
---------------------------------------------------------------------------

    \39\ Ibid.
---------------------------------------------------------------------------

c. Secondary Air System Monitoring Conditions
    We are proposing that manufacturers define the monitoring 
conditions for malfunctions of the secondary air system such that the 
minimum performance ratio requirements discussed in section II.E would 
be met. For purposes of tracking and reporting as required in section 
II.E, all monitors used to detect malfunctions of the secondary air 
system during its normal operation must be tracked separately but 
reported as a single set of values as specified in section II.E
d. Secondary Air System MIL Illumination and DTC Storage
    We are proposing the general requirements for MIL illumination and 
DTC storage as discussed in section II.A.2.
6. Catalytic Converter Monitoring
a. Background
    Three-way catalysts are one of the most important emission-control 
components on gasoline engines. They consist of ceramic or metal 
substrates coated with the one or more of the platinum group metals 
(PGM) platinum, palladium, and rhodium. These PGMs are dispersed within 
an alumina washcoat containing ceria, and the substrates are mounted in 
a stainless steel container in the vehicle exhaust system. Three-way 
catalysts are capable of oxidizing HC emissions, oxidizing CO 
emissions, and reducing NOX emissions, hence the term three-
way.
    While continuous improvements to catalysts have increased their 
durability, their performance still deteriorates, especially when 
subjected to very high temperatures. Such high temperatures can be 
caused by, among other factors, engine misfire which results in 
unburned fuel and air entering and igniting in the catalyst. Exposure 
to such high temperatures will result in reduced catalyst conversion 
efficiency. Catalyst efficiency can also deteriorate via poisoning if 
exposed to lead, phosphorus, or high sulfur levels. Catalysts can also 
fail by mechanical means such as excessive vibration. Given its 
importance to emissions control and the many factors that can reduce 
its effectiveness, the catalyst is one of the most important components 
to be monitored.
b. Catalytic Converter Monitoring Requirements
    We are proposing that the OBD system monitor the catalyst system 
for proper conversion capability. Specifically, the OBD system would be 
required to detect a catalyst system malfunction when the catalyst 
system's conversion capability decreases to the point that any of the 
following occurs:
     NMHC and/or NOX emissions exceed the emissions 
thresholds for the ``catalytic converter system'' as shown in Table 
II.C-1.
    For purposes of determining the catalyst system malfunction 
criteria the manufacturer would be required to use a catalyst system 
deteriorated to the malfunction criteria using methods established by 
the manufacturer to

[[Page 3230]]

represent real world catalyst deterioration under normal and 
malfunctioning operating conditions. The malfunction criteria must be 
established by using a catalyst system with all monitored and 
unmonitored catalysts simultaneously deteriorated to the malfunction 
criteria.\40\ For engines using fuel shutoff to prevent over-fueling 
during misfire conditions (see section II.C.2), the malfunction 
criteria could be established using a catalyst system with all 
monitored catalysts simultaneously deteriorated to the malfunction 
criteria and all unmonitored catalysts deteriorated to the end of the 
engine's useful life.
---------------------------------------------------------------------------

    \40\ The unmonitored portion of the catalyst system would be 
that portion downstream of the sensor(s) used for catalyst 
monitoring.
---------------------------------------------------------------------------

c. Catalytic Converter Monitoring Conditions
    We are proposing that manufacturers define the monitoring 
conditions for malfunctions of the catalytic converter system such that 
the minimum performance ratio requirements discussed in section II.E 
would be met. For purposes of tracking and reporting as required in 
section II.E, all monitors used to detect malfunctions of the catalytic 
converter system during its normal operation must be tracked separately 
but reported as a single set of values as specified in section II.E.
d. Catalytic Converter MIL Illumination and DTC Storage
    We are proposing the general requirements for MIL illumination and 
DTC storage as discussed in section II.A.2. Note that the monitoring 
method for the catalyst(s) would have to be capable of detecting all 
instances, except diagnostic self-clearing, when a catalyst DTC has 
been cleared but the catalyst has not been replaced (e.g., catalyst 
over temperature histogram approaches are not acceptable).
7. Evaporative Emission Control System Monitoring
a. Background
    The evaporative emission control system controls HC emissions that 
would otherwise evaporate from the vehicle's fuel tank and fuel lines. 
Should any leak develop in the evaporative emission control system--
e.g., a disconnected hose--the HC emissions can be quite high and well 
over the evaporative emissions standards. Additionally, evaporative 
purge system defects--e.g., deteriorated vacuum lines, damaged 
canisters, non-functioning purge control valves--may occur which could 
also result in very high evaporative emissions.
b. Evaporative System Monitoring Requirements
    We are proposing that the OBD system verify purge flow from the 
evaporative system and detect any vapor leaks from the complete 
evaporative system, excluding the tubing and connections between the 
purge valve and the intake manifold. Individual components of the 
evaporative system (e.g. valves, sensors) must be monitored in 
accordance with the comprehensive components requirements discussed in 
section II.D.4.
    The OBD system would be required to detect an evaporative system 
malfunction when any of the following conditions exist:
     No purge flow from the evaporative system to the engine 
can be detected by the OBD system (i.e., the ``purge flow'' 
requirement); or
     For the 2010 and later model years, the complete 
evaporative system contains a leak or leaks that cumulatively are 
greater than or equal to a leak caused by a 0.150 inch diameter orifice 
(i.e., the ``system leak'' requirement).\41\
---------------------------------------------------------------------------

    \41\ In their HDOBD regulation, 13 CCR 1971.1, CARB defines 
``orifice'' as an O'Keefe Controls Co. precision metal ``Type B'' 
orifice with NPT connections with a diameter of the specified 
dimension (e.g., part number B-31-SS for a stainless steel 0.031 
inch diameter orifice).
---------------------------------------------------------------------------

    If the most reliable monitoring method available cannot reliably 
detect a system leak as specified above, a manufacturer may design 
their system to detect a larger leak. The manufacturer would be 
required to provide data and/or engineering analyses that demonstrate 
the inability of the monitor to reliably detect the required leak and 
their justification for detecting at their proposed orifice size. 
Further, if the manufacturer can demonstrate that leaks of the required 
size cannot cause evaporative or running loss emissions to exceed 1.5 
times the applicable evaporative emissions standards, the Administrator 
would revise upward the required leak size to the size demonstrated by 
the manufacturer that would result in emissions exceeding 1.5 times the 
standards.
c. Evaporative System Monitoring Conditions
    We are proposing that manufacturers define the monitoring 
conditions for both purge flow and system leak malfunctions such that 
the minimum performance ratio requirements discussed in section II.E 
would be met. For purposes of tracking and reporting as required in 
section II.E, all monitors used to detect system leak malfunctions must 
be tracked separately but reported as a single set of values as 
specified in section II.E.
    Manufacturers may disable or abort an evaporative emission control 
system monitor when the fuel tank level is over 85 percent of nominal 
tank capacity or during a refueling event. Manufacturers may design 
their evaporative emission control system monitor such that it executes 
only during drive cycles determined by the manufacturer to be cold 
starts if such a condition is needed to ensure reliable monitoring. The 
manufacturer would have to provide data and/or an engineering 
evaluation demonstrating that a reliable check can only be made on 
drive cycles when the cold start criteria are satisfied. However, the 
manufacturer may not determine a cold start solely on the basis that 
ambient temperature is higher than engine coolant temperature at engine 
start. Lastly, manufacturers would be allowed to disable temporarily 
the evaporative purge system to perform an evaporative system leak 
check.
d. Evaporative System MIL Illumination and DTC Storage
    We are proposing the general requirements for MIL illumination and 
DTC storage as discussed in section II.A.2, with an exception for leaks 
associated with the fuel filler cap. If the OBD system is capable of 
discerning that a system leak is being caused by a missing or 
improperly secured fuel filler cap, the manufacturer is not required to 
illuminate the MIL or store a DTC provided the vehicle is equipped with 
an alternative indicator for notifying the vehicle operator of the fuel 
filler cap ``malfunction.'' The alternative indicator would have to be 
of sufficient illumination and location to be readily visible to the 
vehicle operator under all lighting conditions. However, if the vehicle 
is not equipped with an alternative indicator and, instead, the MIL is 
illuminated to inform the operator of the ``malfunction,'' the MIL may 
be extinguished and the corresponding DTC(s) erased once the OBD system 
has verified that the fuel filler cap has been securely fastened and 
the MIL has not been commanded ON for any other type of malfunction. 
The Administrator may approve other strategies provided the 
manufacturer was able to demonstrate that the vehicle operator would be 
promptly notified of the missing or improperly secured fuel filler cap 
and that the notification would reasonably result in corrective action 
being undertaken.

[[Page 3231]]

8. Exhaust Gas Sensor Monitoring
a. Background
    Exhaust gas sensors (e.g., oxygen sensors, air-fuel ratio (A/F) 
sensors) are a critical element of the emissions control system on 
gasoline engines. In addition to maintaining a stoichiometric air-fuel 
mixture and, thus, helping to achieve the lowest possible emissions, 
these sensors are also used for enhancing the performance of several 
emission control technologies--e.g., catalysts, EGR systems). Many 
modern vehicles control the fuel supply with an oxygen sensor feedback 
system to maintain stoichiometry. Oxygen sensors are located typically 
in the exhaust system upstream and downstream of the catalytic 
converters. The front, or upstream, oxygen sensor is used generally for 
fuel control. The rear, or downstream, oxygen sensor is used generally 
for adjusting the front oxygen sensor signal as it drifts slightly with 
age related deterioration--often referred to as fuel trimming--and for 
onboard monitoring the catalyst system. Many vehicles use A/F sensors 
in lieu of the more conventional oxygen sensors since A/F sensors 
provide a precise reading of the actual air-fuel ratio.
    We expect that heavy-duty gasoline manufacturers will use both of 
these types of sensors to optimize their emissions control strategies 
and to satisfy many of the proposed heavy-duty OBD monitoring 
requirements--fuel system monitoring, catalyst monitoring, EGR system 
monitoring. Since exhaust gas sensors can be a critical component of an 
engine's fuel and emissions control system, their proper performance 
needs to be assured to maintain low emissions. Thus, any malfunction 
that adversely affects the performance of any of these exhaust gas 
sensors should be detected by the OBD system.
b. Exhaust Gas Sensor Monitoring Requirements
    We are proposing that the OBD system monitor the output signal, 
response rate, and any other parameter that could affect emissions of 
all primary (i.e., fuel control) exhaust gas sensors for malfunction. 
Both the lean to rich and rich to lean response rates must be 
monitored. In addition, we are proposing that the OBD system monitor 
all secondary exhaust gas sensors (i.e., those used for fuel trimming 
or as a monitoring device for another system) for proper output signal, 
activity, and response rate. For engines equipped with heated exhaust 
gas sensors, the OBD system would be required to monitor the sensor 
heater for proper performance.
i. Primary Exhaust Gas Sensors
    We are proposing that the OBD system detect a malfunction prior to 
any failure or deterioration of the exhaust gas sensor output voltage, 
resistance, impedance, current, response rate, amplitude, offset, or 
other characteristic(s) (including drift or bias corrected for by 
secondary sensors) that would cause an engine's emissions to exceed the 
emissions thresholds for ``other monitors'' as shown in Table II.C-1. 
The OBD system would also be required to detect the following exhaust 
gas sensor malfunctions:
     Those caused by either a lack of circuit continuity or 
out-of-range values.
     Those where a sensor failure or deterioration causes the 
fuel system to stop using that sensor as a feedback input (e.g., causes 
default or open-loop operation).
     Those where the sensor output voltage, resistance, 
impedance, current, amplitude, activity, or other characteristics are 
no longer sufficient for use as an OBD system monitoring device (e.g., 
for catalyst monitoring).
ii. Secondary Exhaust Gas Sensors
    We are proposing that the OBD system detect a malfunction prior to 
any failure or deterioration of the exhaust gas sensor voltage, 
resistance, impedance, current, response rate, amplitude, offset, or 
other characteristic(s) that would cause an engine's emissions to 
exceed the emissions thresholds for ``other monitors'' as shown in 
Table II.C-1. The OBD system would also be required to detect the 
following exhaust gas sensor malfunctions:
     Those caused by either a lack of circuit continuity or 
out-of-range values.
     Those where a sensor failure or deterioration causes the 
fuel system to stop using that sensor as a feedback input (e.g., causes 
default or open-loop operation).
     Those where the sensor output voltage, resistance, 
impedance, current, amplitude, activity, or other characteristics are 
no longer sufficient for use as an OBD system monitoring device (e.g., 
for catalyst monitoring).
iii. Exhaust Gas Sensor Heaters
    We are proposing that the OBD system detect a malfunction of the 
sensor heater performance when the current or voltage drop in the 
heater circuit is no longer within the manufacturer's specified limits 
for normal operation (i.e., within the criteria required by the 
component vendor for heater circuit performance at high mileage). The 
manufacturer may use other malfunction criteria for heater performance 
malfunctions. To do so, the manufacturer would be required to submit 
data and/or engineering analyses that demonstrate that the monitoring 
reliability and timeliness would be equivalent to the criteria stated 
here.
    In addition, the OBD system would be required to detect 
malfunctions of the heater circuit including open or short circuits 
that conflict with the commanded state of the heater (e.g., shorted to 
12 Volts when commanded to 0 Volts (ground)).
c. Exhaust Gas Sensor Monitoring Conditions
i. Primary Exhaust Gas Sensors
    We are proposing that manufacturers define the monitoring 
conditions for primary exhaust gas sensor malfunctions causing 
exceedance of the applicable thresholds and/or inability to perform as 
an OBD monitoring device such that the minimum performance ratio 
requirements discussed in section II.E would be met. For purposes of 
tracking and reporting as required in section II.E, all such monitors 
must be tracked separately but reported as a single set of values as 
specified in section II.E.
    Monitoring for primary exhaust gas sensor malfunctions related to 
circuit continuity, out-of-range, and open-loop operation must be done 
continuously with the exception that manufacturers may disable 
continuous exhaust gas sensor monitoring when an exhaust gas sensor 
malfunction cannot be distinguished from other effects. As an example, 
a manufacturer may disable monitoring for out-of-range on the low side 
during conditions where fuel has been cut (i.e., shut off temporarily). 
To do so, the manufacturer would have to submit data and/or engineering 
analyses that demonstrate that a properly functioning sensor cannot be 
distinguished from a malfunctioning sensor and that the disablement 
interval is limited only to that necessary for avoiding a false 
detection.
ii. Secondary Exhaust Gas Sensors
    We are proposing that manufacturers define the monitoring 
conditions for secondary exhaust gas sensor malfunctions causing 
exceedance of the applicable emissions thresholds, lack of circuit 
continuity, and/or inability to perform as an OBD monitoring device 
such that the minimum performance ratio requirements discussed in 
section II.E would be met.
    Monitoring for secondary exhaust gas sensor malfunctions related to 
out-of-

[[Page 3232]]

range and open loop operation must be done continuously with the 
exception that manufacturers may disable continuous exhaust gas sensor 
monitoring when an exhaust gas sensor malfunction cannot be 
distinguished from other effects. As an example, a manufacturer may 
disable monitoring for out-of-range on the low side during conditions 
where fuel has been cut (i.e., shut off temporarily). To do so, the 
manufacturer would have to submit data and/or engineering analyses that 
demonstrate that a properly functioning sensor cannot be distinguished 
from a malfunctioning sensor and that the disablement interval is 
limited only to that necessary for avoiding a false detection.
iii. Sensor Heaters
    We are proposing that manufacturers define monitoring conditions 
for sensor heater performance malfunctions such that the minimum 
performance ratio requirements discussed in section II.E would be met. 
Monitoring for sensor heater circuit malfunctions must be done 
continuously.
d. Exhaust Gas Sensor MIL Illumination and DTC Storage
    We are proposing the general requirements for MIL illumination and 
DTC storage as discussed in section II.A.2.

D. Monitoring Requirements and Timelines for Other Diesel and Gasoline 
Systems

1. Variable Valve Timing and/or Control (VVT) System Monitoring
a. Background
    Variable valve timing (VVT) and/or control systems are used 
primarily to optimize engine performance and have many advantages over 
conventional valve control. Instead of opening and closing the valves 
by fixed amounts and at fixed times, VVT controls can vary the timing 
of valve opening/closing and vary the effective size of the valve 
opening itself (in some systems) depending on the driving conditions 
(e.g., high engine speed and load). This feature permits a better 
compromise between performance, driveability, and emissions than 
conventional systems. With more stringent NOX emission 
standards being phased in, more vehicles are anticipated to use VVT. By 
doing so, some exhaust gas can be retained in the combustion chamber 
thereby reducing peak combustion temperatures and, hence, 
NOX emissions (known as ``internal EGR'').
b. VVT and/or Control System Monitoring Requirements
    We are proposing that the OBD system monitor the VVT system on 
engines so equipped for target error and slow response malfunctions. 
The individual electronic components (e.g., actuators, valves, sensors) 
that are used in the VVT system must be monitored in accordance with 
the comprehensive components requirements in section II.D.4.
i. VVT Target Error Malfunctions
    We are proposing that the OBD system detect a malfunction prior to 
any failure or deterioration in the capability of the VVT system to 
achieve the commanded valve timing and/or control within a crank angle 
and/or lift tolerance that would cause an engine's emissions to exceed 
the emissions thresholds for ``other monitors'' as shown in Table II.B-
1 for diesel engines or Table II.C-1 for gasoline engines. For engines 
in which no failure or deterioration of the VVT system could result in 
an engine's emissions exceeding the applicable emissions thresholds, 
the OBD system would have to detect a malfunction of the VVT system 
when proper functional response of the system to computer commands does 
not occur.
ii. VVT Slow Response Malfunctions
    We are proposing that the OBD system detect a malfunction prior to 
any failure or deterioration in the capability of the VVT system to 
achieve the commanded valve timing and/or control within a 
manufacturer-specified time that would cause an engine's emissions to 
exceed the emissions thresholds for ``other monitors'' as shown in 
Table II.B-1 for diesel engines or Table II.C-1 for gasoline engines. 
For engines in which no failure or deterioration of the VVT system 
could result in an engine's emissions exceeding the applicable 
emissions thresholds, the OBD system would have to detect a malfunction 
of the VVT system when proper functional response of the system to 
computer commands does not occur.
c. VVT and/or Control System Monitoring Conditions
    We are proposing that manufacturers define the monitoring 
conditions for VVT target error or slow response malfunctions such that 
the minimum performance ratio requirements discussed in section II.E 
would be met with the exception that monitoring shall occur every time 
the monitoring conditions are met during the driving cycle rather than 
once per driving cycle as required for most monitors. For purposes of 
tracking and reporting as required in section II.E, all monitors used 
to detect all VVT related malfunctions would have to be tracked 
separately but reported as a single set of values as specified in 
section II.E.\42\
---------------------------------------------------------------------------

    \42\ For specific components or systems that have multiple 
monitors that are required to be reported (e.g., exhaust gas sensor 
bank 1 may have multiple monitors for sensor response or other 
sensor characteristics), the OBD system must separately track 
numerators and denominators for each of the specific monitors and 
report only the corresponding numerator and denominator for the 
specific monitor that has the lowest numerical ratio. If two or more 
specific monitors have identical ratios, the corresponding numerator 
and denominator for the specific monitor that has the highest 
denominator shall be reported for the specific component.
---------------------------------------------------------------------------

d. VVT and/or Control System MIL Illumination and DTC Storage
    We are proposing the general requirements for MIL illumination and 
DTC storage as discussed in section II.A.2.
2. Engine Cooling System Monitoring
a. Background
    We are concerned about two elements of the engine cooling system. 
These elements are the thermostat and the engine coolant temperature 
sensor. Manufacturers typically use a thermostat to control the flow of 
coolant through the radiator and around the engine. During a cold 
engine start, the thermostat is closed typically which prevents the 
flow of coolant and serves to promote more rapid warm-up of the engine. 
As the coolant approaches a specific temperature, the thermostat begins 
to open allowing circulation of coolant through the radiator and around 
the engine. The thermostat then acts to regulate the coolant to the 
specified temperature. If the temperature rises above the regulated 
temperature, the thermostat opens further to allow more coolant to 
circulate, thus reducing the temperature. If the temperature drops 
below the regulated temperature, the thermostat partially closes to 
reduce the amount of coolant circulating, thereby increasing the 
temperature. If a thermostat malfunctions in such a manner that it does 
not adequately restrict coolant flow during vehicle warm-up, an 
increase in emissions could occur due to prolonged operation of the 
vehicle at temperatures below the stabilized, warmed-up value. This is 
particularly true at lower ambient temperatures--50 degrees Fahrenheit 
and below--but not so low that they are rare in the U.S. Equally 
important is that the engine coolant temperature is often used as an 
enable criterion for many OBD monitors. If the engine's coolant 
temperature does not reach the

[[Page 3233]]

manufacturer-specified warmed-up value, such monitors would be 
effectively disabled, perhaps indefinitely, and would, therefore, never 
detect malfunctions.
    Closely linked with the thermostat is the engine coolant 
temperature (ECT) sensor. Manufacturers typically use an ECT sensor as 
an input for many of the emission-related engine control systems. For 
gasoline engines, the ECT sensor is often one of the most important 
factors in determining when to begin closed-loop fuel control. If the 
engine coolant does not warm-up sufficiently, closed-loop fuel control 
is usually not engaged and the vehicle remains in open-loop fuel 
control. Since open-loop fuel control does not provide the precision of 
closed-loop control, the result is increased emissions levels. For 
diesel engines, the ECT sensor is often used to engage closed-loop 
control of the EGR system. Similar to closed-loop fuel control on 
gasoline engines, if the coolant temperature does not warm up, closed-
loop control of the EGR system would not engage which would result in 
increased emissions levels. In addition, for both gasoline and diesel 
engines, the ECT sensor may be used to enable many of the monitors that 
are being proposed. Such monitors would be effectively disabled and 
incapable of detecting malfunctions should the ECT sensor itself 
malfunction.
b. Engine Cooling System Monitoring Requirements
    We are proposing that the OBD system monitor the thermostat on 
engines so equipped for proper operation. We are also proposing that 
the OBD system monitor the ECT sensor for circuit continuity, out-of-
range values, and rationality faults. For engines that use an approach 
other than the cooling system and ECT sensor--e.g., oil temperature, 
cylinder head temperature--for an indication of engine operating 
temperature for emission control purposes (e.g., to modify spark or 
fuel injection timing or quantity), the manufacturer may forego cooling 
system monitoring in favor of monitoring the components or systems used 
in their approach. To do so, the manufacturer would be required to 
submit data and/or engineering analyses that demonstrate that their 
monitoring plan is as reliable and effective as the monitoring required 
for the engine cooling system.
i. Thermostat Monitoring Requirements
    We are proposing that the OBD system detect a thermostat 
malfunction if, within the manufacturer specified time interval 
following engine start, any of the following conditions occur:
     The coolant temperature does not reach the highest 
temperature required by the OBD system to enable other diagnostics;
     The coolant temperature does not reach a warmed-up 
temperature within 20 degrees Fahrenheit of the manufacturer's nominal 
thermostat regulating temperature. The manufacturer may use a lower 
temperature for this criterion provided the manufacturer can 
demonstrate that the fuel, spark timing, and/or other coolant 
temperature-based modification to the engine control strategies would 
not cause an emissions increase greater than or equal to 50 percent of 
any of the applicable emissions standards.
    The time interval specified by the manufacturer would have to be 
supported by the manufacturer via data and/or engineering analyses 
demonstrating that it provides robust monitoring and minimizes the 
likelihood of other OBD monitors being disabled. The manufacturer may 
use alternative malfunction criteria that are a function of temperature 
at engine start on engines that do not reach the temperatures specified 
in the malfunction criteria when the thermostat is functioning 
properly. To do so, the manufacturer would be required to submit data 
and/or engineering analyses that demonstrate that a properly operating 
system does not reach the specified temperatures and that the 
possibility is minimized for cooling system malfunctions to go 
undetected and disable other OBD monitors. In some cases, a 
manufacturer may forgo thermostat monitoring if the manufacturer can 
demonstrate that a malfunctioning thermostat cannot cause a measurable 
increase in emissions during any reasonable driving condition nor cause 
any disablement of other OBD monitors.
ii. Engine Coolant Temperature Sensor Monitoring Requirements
    We are proposing that the OBD system detect an ECT sensor 
malfunction when a lack of circuit continuity or an out-of-range value 
occurs. We are also proposing that the OBD system detect if, within the 
manufacturer specified time interval following engine start, the ECT 
sensor does not achieve the highest stabilized minimum temperature that 
is needed to initiate closed-loop/feedback control of all affected 
emission control systems (e.g., fuel system, EGR system). The 
manufacturer specified time interval would have to be a function of the 
engine coolant temperature and/or intake air temperature at startup. 
The manufacturer time interval would also have to be supported by the 
manufacturer via data and/or engineering analyses demonstrating that it 
provides robust monitoring and minimizes the likelihood of other OBD 
monitors being disabled. Manufacturers may forego the requirement to 
detect the ``time to closed loop/feedback enable temperature'' 
malfunction if the manufacturer does not use engine coolant temperature 
or the ECT sensor to enable closed-loop/feedback control of any 
emission control systems.
    We are also proposing that, to the extent feasible when using all 
available information, the OBD system must detect a malfunction if the 
ECT sensor inappropriately indicates a temperature below the highest 
minimum enable temperature required by the OBD system to enable other 
monitors. For example, an OBD system that requires an engine coolant 
temperature greater than 140 degrees Fahrenheit prior to enabling an 
OBD monitor must detect malfunctions that cause the ECT sensor to 
indicate inappropriately a temperature below 140 degrees Fahrenheit. 
Manufacturers may forego such monitoring within temperature regions in 
which the thermostat monitor or the ECT sensor ``time to reach closed-
loop/feedback enable temperature'' monitor would detect this ``stuck in 
a range below the highest minimum enable temperature'' ECT sensor 
malfunction.
    Lastly, we are proposing that, to the extent feasible when using 
all available information, the OBD system must detect a malfunction if 
the ECT sensor inappropriately indicates a temperature above the lowest 
maximum enable temperature required by the OBD system to enable other 
monitors. For example, an OBD system that requires an engine coolant 
temperature less than 90 degrees Fahrenheit at startup prior to 
enabling an OBD monitor must detect malfunctions that cause the ECT 
sensor to indicate inappropriately a temperature above 90 degrees 
Fahrenheit. Manufacturers may forego such monitoring within temperature 
regions in which the thermostat monitor, the ECT sensor ``time to reach 
closed-loop/feedback enable temperature'' monitor, or the ECT sensor 
``stuck in a range below the highest minimum enable temperature'' 
monitor would detect this ECT sensor ``stuck in a range above the 
lowest maximum enable temperature'' ECT sensor malfunction. The 
manufacturer may also forego such monitoring if the MIL would be 
illuminated for entering a ``limp home'' or default mode of

[[Page 3234]]

operation--e.g., for an over temperature protection strategy--as 
discussed in section II.A.2. Manufacturers may also forego this 
monitoring within temperature regions where the temperature gauge 
indicates a temperature in the engine overheating ``red zone'' should 
the vehicle have a temperature gauge on the instrument panel that 
displays the same temperature information as used by the OBD system 
(note that a temperature gauge would be required, not a temperature 
warning light).
c. Engine Cooling System Monitoring Conditions
i. Thermostat Monitoring Conditions
    We are proposing that manufacturers define the monitoring 
conditions for thermostat malfunctions in accordance with the general 
monitoring conditions for all engines described in section II.A.3. 
Additionally, monitoring for thermostat malfunctions would have to be 
done once per drive cycle on every drive cycle in which the ECT sensor 
indicates, at engine start, a temperature lower than the temperature 
established as the malfunction criteria in section II.D.2.b.i. 
Manufacturers would be allowed to disable thermostat monitoring at 
ambient engine start temperatures below 20 degrees Fahrenheit. 
Manufacturers may suspend or disable thermostat monitoring if the 
engine is subjected to conditions that could lead to false diagnosis 
(e.g., engine operation at idle for more than 50 percent of the warm-up 
time and/or hot restart conditions). To do so, the manufacturer must 
submit data and/or engineering analyses that demonstrate that the 
suspension or disablement is necessary. In general, the manufacturer 
would not be allowed to suspend or disable the thermostat monitor on 
engine starts where the engine coolant temperature at engine start is 
more than 35 degrees Fahrenheit lower than the thermostat malfunction 
threshold temperature.
ii. Engine Coolant Temperature Sensor Monitoring Conditions
    We are proposing that monitoring for ECT sensor circuit continuity 
and out-of-range malfunctions be done continuously. Manufacturers would 
be allowed to disable continuous ECT sensor monitoring when an ECT 
sensor malfunction cannot be distinguished from other effects. To do 
so, the manufacturer would have to submit test data and/or engineering 
evaluation that demonstrate that a properly functioning sensor cannot 
be distinguished from a malfunctioning sensor and that the disablement 
interval is limited only to that necessary for avoiding false 
detection.
    We are also proposing that manufacturers define the monitoring 
conditions for ``time to reach closed-loop/feedback enable 
temperature'' malfunctions in accordance with the general monitoring 
conditions for all engines described in section II.A.3. Additionally, 
monitoring for ``time to reach closed-loop/feedback enable 
temperature'' malfunctions would have to be conducted once per drive 
cycle on every drive cycle in which the ECT sensor at engine start 
indicates a temperature lower than the closed-loop enable temperature 
(i.e., all engine start temperatures greater than the ECT sensor out-
of-range low temperature and less than the closed-loop enable 
temperature). Manufacturers would be allowed to suspend or delay the 
``time to reach closed-loop/feedback enable temperature'' monitor if 
the engine is subjected to conditions that could lead to false 
diagnosis (e.g., vehicle operation at idle for more than 50 to 75 
percent of the warm-up time).
    We are also proposing that manufacturers define the monitoring 
conditions for ECT sensor ``stuck in a range below the highest minimum 
enable temperature'' and ``stuck in a range above the lowest maximum 
enable temperature'' malfunctions in accordance with the general 
monitoring conditions for all engines described in section II.A.3 and 
in accordance with the minimum performance ratio requirements discussed 
in section II.E.
d. Engine Cooling System MIL Illumination and DTC Storage
    We are proposing the general requirements for MIL illumination and 
DTC storage as discussed in section II.A.2.
3. Crankcase Ventilation System Monitoring
a. Background
    Crankcase emissions are the pollutants emitted in the gases that 
are vented from an engine's crankcase. These gases are also referred to 
as ``blowby gases'' because they result from engine exhaust from the 
combustion chamber ``blowing by'' the piston rings into the crankcase. 
These gases are vented to prevent high pressures from occurring in the 
crankcase. Our emission standards have historically prohibited 
crankcase emissions from all highway engines except turbocharged heavy-
duty diesel engines. The most common way to eliminate crankcase 
emissions has been to vent the blowby gases into the engine air intake 
system, so that the gases can be recombusted. We made the exception for 
turbocharged heavy-duty diesel engines in the past because of concerns 
about fouling that could occur by routing the diesel particulates 
(including engine oil) into the turbocharger and aftercooler. Newly 
developed closed crankcase filtration systems specifically designed for 
turbocharged heavy-duty diesel engines now allow the crankcase gases to 
be captured.
    In general, the crankcase ventilation system consists of a fresh 
air inlet hose, a crankcase vapor outlet hose, and a crankcase 
ventilation valve to control the flow through the system. Fresh air is 
introduced to the crankcase via the inlet (typically a connection from 
the intake air cleaner assembly). On the opposite side of the 
crankcase, vapors are vented from the crankcase through the valve by 
way of the outlet hose and then to the intake manifold. On gasoline 
engines, the intake manifold provides the vacuum that is needed to 
accomplish the circulation while the engine is running.
    For gasoline engines, the valve is used to regulate the amount of 
flow based on engine speed. During low engine load operation (e.g., 
idle), the valve is nearly closed allowing only a small portion of air 
to flow through the system. With open throttle conditions, the valve 
opens to allow more air into the system. At high engine load operation 
(i.e., hard accelerations), the valve begins to close again, limiting 
air flow to a small amount. For most systems, a mechanical valve is all 
that is necessary to adequately regulate crankcase ventilation system 
air flow. The crankcase ventilation system on diesel engines, while 
slightly different than that for gasoline engines, has essentially the 
same purpose and function.
    We do not believe that failures involving cracked or deteriorated 
hoses have a significant impact on crankcase emissions because vapors 
are drawn into the engine by intake manifold vacuum which suggests that 
fresh air would be drawn into the cracked hose rather than dirty 
exhaust being blown out of the cracked hose. The more likely cause of 
crankcase ventilation system malfunctions and excess emissions is 
improper service or tampering of the system. Such failures include 
misrouted or disconnected hoses and missing valves. Of these failures, 
hose disconnections on the vapor vent side of the system and/or missing 
valves can cause harmful crankcase emissions to be vented directly to 
the atmosphere.

[[Page 3235]]

b. Crankcase Ventilation System Monitoring Requirements
    We are proposing that the OBD system monitor the crankcase 
ventilation system on engines so equipped for system integrity. Engines 
not equipped with crankcase ventilation systems would be exempt from 
monitoring the crankcase ventilation system.
    Specifically for diesel engines, the manufacturer would be required 
to submit a plan for the monitoring strategy, malfunction criteria, and 
monitoring conditions prior to OBD certification. The plan would have 
to demonstrate the effectiveness of the strategy to monitor the 
performance of the crankcase ventilation system to the extent feasible 
with respect to the malfunction criteria below and the monitoring 
conditions required by the monitor.
    We are proposing that the OBD system detect a malfunction of the 
crankcase ventilation system when a disconnection of the system occurs 
between either the crankcase and the crankcase ventilation valve, or 
between the crankcase ventilation valve and the intake manifold. 
Manufacturers may forego detecting a disconnection between the 
crankcase and the crankcase ventilation valve provided the manufacturer 
can demonstrate that the crankcase ventilation system is designed such 
that the crankcase ventilation valve is fastened directly to the 
crankcase in a manner that makes it significantly more difficult to 
remove the valve from the crankcase than to disconnect the line between 
the valve and the intake manifold (aging effects must be taken into 
consideration). Manufacturers may also forego detecting a disconnection 
between the crankcase and the crankcase ventilation valve for system 
designs that use tubing between the valve and the crankcase provided 
the manufacturer can demonstrate that the connections between the valve 
and the crankcase are: (1) Resistant to deterioration or accidental 
disconnection; (2) significantly more difficult to disconnect than the 
line between the valve and the intake manifold; and, (3) not subject to 
disconnection per the manufacturer's repair procedures for non-
crankcase ventilation system repair work. Lastly, manufacturers may 
forego detecting a disconnection between the crankcase ventilation 
valve and the intake manifold upon determining that the disconnection: 
(1) Causes the vehicle to stall immediately during idle operation; or, 
(2) is unlikely to occur due to a crankcase ventilation system design 
that is integral to the induction system (e.g., machined passages 
rather than tubing or hoses).
c. Crankcase Ventilation System Monitoring Conditions
    We are proposing that manufacturers define the monitoring 
conditions for crankcase ventilation system malfunctions in accordance 
with the general monitoring conditions for all engines described in 
section II.A.3, and the minimum performance ratio requirements 
discussed in section II.E.
d. Crankcase Ventilation System MIL Illumination and DTC Storage
    We are proposing the general requirements for MIL illumination and 
DTC storage as discussed in section II.A.2. The stored DTC need not 
specifically identify the crankcase ventilation system (e.g., a DTC for 
idle speed control or fuel system monitoring can be stored) if the 
manufacturer can demonstrate that additional monitoring hardware would 
be necessary to make this identification, and provided the 
manufacturer's diagnostic and repair procedures for the detected 
malfunction include directions to check the integrity of the crankcase 
ventilation system.
4. Comprehensive Component Monitors
a. Background
    Comprehensive components is a term meant to capture essentially 
every other emissions related component not discussed above. 
Specifically, it covers all other electronic engine components or 
systems not mentioned above that either can affect vehicle emissions or 
are used as part of the OBD diagnostic strategy for another monitored 
component or system. Comprehensive components are generally identified 
as input components--i.e., those that provide input directly or 
indirectly to the onboard computer--or as output components and/or 
systems--i.e., those that receive commands from the onboard computer. 
Typical examples of input components include temperature sensors and 
pressure sensors, while examples of output components and/or systems 
include the idle control system, glow plugs, and wait-to-start lamps.
    While a malfunctioning comprehensive component may not have as much 
impact on emissions as a malfunctioning major emissions-related 
component, it still could result in a measurable increase in emissions. 
The proper performance of these components can be critical to both the 
proper functioning of major emissions-related components, and to the 
accurate monitoring of those components or systems. Malfunctions of 
comprehensive components that go undetected by the OBD system may 
disable or adversely affect the robustness of other OBD monitors 
without any awareness by the operator that a problem exists. Due to the 
vital role these components play, monitoring them properly is extremely 
important.
b. Comprehensive Component Monitoring Requirements
    We are proposing that the OBD system monitor for malfunction any 
electronic engine components/systems not otherwise described in 
sections above that either provides input to (directly or indirectly) 
or receives commands from the onboard computer(s), and: (1) Can affect 
emissions during any reasonable in-use driving condition; or, (2) is 
used as part of the diagnostic strategy for any other monitored system 
or component.\43\
---------------------------------------------------------------------------

    \43\ When referring to ``comprehensive components'' and their 
monitors, ``electronic engine components/systems'' is not meant to 
include components/systems that are driven by the engine yet are not 
related to the control of the fueling, air handling, or emissions of 
the engine (e.g., PTO components, air conditioning system 
components, and power steering components are not included).
---------------------------------------------------------------------------

    Input components required to be monitored may include the crank 
angle sensor, knock sensor, throttle position sensor, cam position 
sensor, intake air temperature sensor, boost pressure sensor, manifold 
pressure sensor, mass air flow sensor, exhaust temperature sensor, 
exhaust pressure sensor, fuel pressure sensor, and fuel composition 
sensor (e.g., flexible fuel vehicles). Output components/systems 
required to be monitored may include the idle speed control system, 
glow plug system, variable length intake manifold runner systems, 
supercharger or turbocharger electronic components, heated fuel 
preparation systems, the wait-to-start lamp on diesel applications, and 
the MIL. The manufacturer would be responsible for determining which 
input and output components/systems could affect emissions during any 
reasonable in-use driving condition. The manufacturer would be allowed 
to make this determination based on data or engineering judgment. 
However, if the Administrator reasonably believes that a manufacturer 
has incorrectly determined that a component/system cannot affect 
emissions, the manufacturer may be required to provide emissions data 
showing that the component/system, when malfunctioning and installed in 
a suitable test engine, does not have an emissions effect. Such 
emissions data may be requested for any reasonable driving condition.

[[Page 3236]]

i. Input Components
    We are proposing that the OBD system detect malfunctions of input 
components caused by a lack of circuit continuity, out-of-range values, 
and, where feasible, improper rationality. To the extent feasible, the 
rationality diagnostics should verify that a sensor's input to the 
onboard computer is neither inappropriately high nor inappropriately 
low (i.e., ``two-sided'' diagnostics should be used). Also to the 
extent feasible, the OBD system should detect and store different DTCs 
that distinguish rationality malfunctions from lack of circuit 
continuity malfunctions and out-of-range values. For lack of circuit 
continuity malfunctions and out-of-range values, the OBD system should 
detect and store different DTCs for each distinct malfunction (e.g., 
out-of-range low, out-of-range high, open circuit). The OBD system is 
not required to store separate DTCs for lack of circuit continuity 
malfunctions that cannot be distinguished from malfunctions associated 
with out-of-range values.
    For input components that are used to activate alternative 
strategies that can affect emissions (e.g., AECDs, engine shutdown 
systems), the OBD system would be required to detect rationality 
malfunctions that cause the system to erroneously activate or 
deactivate the alternative strategy. To the extent feasible when using 
all available information, the rationality diagnostics should detect a 
malfunction if the input component inappropriately indicates a value 
that activates or deactivates the alternative strategy. For example, if 
an alternative strategy requires an intake air temperature greater than 
120 degrees Fahrenheit prior to activating, the OBD system should 
detect malfunctions that cause the intake air temperature sensor to 
inappropriately indicate a temperature above 120 degrees Fahrenheit.
    For engines that require precise alignment between the camshaft and 
the crankshaft, the OBD system would be required to monitor the 
crankshaft position sensor(s) and camshaft position sensor(s) to verify 
proper alignment between the camshaft and crankshaft. The OBD system 
would also have to monitor the sensors for circuit continuity and 
rationality malfunctions. Such monitoring for proper alignment between 
a camshaft and a crankshaft would only be required in cases where both 
are equipped with position sensors.
    For engines equipped with VVT systems and a timing belt or chain, 
the OBD system must detect a malfunction if the alignment between the 
camshaft and crankshaft is off by one or more cam/crank sprocket cogs 
(e.g., the timing belt/chain has slipped by one or more teeth/cogs). If 
a manufacturer demonstrates that a single tooth/cog misalignment cannot 
cause a measurable increase in emissions during any reasonable driving 
condition, the OBD system would be required to detect a malfunction 
when the minimum number of teeth/cogs misalignment needed to cause a 
measurable emission increase has occurred.
ii. Output Components/Systems
    We are proposing that the OBD system detect a malfunction of an 
output component/system when proper functional response of the 
component/system to computer commands does not occur. If a functional 
check is not feasible, the OBD system would be required to detect 
malfunctions caused by a lack of circuit continuity (e.g., short to 
ground or high voltage). For output component malfunctions associated 
with the lack of circuit continuity, the OBD system is not required to 
store different DTCs for each distinct malfunction (e.g., open circuit, 
shorted low). Further, manufacturers would not be required to activate 
an output component/system when it would not normally be active for the 
exclusive purpose of performing functional monitoring of output 
components/systems.
    Additionally, the idle control system would have to be monitored 
for proper functional response to computer commands. For gasoline 
engines that use monitoring strategies based on deviation from target 
idle speed, a malfunction would have to be detected when either of the 
following conditions occur: (a) The idle speed control system cannot 
achieve the target idle speed within 200 revolutions per minute (rpm) 
above the target speed or 100 rpm below the target speed--the OBD 
system could use larger engine speed tolerances provided the 
manufacturer is able to demonstrate via data and/or engineering 
analyses that the tolerances can be exceeded without a malfunction 
being present; or, (b) the idle speed control system cannot achieve the 
target idle speed within the smallest engine speed tolerance range 
required by the OBD system to enable any other OBD monitors. For diesel 
engines, a malfunction would have to be detected when either of the 
following conditions occur: (a) The idle fuel control system cannot 
achieve the target idle speed or fuel injection quantity within +/-50 
percent of the manufacturer-specified fuel quantity and engine speed 
tolerances; or, (b) the idle fuel control system cannot achieve the 
target idle speed or fueling quantity within the smallest engine speed 
or fueling quantity tolerance range required by the OBD system to 
enable any other OBD monitors.
    Glow plugs and intake air heater systems would also have to be 
monitored for proper functional response to computer commands and for 
malfunctions associated with circuit continuity. The glow plug and 
intake air heater circuit(s) would have to be monitored for proper 
current and voltage drop. The manufacturer may use other monitoring 
strategies by submitting data and/or engineering analyses that 
demonstrate that the strategy provides equally reliable and timely 
detection of malfunctions. In general, the OBD system would have to 
detect a malfunction when a single glow plug no longer operates within 
the manufacturer's specified limits for normal operation. If a 
manufacturer demonstrates that a single glow plug malfunction cannot 
cause a measurable increase in emissions during any reasonable driving 
condition, the OBD system must detect a malfunction for the minimum 
number of glow plugs needed to cause an emissions increase. Further, to 
the extent feasible without adding additional hardware for this 
purpose, the stored DTC must identify the specific malfunctioning glow 
plug(s).
    Lastly, the wait-to-start lamp circuit and the MIL circuit would 
have to be monitored for malfunctions that cause either lamp to fail to 
illuminate when commanded on (e.g., burned out bulb).
c. Comprehensive Component Monitoring Conditions
i. Input Components
    We are proposing that input components be monitored continuously 
for circuit continuity and for providing values within the proper 
range. For rationality monitoring, where applicable, manufacturers 
would define the monitoring conditions for detecting malfunctions in 
accordance with the general monitoring conditions for all engines 
described in section II.A.3 and the minimum performance ratio 
requirements described in section II.E except that rationality 
monitoring would have to occur every time the monitoring conditions are 
met during the drive cycle rather than once per drive cycle as required 
in section II.A.3.
    A manufacturer may disable continuous monitoring for circuit 
continuity, and for providing values within the proper range, when a

[[Page 3237]]

malfunction cannot be distinguished from other effects. To do so, the 
manufacturer would have to submit data and/or engineering analyses that 
demonstrate that a properly functioning input component cannot be 
distinguished from a malfunctioning input component and that the 
disablement interval is limited only to that necessary for avoiding 
false detection.
ii. Output Components/Systems
    We are proposing that output components/systems be monitored 
continuously for circuit continuity. For functional monitoring, 
manufacturers would define the monitoring conditions for detecting 
malfunctions in accordance with the general monitoring conditions for 
all engines described in section II.A.3 and the minimum performance 
ratio requirements described in section II.E.
    For the idle control system, we are proposing that manufacturers 
define the monitoring conditions for functional monitoring in 
accordance with the general monitoring conditions for all engines 
described in section II.A.3 and the minimum performance ratio 
requirements described in section II.E except that functional 
monitoring would have to occur every time the monitoring conditions are 
met during the drive cycle rather than once per drive cycle as required 
in section II.A.3.
    A manufacturer may disable continuous monitoring for circuit 
continuity when a malfunction cannot be distinguished from other 
effects. To do so, the manufacturer would have to submit data and/or 
engineering analyses that demonstrate that a properly functioning 
output component cannot be distinguished from a malfunctioning output 
component and that the disablement interval is limited only to that 
necessary for avoiding false detection.
d. Comprehensive Component MIL Illumination and DTC Storage
    With a couple of exceptions, we are proposing the general 
requirements for MIL illumination and DTC storage as discussed in 
section II.A.2. The exceptions to this being that MIL illumination 
would not be required in conjunction with storing a MIL-on DTC for any 
comprehensive component if: (a) The component or system, when 
malfunctioning, could not cause engine emissions to increase by 15 
percent or more of the FTP standard during any reasonable driving 
condition; and, (b) the component or system is not used as part of the 
diagnostic strategy for any other monitored system or component. MIL 
illumination is also not required if a malfunction has been detected in 
the MIL circuit that prevents the MIL from illuminating (e.g., burned 
out bulb or light emitting diode (LED)). However, the electronic MIL 
status must be reported as ``commanded on'' and a MIL-on DTC would have 
to be stored.
5. Other Emissions Control System Monitoring
a. Background
    As noted above, the primary purpose of OBD is to detect 
malfunctions in the engine and/or emissions control system. Therefore, 
we are proposing that manufacturers be required to submit to the 
Administrator a monitoring plan for any new engine and/or emissions 
control technology not otherwise described above. Such technology might 
include hydrocarbon traps or homogeneous charge compression ignition 
(HCCI) systems. This would allow manufacturers and EPA to evaluate the 
new technology and determine an appropriate level of monitoring that 
would be both technologically feasible and consistent with the 
monitoring requirements for the other emissions control devices 
described above.
    As proposed, the Administrator would provide guidance as to what 
type of components would fall under the ``other emissions control 
system'' requirements and which would fall under the comprehensive 
component requirements. Specifically, we are concerned that uncertainty 
may arise for those emission control components or systems that also 
meet the definition of electronic engine components. As such, the 
proposal would delineate the two by requiring components/systems that 
fit both definitions but are not corrected or compensated for by the 
adaptive fuel control system to be monitored as ``other emissions 
control devices'' rather than as comprehensive components. A typical 
device that would fall under this category instead of the comprehensive 
components category because of this delineation would be a swirl 
control valve system. Such delineation is necessary because such 
emissions control components generally require more thorough monitoring 
than comprehensive components to ensure low emissions levels throughout 
an engine's life. Further, emissions control components that are not 
compensated for by the fuel control system as they age or deteriorate 
can have a larger impact on tailpipe emissions than is typical of 
comprehensive components that are corrected for by the fuel control 
system as they deteriorate.
    Note that the Administrator does not foresee any outcome where a 
promising new emissions control technology would be prohibited based 
solely on the lack of an OBD monitoring strategy for it. Instead, we 
want to instill in manufacturers the need to consider OBD monitoring 
when developing any new emissions control technology. Further, we want 
to instill in manufacturers the sense that an OBD monitoring strategy 
will, one day, be necessary so a plan for such should exist prior to 
introducing the technology on new products.
b. Other Emissions Control System Monitoring Requirements/Conditions
    We are proposing that, for other emission control systems that are: 
(1) Not identified or addressed in sections II.B through II.D.4--e.g., 
hydrocarbon traps, HCCI control systems; or, (2) identified or 
addressed in section II.D.4 but not corrected or compensated for by an 
adaptive control system--e.g., swirl control valves, manufacturers 
would be required to submit a plan for Administrator approval of the 
monitoring strategy, the malfunction criteria, and the monitoring 
conditions prior to introduction on a production engine. Administrator 
approval of the plan would be based on the effectiveness of the 
monitoring strategy, the robustness of the malfunction criteria, and 
the frequency of meeting the necessary monitoring conditions.
    We are also proposing that, for engines that use emissions control 
systems that alter intake air flow or cylinder charge characteristics 
by actuating valve(s), flap(s), etc., in the intake air delivery system 
(e.g., swirl control valve systems), the manufacturers, in addition to 
meeting the requirements above, may elect to have the OBD system 
monitor the shaft to which all valves in one intake bank are physically 
attached rather than monitoring the intake air flow, cylinder charge, 
or individual valve(s)/flap(s) for proper functional response. For non-
metal shafts or segmented shafts, the monitor must verify all shaft 
segments for proper functional response (e.g., by verifying the segment 
or portion of the shaft furthest from the actuator functions properly). 
For systems that have more than one shaft to operate valves in multiple 
intake banks, manufacturers are not required to add more than one set 
of detection hardware (e.g., sensor, switch) per intake bank to meet 
this requirement.

[[Page 3238]]

6. Exceptions to Monitoring Requirements
a. Background
    Under some conditions, the reliability of specific monitors may be 
diminished significantly. Therefore, we are proposing to allow 
manufacturers to disable the affected monitors when these conditions 
are encountered in-use. These include situations of extreme conditions 
(e.g., very low ambient temperatures, high altitudes) and of periods 
where default modes of operation are active (e.g., when a tire pressure 
problem is detected). In some of these cases, we may allow 
manufacturers to revise the emission malfunction threshold to ensure 
the most reliable monitoring performance.
b. Requirements for Exceptions to Monitoring
    The Administrator may revise the emission threshold for any 
monitor, or revise the PM filtering performance malfunction criteria 
for DPFs to exclude detection of specific failure modes such as 
partially melted substrates, if the most reliable monitoring method 
developed requires a higher threshold or, in the case of PM filtering 
performance, the exclusion of specific failure modes, to prevent 
significant errors of commission in detecting a malfunction. The 
Administrator would notify the industry of any such revisions to ensure 
that all manufacturers would be able to implement OBD on an equal 
basis. In other words, we would not allow one manufacturer to revise a 
specific monitoring threshold upwards while insisting that another meet 
the proposed threshold.
    Manufacturers may disable an OBD system monitor at ambient engine 
start temperatures below 20 degrees Fahrenheit (low ambient temperature 
conditions may be determined based on intake air or engine coolant 
temperature at engine start) or at elevations higher than 8000 feet 
above sea level. To do so, the manufacturer would have to submit data 
and/or engineering analyses that demonstrate that monitoring would be 
unreliable during the disable conditions. A manufacturer may request 
that an OBD system monitor be disabled at other ambient engine start 
temperatures by submitting data and/or engineering analyses 
demonstrating that misdiagnosis would occur at the given ambient 
temperatures due to their effect on the component itself (e.g., 
component freezing).
    Manufacturers may disable an OBD system monitor when the fuel level 
is 15 percent or less of the nominal fuel tank capacity for those 
monitors that can be affected by low fuel level or running out of fuel 
(e.g., misfire detection). To do so, the manufacturer would have to 
submit data and/or engineering analyses that demonstrate that both 
monitoring at the given fuel levels would be unreliable, and the OBD 
system is still able to detect a malfunction if the component(s) used 
to determine fuel level indicates erroneously a fuel level that causes 
the disablement.
    Manufacturers may disable OBD monitors that can be affected by 
vehicle battery or system voltage levels. For an OBD monitor affected 
by low vehicle battery or system voltages, manufacturers may disable 
monitoring when the battery or system voltage is below 11.0 Volts. 
Manufacturers may use a voltage threshold higher than 11.0 Volts to 
disable monitors but would have to submit data and/or engineering 
analyses that demonstrate that monitoring at those voltages would be 
unreliable and that either operation of a vehicle below the disablement 
criteria for extended periods of time is unlikely or the OBD system 
monitors the battery or system voltage and would detect a malfunction 
at the voltage used to disable other monitors.
    For monitoring systems affected by high vehicle battery or system 
voltages, manufacturers may disable monitoring when the battery or 
system voltage exceeds a manufacturer-defined voltage. To do so, the 
manufacturer would have to submit data and/or engineering analyses that 
demonstrate that monitoring above the manufacturer-defined voltage 
would be unreliable and that either the electrical charging system/
alternator warning light would be illuminated (or voltage gauge would 
be in the ``red zone'') or the OBD system monitors the battery or 
system voltage and would detect a malfunction at the voltage used to 
disable other monitors.
    A manufacturer may also disable affected OBD monitors in vehicles 
designed to accommodate the installation of power take off (PTO) units 
provided disablement occurs only while the PTO unit is active and the 
OBD readiness status is cleared by the onboard computer (i.e., all 
monitors set to indicate ``not complete'') while the PTO unit is 
activated (see section II.F.4 below). If the disablement occurs, the 
readiness status may be restored, when the disablement ends, to its 
state prior to PTO activation.

E. A Standardized Method To Measure Real World Monitoring Performance

    As was noted in section II.A.3, manufacturers determine the most 
appropriate times to run the non-continuous OBD monitors. This way, 
they are able to make their OBD evaluation either at the operating 
condition when an emissions control system is active and its 
operational status can best be evaluated, and/or at the operating 
condition when the most accurate evaluation can be made (e.g., highly 
transient conditions or extreme conditions can make evaluation 
difficult). Importantly, manufacturers are prohibited from using a 
monitoring strategy that is so restrictive such that it rarely or never 
runs. To help protect against monitors that rarely run, we are 
proposing an ``in-use monitor performance ratio'' requirement as 
described here.
    The set of operating conditions that must be met so that an OBD 
monitor can run are called the ``enable criteria'' for that given 
monitor. These enable criteria are often different for different 
monitors and may well be different for different types of engines. A 
large diesel engine intended for use in a Class 8 truck would be 
expected to see long periods of relatively steady-state operation while 
a smaller engine intended for use in an urban delivery truck would be 
expected to see a lot of transient operation. Manufacturers will need 
to balance between a rather loose set of enable criteria for their 
engines and vehicles given the very broad range of operation HD highway 
engines see and a tight set of enable criteria given the desire for 
greater monitor accuracy. Manufacturers would be required to design 
these enable criteria so that the monitor:
     Is robust (i.e., accurate at making pass/fail decisions);
     Runs frequently in the real world; and,
     In general, also runs during the FTP heavy-duty transient 
cycle.
    If designed incorrectly, these enable criteria may be either too 
broad and result in inaccurate monitors, or overly restrictive thereby 
preventing the monitor from executing frequently in the real world.
    Since the primary purpose of an OBD system is to monitor for and 
detect emission-related malfunctions while the engine is operating in 
the real world, a standardized methodology for quantifying real world 
performance would be beneficial to both EPA and manufacturers. 
Generally, in determining whether a manufacturer's monitoring 
conditions are sufficient, a manufacturer would discuss the proposed 
monitoring conditions with EPA staff. The finalized conditions would be 
included in the certification applications and submitted to EPA staff 
who would review the conditions and make determinations on a case-by-
case

[[Page 3239]]

basis based on the engineering judgment of the staff. In cases where we 
are concerned that the documented conditions may not be met during 
reasonable in-use driving conditions, we would most likely ask the 
manufacturer for data or other engineering analyses used by the 
manufacturer to determine that the conditions would occur in-use. In 
proposing a standardized methodology for quantifying real world 
performance, we believe this review process can be done more 
efficiently than would occur otherwise. Furthermore, it would serve to 
ensure that all manufacturers are held to the same standard for real 
world performance. Lastly, we want review procedures that will ensure 
that monitors operate properly and frequently in the field.
    Therefore, we are proposing that all manufacturers be required to 
use a standardized method for determining real world monitoring 
performance and to hold manufacturers liable if monitoring occurs less 
frequently than a minimum acceptable level, expressed as minimum 
acceptable in-use performance ratio. We are also proposing that 
manufacturers be required to implement software in the onboard computer 
to track how often several of the major monitors (e.g., catalyst, EGR, 
CDPF, other diesel aftertreatment devices) execute during real world 
driving. The onboard computer would keep track of how many times each 
of these monitors has executed and how much the engine has been 
operated. By measuring both of these values, the ratio of monitor 
operation relative to engine operation can be calculated to determine 
monitoring frequency.
    The proposed minimum acceptable frequency requirement would apply 
to many but not all of the OBD monitors. We are proposing that monitors 
be required to operate either continuously, once per drive cycle, or, 
in a few cases, multiple times per drive cycle (i.e., whenever the 
proper monitoring conditions are present). For components or systems 
that are more likely to experience intermittent failures or failures 
that can routinely happen in distinct portions of an engine's operating 
range (e.g., only at high engine speed and load, only when the engine 
is cold or hot), monitors would be required to operate continuously. 
Examples of continuous monitors include the fuel system monitor and 
most electrical/circuit continuity monitors. For components or systems 
that are less likely to experience intermittent failures or failures 
that only occur in specific vehicle operating regions or for components 
or systems where accurate monitoring can only be performed under 
limited operating conditions, monitors would be required to run once 
per drive cycle. Examples of once per drive cycle monitors typically 
include gasoline catalyst monitors, evaporative system leak detection 
monitors, and output comprehensive component functional monitors. For 
components or systems that are routinely used to perform functions that 
are crucial to maintaining low emissions but may still require 
monitoring under fairly limited conditions, monitors would be required 
to run each and every time the manufacturer-defined enable conditions 
are present. Examples of multiple times per drive cycle monitors 
typically include input comprehensive component rationality monitors 
and some exhaust aftertreatment monitors.
    Monitors required to run continuously, by definition, would always 
be running, thereby making a minimum frequency requirement moot. The 
new frequency requirement would essentially apply only to those 
monitors that are designated as once per drive cycle or multiple times 
per drive cycle monitors. For all of these monitors, manufacturers 
would be required to define monitoring conditions that ensure adequate 
frequency in-use. Specifically, the monitors would need to run often 
enough so that the measured monitor frequency on in-use engines would 
exceed the minimum acceptable frequency. However, even though the 
minimum frequency requirement would apply to nearly all once per drive 
cycle and multiple times per drive cycle monitors, manufacturers would 
only be required to implement software to track and report the in-use 
frequency for a few of the major monitors. These few monitors generally 
represent the major emissions control components and the ones with the 
most limited enable criteria.
    We believe that OBD monitors should run frequently to ensure early 
detection of emissions-related malfunctions and, consequently, to 
maintain low emissions. Allowing malfunctions to continue undetected 
and unrepaired for long periods of time allows emissions to increase 
unnecessarily. Frequent monitoring can also help to ensure detection of 
intermittent emissions-related malfunctions (i.e., those that are not 
continuously present but occur sporadically for days and even weeks at 
a time). The nature of mechanical and electrical systems is that 
intermittent malfunctions can and do occur. The less frequent the 
monitoring, the less likely these malfunctions will be detected and 
repaired. Additionally, for both intermittent and continuous 
malfunctions, earlier detection is equivalent to preventative 
maintenance in that the original malfunction can be detected and 
repaired prior to it causing subsequent damage to other components. 
This can help vehicle operators avoid more costly repairs that could 
have resulted had the first malfunction gone undetected.
    Infrequent monitoring can also have an impact on the service and 
repair industry. Specifically, monitors that have unreasonable or 
overly restrictive enable conditions could hinder vehicle repair 
services. In general, upon completing an OBD-related repair to an 
engine, a technician will attempt to verify that the repair has indeed 
fixed the problem. Ideally, a technician will operate the vehicle in a 
manner that will exercise the appropriate OBD monitor and allow the OBD 
system to confirm that the malfunction is no longer present. This 
affords a technician the highest level of assurance that the repair was 
indeed successful. However, OBD monitors that operate infrequently are 
difficult to exercise and, therefore, technicians may not be able (or 
may not be likely) to perform such post-repair evaluations. Despite the 
service information availability requirements we are proposing--
requirements that manufacturers make all of their service and repair 
information available to all technicians, including the information 
necessary to exercise OBD monitors--technicians would still find it 
difficult to exercise monitors that require infrequently encountered 
engine operating conditions (e.g., abnormally steady constant speed 
operation for an extended period of time). Additionally, to execute OBD 
monitors in an expeditious manner or to execute monitors that would 
require unusual or infrequently encountered conditions, technicians may 
be required to operate the vehicle in an unsafe manner (e.g., at 
freeway speeds on residential streets or during heavy traffic). If 
unsuccessful in executing these monitors, technicians may even take 
shortcuts in attempting to validate the repair while maintaining a 
reasonable cost for customers. These shortcuts would likely not be as 
thorough in verifying repairs and could increase the chance that 
improperly repaired engines would be returned to the vehicle owner or 
additional repairs would be performed just to ensure the problem is 
fixed. In the end, monitors that operate less frequently can result in 
unnecessary costs and inconvenience to both vehicle owners and 
technicians.

[[Page 3240]]

1. Description of Software Counters to Track Real World Performance
    As stated above, manufacturers would be required to track monitor 
peformance by comparing the number of monitoring events (i.e., how 
often each monitor has run) to the number of driving events (i.e., how 
often has the vehicle been operated). The ratio of these two numbers 
would give an indication of how often the monitor is operating relative 
to vehicle operation. In equation form, this can be stated as:
[GRAPHIC] [TIFF OMITTED] TP24JA07.004

    To ensure that all manufacturers are tracking in-use performance in 
the same manner, we are proposing very detailed requirements for 
defining and incrementing both the numerator and denominator of this 
ratio. Manufacturers would be required to keep track of separate 
numerators and denominators for each of the major monitors, and to 
ensure that the data are saved every time the engine is shut off. The 
numerators and denominators would be reset to zero only in extreme 
circumstances when the non-volatile memory has been cleared (e.g., when 
the onboard computer has been reprogrammed in the field or when the 
onboard computer memory has been corrupted). The values would not be 
reset to zero during normal occurrences such as clearing of stored DTCs 
or performing routine service or maintenance.
    Further, the numerator and denominator would be structured such 
that their maximum values would be 65,535 which is the maximum number 
that can be stored in a 2-byte location. This would ensure that 
manufacturers allocate sufficient and consistent memory space in the 
onboard computer. If either the numerator or denominator for a 
particular monitor reaches the maximum value, both values for that 
particular monitor would be divided by two before counting resumes. In 
general, the numerator and denominator would only be allowed to 
increment a maximum of once per drive cycle because most of the major 
monitors are designed to operate only once per drive cycle. 
Additionally, incrementing of both the numerator and denominator for a 
particular monitor would be disabled (i.e., paused but the stored 
values would not be erased or reset) only when a problem has been 
detected (i.e., a pending or MIL-on DTC has been stored) that prevents 
the monitor from executing. Once the problem is no longer detected and 
any stored DTCs associated with the problem have been erased, either 
through the allowable self-clearing process or upon command by a 
technician via a scan tool, incrementing of both the numerator and 
denominator would resume.
    SAE has developed standards for storing and reporting the data to a 
generic scan tool. This would help ensure that all manufacturers report 
the data in an identical manner which should ease data collection in 
the field.
a. Number of Monitoring Events (``Numerator'')
    For the numerator, manufacturers would be required to keep a 
separate numeric count of how often each of the particular monitors has 
operated. More specifically, manufacturers would have to implement a 
software counter that increments by one every time the particular 
monitor meets all of the enable/monitoring conditions for a long enough 
period of time such that a malfunctioning component would have been 
detected. For example, if a manufacturer requires a vehicle to be 
warmed-up and at idle for 20 seconds continuously to detect a 
malfunctioning catalyst, the catalyst monitor numerator could only be 
incremented if the vehicle actually operates simultaneously in all of 
those conditions. If the vehicle is operated in some but not all of the 
conditions (e.g., at idle but not warmed-up), the numerator would not 
be allowed to increment because the monitor would not have been able to 
detect a malfunctioning catalyst since all of the conditions were not 
satisfied simultaneously.
    Another complication is the difference between a monitor reaching a 
``pass'' or ``fail'' decision. At first glance, it would appear that a 
manufacturer should simply increment the numerator anytime the 
particular monitor reaches a decision, be it ``pass'' or ``fail''. 
However, monitoring strategies may have a different set of criteria 
that must be met to reach a ``pass'' decision versus a ``fail'' 
decision. As a simple example, a manufacturer may appropriately require 
only 10 seconds of operation at idle to reach a ``pass'' decision but 
require 30 seconds of operation at idle to reach a ``fail'' decision. 
Manufacturers would not be allowed to increment the numerator if the 
vehicle had idled for 10 seconds and reached a ``pass'' decision since 
insufficient time had passed to allow for a possible ``fail'' decision. 
This is necessary because the primary function of OBD systems is to 
detect malfunctions (i.e., to correctly reach ``fail'' decisions, not 
``pass'' decisions) and, thus, the real world ability of the monitors 
to detect malfunctions is the parameter we want most to measure. 
Therefore, monitors with different criteria to reach a ``pass'' 
decision versus a ``fail'' decision would not be allowed to increment 
the numerator solely upon satisfying the ``pass'' criteria.
    The correct implementation of the numerator counters by 
manufacturers is imperative to ensure a reliable measure for 
determining real world performance. ``Overcounting'' would falsely 
indicate the monitor is executing more often than it really is, while 
``undercounting'' would make it appear as if the monitor is not running 
as often as it really is. Manufacturers would be required to describe 
their numerator incrementing strategy in their certification 
documentation and to verify the proper performance of their strategy 
during production vehicle evaluation testing.
b. Number of Driving Events (``Denominator'')
    We are also proposing that manufacturers separately track how often 
the engine is operated. Basically, the denominator would be a counter 
that increments by one each time the engine is operated. We are 
proposing that the denominator counter be incremented by one only if 
several criteria are satisfied during a single drive cycle. This allows 
very short trips or trips during extreme conditions such as very cold 
temperatures or very high altitude to be filtered out and excluded from 
the count. This is appropriate because these are also conditions where 
most OBD monitors are neither expected nor required to operate.
    Specifically, the denominator would be incremented if, on a single 
key start, the following criteria were satisfied while ambient 
temperature remained above 20 degrees Fahrenheit and altitude remained 
below 8,000 feet:
     Minimum engine run time of 10 minutes;

[[Page 3241]]

     Minimum of 5 minutes, cumulatively, of operation at 
vehicle speeds greater than 25 miles-per-hour for gasoline engines or 
calculated load greater than 15 percent for diesel engines; and
     At least one continuous idle for a minimum of 30 seconds 
encountered.
    We intend to work with industry to collect data during the first 
few years of implementation and make any adjustments, if necessary, to 
the criteria used to increment the denominator to ensure that the in-
use performance ratio provides a meaningful measure of in-use 
monitoring performance.
2. Proposed Performance Tracking Requirements
a. In-use Monitoring Performance Ratio Definition
    For monitors required to meet the in-use performance tracking 
requirements,\44\ we are proposing that the incrementing of numerators 
and denominators and the calculation of the in-use performance ratio be 
done in accordance with the following specifications.
---------------------------------------------------------------------------

    \44\ These monitors, as presented in section II.A.3, are, for 
diesel engines: the NMHC catalyst, the CDPF system, the 
NOX adsorber system, the NOX converting 
catalyst system, and the boost system; and, for gasoline engines: 
the catalyst, the evaporative system, and the secondary air system; 
and, for all engines, the exhaust gas sensors, the EGR system, and 
the VVT system.
---------------------------------------------------------------------------

    The numerator(s) would be defined as a measure of the number of 
times a vehicle has been operated such that all monitoring conditions 
necessary for a specific monitor to detect a malfunction have been 
encountered. Except for systems using alternative statistical MIL 
illumination protocols, the numerator is to be incremented by an 
integer of one. The numerator(s) may not be incremented more than once 
per drive cycle. The numerator(s) for a specific monitor would be 
incremented within 10 seconds if and only if the following criteria are 
satisfied on a single drive cycle:
     Every monitoring condition necessary for the monitor of 
the specific component to detect a malfunction and store a pending DTC 
has been satisfied, including enable criteria, presence or absence of 
related DTCs, sufficient length of monitoring time, and diagnostic 
executive priority assignments (e.g., diagnostic ``A'' must execute 
prior to diagnostic ``B''). For the purpose of incrementing the 
numerator, satisfying all the monitoring conditions necessary for a 
monitor to determine that the component is passing may not, by itself, 
be sufficient to meet this criteria.
     For monitors that require multiple stages or events in a 
single drive cycle to detect a malfunction, every monitoring condition 
necessary for all events to have completed must be satisfied.
     For monitors that require intrusive operation of 
components to detect a malfunction, a manufacturer would be required to 
request Administrator approval of the strategy used to determine that, 
had a malfunction been present, the monitor would have detected the 
malfunction. Administrator approval of the request would be based on 
the equivalence of the strategy to actual intrusive operation and the 
ability of the strategy to determine accurately if every monitoring 
condition was satisfied as necessary for the intrusive event to occur.
     For the secondary air system monitor, the three criteria 
above are satisfied during normal operation of the secondary air 
system. Monitoring during intrusive operation of the secondary air 
system later in the same drive cycle solely for the purpose of 
monitoring may not, by itself, be sufficient to meet these criteria.
    The third bullet item above requires explanation. There may be 
monitors, and there have been monitors in light-duty, designed to use 
what could be termed a two stage or two step process. The first step is 
usually a passive and/or short evaluation that can be used to ``pass'' 
a properly working component where ``pass'' refers to evaluating the 
component and determining that it is not malfunctioning. The second 
step is usually an intrusive and/or longer evaluation that is necessary 
to ``fail'' a malfunctioning component or ``pass'' a component nearing 
the point of failure. An example of such an approach might be an 
evaporative leak detection monitor that uses an intrusive vacuum pull-
down/bleed-up evaluation during highway cruise conditions. If the 
evaporative system is sealed tight, the monitor ``passes'' and is done 
with testing for the given drive cycle. If the monitor senses a leak 
close to the required detection limit, the monitor does not ``pass'' 
and an internal flag is stored that will trigger the second stage of 
the test during the next cold start when a more accurate evaluation can 
be conducted. On the next cold start, provided the internal flag is 
set, an intrusive vacuum pull-down/bleed up monitor might be conducted 
during engine idle a very short time after the cold start. This second 
evaluation stage, being at idle and cold, gives a more accurate 
indication of the evaporative system's integrity and provides for a 
more accurate decision regarding the presence and size of a leak.
    In this example, the second stage of this monitor would run less 
frequently in real use than the first stage since it is activated only 
on those occasions where the first stage suggests that a leak may be 
present (which most cars will not have). The rate-based tracking 
requirements are meant to give a measure of how often a monitor could 
detect a malfunction. To know the right answer, we need to know how 
often the first stage is running and could ``fail'', thus triggering 
the second stage, and then how often the second stage is completing. If 
we track only the first stage, we would get a false indication of how 
often the monitor could really detect a leak. But, if we track only the 
second stage, most cars would never increment the counter since most 
cars do not have leaks and would not trigger stage two.
    In considering this, we see two possible solutions: (1) Always 
activate the second stage evaluation in which case there would be an 
intrusive monitor being performed that does not really need to be 
performed; or, (2) implement a ``ghost'' monitor that pretends that the 
first stage evaluation triggers the second stage evaluation and then 
also looks for when the second stage evaluation could have completed 
had it been necessary. The third bullet item in the list above requires 
that, if a manufacturer intends to implement a two stage monitor and 
intends to implement such a ``ghost'' monitor as described here for 
rate based tracking, approval must be sought for doing so to make sure 
we agree that you are doing it correctly and properly.
    For monitors that can generate results in a ``gray zone'' or ``non-
detection zone'' (i.e., results that indicate neither a passing system 
nor a malfunctioning system) or in a ``non-decision zone'' (e.g., 
monitors that increment and decrement counters until a pass or fail 
threshold is reached), the manufacturer would be responsible for 
incrementing the numerator appropriately. In general, the numerator 
should not be incremented when the monitor indicates a result in the 
``non-detection zone'' or prior to the monitor reaching a decision. 
When necessary, the Administrator would consider data and/or 
engineering analyses submitted by the manufacturer demonstrating the 
expected frequency of results in the ``non-detection zone'' and the 
ability of the monitor to determine accurately, had an actual 
malfunction been present, whether or not the monitor would have 
detected a malfunction instead of a result in the ``non-detection 
zone.''

[[Page 3242]]

    For monitors that run or complete their evaluation with the engine 
off, the numerator must be incremented either within 10 seconds of the 
monitor completing its evaluation in the engine off state, or during 
the first 10 seconds of engine start on the subsequent drive cycle.
    Manufacturers using alternative statistical MIL illumination 
protocols for any of the monitors that require a numerator would be 
required to increment the numerator(s) appropriately. The manufacturer 
may be required to provide supporting data and/or engineering analyses 
demonstrating both the equivalence of their incrementing approach to 
the incrementing specified above for monitors using the standard MIL 
illumination protocol, and the overall equivalence of their 
incrementing approach in determining that the minimum acceptable in-use 
performance ratio has been satisfied.
    Regarding the denominator(s), defined as a measure of the number of 
times a vehicle has been operated, we are proposing that it also be 
incremented by an integer of one. The denominator(s) may not be 
incremented more than once per drive cycle. The general denominator and 
the denominators for each monitor would be incremented within 10 
seconds if and only if the following criteria are satisfied on a single 
drive cycle during which ambient temperature remained at or above 20 
degrees Fahrenheit and altitude remained below 8,000 feet:
     Cumulative time since the start of the drive cycle is 
greater than or equal to 600 seconds (10 minutes);
     Cumulative gasoline engine operation at or above 25 miles 
per hour or diesel engine operation at or above 15 percent calculated 
load, either of which occurs for greater than or equal to 300 seconds 
(5 minutes); and
     Continuous engine operation at idle (e.g., accelerator 
pedal released by driver and vehicle speed less than or equal to one 
mile per hour) for greater than or equal to 30 seconds.
    In addition to the requirements above, the evaporative system 
monitor denominator(s) must be incremented if and only if:
     Cumulative time since the start of the drive cycle is 
greater than or equal to 600 seconds (10 minutes) while at an ambient 
temperature of greater than or equal to 40 degrees Fahrenheit but less 
than or equal to 95 degrees Fahrenheit; and
     Engine cold start occurs with engine coolant temperature 
at engine start greater than or equal to 40 degrees Fahrenheit but less 
than or equal to 95 degrees Fahrenheit and less than or equal to 12 
degrees Fahrenheit higher than ambient temperature at engine start.
    In addition to the requirements above, the denominator(s) for the 
following monitors must be incremented if and only if the component or 
strategy is commanded ``on'' for a time greater than or equal to 10 
seconds:
     Gasoline secondary air system;
     Cold start emission reduction strategy;
     Components or systems that operate only at engine start-up 
(e.g., glow plugs, intake air heaters) and are subject to monitoring 
under ``other emission control systems'' (section II.D.5) or 
comprehensive component output components (see section II.D.4).
    For purposes of determining this commanded ``on'' time, the OBD 
system may not include time during intrusive operation of any of the 
components or strategies later in the same drive cycle solely for the 
purposes of monitoring.
    In addition to the requirements above, the denominator(s) for the 
monitors of the following output components (except those operated only 
at engine start-up as outlined above) must be incremented if and only 
if the component is commanded to function (e.g., commanded ``on'', 
``open'', ``closed'', ``locked'') two or more times during the drive 
cycle or for a time greater than or equal to 10 seconds, whichever 
occurs first:
     Variable valve timing and/or control system
     ``Other emission control systems''
     Comprehensive component (output component only, e.g., 
turbocharger waste-gates, variable length manifold runners)
    For monitors of the following components, the manufacturer may use 
alternative or additional criteria to that set forth above for 
incrementing the denominator. To do so, the manufacturer would need to 
be able to demonstrate that the criteria would be equivalent to the 
criteria outlined above at measuring the frequency of monitor operation 
relative to the amount of engine operation:
     Engine cooling system input components (section II.D.2)
     ``Other emission control systems'' (section II.D.5)
     Comprehensive component input components that require 
extended monitoring evaluation (section II.D.4, e.g., stuck fuel level 
sensor rationality)
    For monitors of the following components or other emission controls 
that experience infrequent regeneration events, the manufacturer may 
use alternative or additional criteria to that set forth above for 
incrementing the denominator. To do so, the manufacturer would need to 
demonstrate that the criteria would be equivalent to the criteria 
outlined above at measuring the frequency of monitor operation relative 
to the amount of engine operation:
     Oxidation catalysts
     Diesel particulate filters
    For hybrid engine systems, engines that employ alternative engine 
start hardware or strategies (e.g., integrated starter and generators), 
or alternative fueled engines (e.g., dedicated, bi-fuel, or dual-fuel 
applications), the manufacturer may request Administrator approval to 
use alternative criteria to that set forth above for incrementing the 
denominator. In general, approval would not be given for alternative 
criteria that only employ engine shut off at or near idle/vehicle 
stationary conditions. Approval of the alternative criteria would be 
based on the equivalence of the alternative criteria at determining the 
amount of engine operation relative to the measure of conventional 
engine operation in accordance with the criteria above.
    The numerators and denominators may need to be disabled at some 
times. To do this, within 10 seconds of a malfunction being detected 
(i.e., a pending, MIL-on, or active DTC being stored) that disables a 
monitor required to meet the performance tracking requirements,\45\ the 
OBD system must disable further incrementing of the corresponding 
numerator and denominator for each monitor that is disabled. When the 
malfunction is no longer detected (e.g., the pending DTC is erased 
through self-clearing or through a scan tool command), incrementing of 
all corresponding numerators and denominators should resume within 10 
seconds. Also, within 10 seconds of the start of a power takeoff unit 
(PTO) that disables a monitor required to meet the performance tracking 
requirements, the OBD system should disable further incrementing of the 
corresponding numerator and denominator for each monitor that is 
disabled. When the PTO operation ends, incrementing of all 
corresponding numerators and denominators should resume within 10 
seconds. The OBD system must disable further incrementing of all 
numerators

[[Page 3243]]

and denominators within 10 seconds if a malfunction has been detected 
in any component used to determine if: vehicle speed/calculated load; 
ambient temperature; elevation; idle operation; engine cold start; or, 
time of operation has been satisfied, and the corresponding pending DTC 
has been stored. Incrementing of all numerators and denominators should 
resume within 10 seconds when the malfunction is no longer present 
(e.g., pending DTC erased through self-clearing or by a scan tool 
command).
---------------------------------------------------------------------------

    \45\ These monitors, as presented in section II.A.3, are, for 
diesel engines: the NMHC catalyst, the CDPF system, the 
NOX adsorber system, the NOX converting 
catalyst system, and the boost system; and, for gasoline engines: 
the catalyst, the evaporative system, and the secondary air system; 
and, for all engines, the exhaust gas sensors, the EGR system, and 
the VVT system.
---------------------------------------------------------------------------

    The in-use performance monitoring ratio itself is defined as the 
numerator for the given monitor divided by the denominator for that 
monitor.
b. Standardized Tracking and Reporting of Monitor Performance
    We are proposing that the OBD system separately report an in-use 
monitor performance numerator and denominator for each of the following 
components:
     For diesel engines: NMHC catalyst bank 1, NMHC catalyst 
bank 2, NOX catalyst bank 1, NOX catalyst bank 2, 
exhaust gas sensor bank 1, exhaust gas sensor bank 2, EGR/VVT system, 
DPF system, turbo boost control system, and the NOX 
adsorber. The OBD system must also report a general denominator and an 
ignition cycle counter in the standardized format discussed below and 
in section II.F.5.
     For gasoline engines: catalyst bank 1, catalyst bank 2, 
oxygen sensor bank 1, oxygen sensor bank 2, evaporative leak detection 
system, EGR/VVT system, and secondary air system. The OBD system must 
also report a general denominator and an ignition cycle counter in the 
standardized format specified below and in section II.F.5.
    The OBD system would be required to report a separate numerator for 
each of the components listed in the above bullet lists. For specific 
components or systems that have multiple monitors that are required to 
be reported under section II.B--e.g., exhaust gas sensor bank 1 may 
have multiple monitors for sensor response or other sensor 
characteristics--the OBD system should separately track numerators and 
denominators for each of the specific monitors and report only the 
corresponding numerator and denominator for the specific monitor that 
has the lowest numerical ratio. If two or more specific monitors have 
identical ratios, the corresponding numerator and denominator for the 
specific monitor that has the highest denominator should be reported 
for the specific component. The numerator(s) must be reported in 
accordance with the specifications in section II.F.5.
    The OBD system would also be required to report a separate 
denominator for each of the components listed in the above bullet 
lists. The denominator(s) must be reported in accordance with the 
specifications in section II.F.5.
    Similarly, for the in-use performance ratio, determining which 
corresponding numerator and denominator to report as required for 
specific components or systems that have multiple monitors that are 
required to be reported--e.g., exhaust gas sensor bank 1 may have 
multiple monitors for sensor response or other sensor 
characteristics'the ratio should be calculated in accordance with the 
specifications in section II.F.5.
    The ignition cycle counter is defined as a counter that indicates 
the number of ignition cycles a vehicle has experienced. The ignition 
cycle counter must also be reported in accordance with the 
specifications in section II.F.5. The ignition cycle counter, when 
incremented, should be incremented by an integer of one. The ignition 
cycle counter may not be incremented more than once per ignition cycle. 
The ignition cycle counter should be incremented within 10 seconds if 
and only if the engine exceeds an engine speed of 50 to 150 rpm below 
the normal, warmed-up idle speed (as determined in the drive position 
for vehicles equipped with an automatic transmission) for at least two 
seconds plus or minus one second. The OBD system should disable further 
incrementing of the ignition cycle counter within 10 seconds if a 
malfunction has been detected in any component used to determine if 
engine speed or time of operation has been satisfied and the 
corresponding pending DTC has been stored. The ignition cycle counter 
may not be disabled from incrementing for any other condition. 
Incrementing of the ignition cycle counter should resume within 10 
seconds after the malfunction is no longer present (e.g., pending DTC 
erased through self-clearing or by a scan tool command).

F. Standardization Requirements

    The heavy-duty OBD regulation would include requirements for 
manufacturers to standardize certain features of the OBD system. 
Effective standardization assists all repair technicians in diagnosing 
and repairing malfunctions by providing equal access to essential 
repair information, and requires structuring the information in a 
common format from manufacturer to manufacturer. Additionally, the 
standardization would help to facilitate the potential use of OBD 
checks in heavy-duty inspection and maintenance programs.
    Among the features that would be standardized under the proposed 
heavy-duty OBD regulation include:
     The diagnostic connector, the computer communication 
protocol;
     The hardware and software specifications for tools used by 
service technicians;
     The information communicated by the onboard computer and 
the methods for accessing that information;
     The numeric designation of the DTCs stored when a 
malfunction is detected; and,
     The terminology used by manufacturers in their service 
manuals.
    Our proposal would require that only a certain minimum set of 
emissions-related information be made available through the 
standardized format, protocol, and connector. We are not limiting 
engine manufacturers as to what protocol they use for engine control, 
communication between onboard computers, or communication to 
manufacturer-specific scan tools or test equipment. Further, we are not 
prohibiting engine manufacturers from equipping the vehicle with 
additional diagnostic connectors or protocols as required by other 
suppliers or purchasers. For example, fleets that use data logging or 
other equipment that requires the use of SAE J1587 communication and 
connectors could still be installed and supported by the engine and 
vehicle manufacturers. The OBD rules would only require that engine 
manufacturers also equip their vehicles with a specific connector and 
communication protocol that meet the standardized requirements to 
communicate a minimum set of emissions-related diagnostic, service and, 
potentially, inspection information.
    Additionally, our proposal includes a phase-in of one engine family 
meeting the requirements of OBD in the model years 2010 through 2012. 
Because non-compliant engines would not require the proposed 
standardization features, truck and coach builders could be faced with 
several integration issues when building product in 2010 through 2012. 
Specifically, they could be faced with designing their vehicles to 
accommodate a standardized MIL, diagnostic connector, and communication 
protocol when using a compliant engine yet to not accommodate those 
features when using a non-compliant engine. This outcome could easily 
arise since only one engine-family per manufacturer would be compliant 
and, therefore, a given truck

[[Page 3244]]

designed to accommodate several engines from several engine 
manufacturers would very likely need to accommodate a compliant engine 
from manufacturer A and a non-compliant engine from manufacturer B. It 
should be noted that engine choices are typically driven by the end 
user--the truck buyer--and not by the truck or coach builder. For that 
reason, the truck builder must accommodate all possible engines for the 
truck size and cannot necessarily demand from the engine manufacturer a 
compliant versus a non-compliant engine.
    As a result, rather than force truck and coach builders to 
accommodate two different systems and risk incompatibilities, we are 
proposing to exempt the 2010 through 2012 model year engines from 
meeting certain standardization requirements of OBD. This should allow 
truck and coach builders to integrate engines in the same manner as 
done currently and then to switch over to integrating a single system 
in 2013 when all engines are required to meet all of the 
standardization requirements of OBD. The proposed implementation 
schedule for standardization features is shown in Table II.G-2.
1. Reference Documents
    We are proposing that OBD systems comply with the following 
provisions laid out in the following Society of Automotive Engineers 
(SAE) and/or International Organization of Standards (ISO) documents 
that are or would be incorporated by reference (IBR) into federal 
regulation:

                          Table II.F--1. Reference Documents for Over 14,000 Pound OBD
----------------------------------------------------------------------------------------------------------------
       Document No.                    Document title                    Date                   Comment
----------------------------------------------------------------------------------------------------------------
SAE J1962.................  ``Diagnostic Connector--Equivalent   April 2002..........  Updated IBR.
                             to ISO/DIS 15031-3: December 14,
                             2001''.
SAE J1930.................  ``Electrical/Electronic Systems      April 2002..........  Updated IBR.
                             Diagnostic Terms, Definitions,
                             Abbreviations, and Acronyms--
                             Equivalent to ISO/TR 15031-2:
                             April 30, 2002''.
SAE J1978.................  ``OBD II Scan Tool--Equivalent to    April 2002..........  Updated IBR.
                             ISO/DIS 15031-4: December 14,
                             2001''.
SAE J1979.................  ``E/E Diagnostic Test Modes--        April 2002..........   Updated IBR.
                             Equivalent to ISO/DIS 15031-5:
                             April 30, 2002''.
SAE J2012.................  ``Diagnostic Trouble Code            April 2002..........  Updated IBR.
                             Definitions--Equivalent to ISO/DIS
                             15031-6: April 30, 2002''.
SAE J1939.................  ``Recommended Practice for a Serial  2005 Edition, March   Updated IBR.
                             Control and Communications Vehicle   2005.
                             Network,'' and the associated
                             subparts included in SAE HS-1939,
                             ``Truck and Bus Control and
                             Communications Network Standards
                             Manual''.
SAE J2403.................  ``Medium/Heavy-Duty E/E Systems      August 2004.........  New IBR.
                             Diagnosis Nomenclature''.
SAE J2534.................  ``Recommended Practice for Pass-     February 2002.......  New IBR.
                             Thru Vehicle Reprogramming''.
ISO 15765-4:2001..........  ``Road Vehicles--Diagnostics on      December 2001.......  New IBR.
                             Controller Area Network (CAN)--
                             Part 4: Requirements for emission-
                             related systems''.
----------------------------------------------------------------------------------------------------------------

    Copies of these SAE materials may be obtained from Society of 
Automotive Engineers International, 400 Commonwealth Dr., Warrendale, 
PA, 15096-0001. Copies of these ISO materials may be obtained from the 
International Organization for Standardization, Case Postale 56, CH-
1211 Geneva 20, Switzerland.
2. Diagnostic Connector Requirements
    We are proposing that a standard data link connector conforming to 
either SAE J1962 or SAE J1939-13 specifications (except as noted below) 
would have to be included in each vehicle. The connector would have to 
be located in the driver's side foot-well region of the vehicle 
interior in the area bound by the driver's side of the vehicle and the 
driver's side edge of the center console (or the vehicle centerline if 
the vehicle does not have a center console) and at a location no higher 
than the bottom of the steering wheel when in the lowest adjustable 
position. The Administrator would not allow the connector to be located 
on or in the center console (i.e., neither on the horizontal faces near 
the floor-mounted gear selector, parking brake lever, or cup-holders, 
nor on the vertical faces near the car stereo, climate system, or 
navigation system controls). The location of the connector must be 
easily identifiable and accessed (e.g., to connect an off-board tool). 
For vehicles equipped with a driver's side door, the connector would 
have to be easily identified and accessed by someone standing (or 
``crouched'') on the ground outside the driver's side of the vehicle 
with the driver's side door open.
    If a manufacturer wants to cover the connector, the cover must be 
removable by hand without the use of any tools and be labeled ``OBD'' 
to aid technicians in identifying the location of the connector. Access 
to the diagnostic connector could not require opening or removing any 
storage accessory (e.g., ashtray, coinbox). The label would have to 
clearly identify that the connector is located behind the cover and is 
consistent with language and/or symbols commonly used in the automobile 
and/or heavy truck industry.
    If the ISO 15765-4 protocol (see section II.F.3) is used for the 
required OBD standardized functions, the connector would have to meet 
the ``Type A'' specifications of SAE J1962. Any pins in the connector 
that provide electrical power must be properly fused to protect the 
integrity and usefulness of the connector for diagnostic purposes and 
may not exceed 20.0 Volts DC regardless of the nominal vehicle system 
or battery voltage (e.g., 12V, 24V, 42V).
    If the SAE J1939 protocol (see section II.F.3)) is used for the 
required OBD standardized functions, the connector must meet the 
specifications of SAE J1939-13. Any pins in the connector that provide 
electrical power must be properly fused to protect the integrity and 
usefulness of the connector for diagnostic purposes.
    Manufacturers would be allowed to equip engines/vehicles with 
additional diagnostic connectors for manufacturer-specific purposes 
(i.e., purposes other than the required OBD functions). However, if the 
additional connector conforms to the ``Type A'' specifications of SAE 
J1962 or the specifications of SAE J1939-13 and is located in the 
vehicle interior near the required connector as described above, the 
connector(s) must be clearly labeled to identify which connector is 
used to access the standardized OBD information proposed below.

[[Page 3245]]

3. Communications to a Scan Tool
a. Background
    In light-duty OBD, manufacturers are allowed to use one of four 
protocols for communication between a generic scan tool and the 
vehicle's onboard computer. A generic scan tool automatically cycles 
through each of the allowable protocols until it hits upon the proper 
one with which to establish communication with the particular onboard 
computer. While this has generally worked successfully in the field, 
some communication problems have arisen.
    In an effort to address these problems, CARB has made recent 
changes to their light-duty OBD II regulation that require all light-
duty vehicle manufacturers to use only one communication protocol by 
the 2008 model year. In making these changes, CARB staff argued that 
their experience with standardization under the OBD II regulation 
showed that having a single set of standards used by all vehicles would 
be desirable. CARB staff argued that a single protocol offers a 
tremendous benefit to both scan tool designers and service technicians. 
Scan tool designers could focus on added feature content and could 
expend much less time and money validating basic functionality of their 
product on all the various permutations of protocol interpretations 
that are implemented. In turn, technicians would likely get a scan tool 
that works properly on all vehicles without the need for repeated 
software updates that incorporate ``work-arounds'' or other patches to 
fix bugs or adapt the tool to accommodate slight variances in how the 
multiple protocols interact with each other or are implemented by 
various manufacturers. Further, a single protocol should also be 
beneficial to fleet operators that use add-on equipment such as data 
loggers, and for vehicle manufacturers that integrate parts from 
various engine and component suppliers all of which must work together.
    Based on our similar experiences at the federal level with 
communication protocols giving rise to service and inspection/
maintenance program issues, we initially wanted to propose a single 
communication protocol for engines used in over 14,000 pound vehicles. 
However, the affected industry has been divided over which single 
protocol should be required and has strongly argued for more than one 
protocol to be allowed. Therefore, for vehicles with diesel engines, we 
are proposing that manufacturers be required to use either the 
standards set forth in SAE J1939, or those set forth in the 500 kbps 
baud rate version of ISO 15765. For vehicles with gasoline engines, we 
are proposing that manufacturers be required to use the 500 kbps baud 
rate version of ISO 15765. Manufacturers would be required to use only 
one standard to meet all the standardization requirements on a single 
vehicle; that is, a vehicle must use only one protocol for all OBD 
modules on the vehicle.
    Several in the heavy-duty industry have argued for options that 
would allow the use of more than these two protocols on heavy-duty 
engines. Some have even argued for combinations of these protocols--
e.g., diagnostic connector and messages of ISO 15765 on an SAE J1939 
physical layer network. However, as described above, experience from 
multiple protocols and multiple variants within the protocols has 
unnecessarily caused a significant number of problems with engine and 
vehicle related computer communications.
b. Requirements for Communications to a Scan Tool
    We are proposing that all OBD control modules--e.g., engine, 
auxiliary emission control module--on a single vehicle be required to 
use the same protocol for communication of required emissions-related 
messages from onboard to off-board network communications to a scan 
tool meeting SAE J1978 specifications or designed to communicate with a 
SAE J1939 network. Engine manufacturers would not be allowed to alter 
normal operation of the engine emissions control system due to the 
presence of off-board test equipment accessing the OBD information 
proposed below. The OBD system would be required to use one of the 
following standardized protocols:
     ISO 15765-4 and all required emission-related messages 
using this protocol would have to use a 500 kbps baud rate.
     SAE J1939 which may only be used on vehicles with diesel 
engines.
4. Required Emissions Related Functions
    Most of the proposed emissions related functions are elements that 
exist in our light-duty OBD requirements. We are proposing several 
required functions, these are:
     Readiness status
     Distance and number of warm-up cycles since DTC clear
     Permanent DTC storage
     Real time indication of monitor status
     Communicating readiness status to the vehicle operator
     Diagnostic trouble codes (DTC)
     Data stream
     Freeze frame
     Test results
     Software calibration identification
     Software calibration verification number
     Vehicle identification number (VIN)
i. Readiness Status
    The main intent of readiness status is to ensure that a vehicle is 
ready for an OBD-based inspection--by indicating that monitors have run 
and operational status of the emissions-control system has been fully 
evaluated--and to prevent fraudulent testing in inspection programs. In 
general, for OBD-based inspections, technicians ``fail'' a vehicle with 
an illuminated MIL since this would indicate the presence of an 
emissions control system malfunction. Without the readiness status 
indicators, technicians would not have a clear indication from the OBD 
system that it had sufficiently evaluated the emissions control system 
prior to the inspection. Since the potential exists for OBD checks to 
be used as part of a heavy truck inspection program, we believe that 
having readiness status indicators as part of this proposal is 
important--waiting for a subsequent OBD-I/M rulemaking to require such 
indicators would unnecessarily delay implementation of such OBD-I/M 
programs.
    Absent such OBD-I/M programs, we still believe that readiness 
indicators are an important OBD tool. Technicians would be expected to 
use the readiness status to verify OBD-related repairs. Specifically, 
technicians would clear the computer memory after repairing an OBD-
detected fault in order to erase the DTC, extinguish the MIL, and reset 
the readiness status to ``incomplete.'' Then the vehicle could be 
operated in such a manner that the monitor of the repaired component 
would run (i.e., the readiness status of the monitor would be set to 
``complete''). The absence of any DTCs or MIL illumination upon 
readiness status indicating ``complete'' would indicate a successful 
repair.
    Therefore, we are proposing that manufacturers be required to 
indicate the readiness status of the OBD monitors. This would serve to 
indicate whether or not engine operation has been sufficient to allow 
certain OBD monitors to perform their system evaluations. The OBD 
system would be required to report a readiness status of either 
``complete'' if the monitor has run a sufficient number of times to 
detect a malfunction since computer memory was last cleared, 
``incomplete'' if the monitor has not yet run a sufficient number of 
times since the memory was last cleared, or ``not applicable'' if the

[[Page 3246]]

monitor is not present or if the specific monitored component is not 
equipped on the vehicle. The readiness status of monitors that are 
required to run continuously would always indicate ``complete.'' The 
details of the proposal discussed below clarify that the readiness 
status would be set to ``incomplete'' whenever memory is cleared either 
by a battery disconnect or by a scan tool but not after a normal 
vehicle shutdown (i.e., key-off).
ii. Distance Traveled and Number of Warm-Up Cycles Since DTC Clear
    As originally envisioned in our OBD-I/M rulemaking (61 FR 40940), 
we intended to require that all readiness status indicators be set to 
``complete'' prior to accepting a vehicle for I/M inspection. However, 
it became clear that some vehicles were being rejected from inspection 
for reasons beyond the driver's control. For example, a vehicle driven 
in extreme ambient conditions would prohibit monitors from running and 
setting readiness status indicators to ``complete.'' Also, a vehicle 
repaired just prior to arriving at the inspection station may not have 
been operated sufficiently to set the readiness status of the monitor 
for the recently repaired component to ``complete.'' The driver of such 
a vehicle would, in essence, be punished unintentionally for having 
taken the time and expense to repair the vehicle just prior to the 
inspection. As a result, we issued guidance (cite) to state inspectors 
recommending that vehicles be accepted for I/M inspection provided two 
or fewer readiness status indicators are ``incomplete.'' Note that most 
light-duty gasoline vehicles--the bulk of the vehicle fleet facing OBD-
I/M checks--have only four monitors for which the readiness status 
indicator is meaningful (all of their other monitors being continuous 
monitors). However, there exists evidence that this policy is perhaps 
accepting vehicles for I/M inspection that should not be accepted due 
to unscrupulous clearing of DTCs and readiness status by people that 
understand how to do so and then operate their vehicles just enough to 
set the required minimum number of readiness indicators to 
``complete.''
    As a result, we are proposing some additional features that should 
better differentiate between vehicles that have been repaired recently 
or have ``incomplete'' readiness indicators through circumstances 
outside the driver's control, and those vehicles operated by drivers 
that are attempting to fraudulently get through an OBD-based 
inspection. We are proposing that the OBD system make available data 
that would report the distance traveled or engine run time for those 
engines that do not use vehicle speed information, and the number of 
warm-up cycles since the fault memory was last cleared.\46\ By 
combining these data with the readiness data, technicians or inspectors 
would better be able to determine if ``incomplete'' readiness status 
indicators or an extinguished MIL are due to unscrupulous memory 
clearing or circumstances beyond the driver's control. For example, a 
vehicle with several ``incomplete'' readiness indicators but with a 
high distance traveled/engine run time and a high number of warm-up 
cycles since the last clearing of fault memory would be unlikely to 
have undergone a recent fault memory clearing for the purpose of 
extinguishing the MIL prior to inspection. On the other hand, a vehicle 
with only one or two ``incomplete'' readiness indicators and a very low 
distance traveled/engine run time and a low number of warm-up cycles 
since fault memory clearing should probably be rejected or failed at an 
inspection. This would better allow an inspection program to be set up 
to reject only those vehicles with recently cleared memories while 
minimizing the chances of rejecting vehicles that driven such that 
monitors rarely run whether by unique driver behaviors or extreme 
ambient conditions.
---------------------------------------------------------------------------

    \46\ The fault memory being any DTCs, readiness status 
indicators, freeze frame information, etc.
---------------------------------------------------------------------------

iii. Permanent Diagnostic Trouble Code Storage
    Consistent with the proposal for distance traveled/engine run time 
and number of warm-up cycles, we are proposing a requirement to make it 
much more difficult for a vehicle owner or technician to clear the 
fault memory and erase all traces of a previously detected malfunction. 
Current OBD systems on under 14,000 pound vehicles allow a technician 
or vehicle owner to erase all DTCs and extinguish the MIL by issuing a 
command from a generic scan tool or, in many cases, simply by 
disconnecting the vehicle battery. This would set to ``incomplete'' the 
readiness status indicators for all monitors and would remove all 
record of the malfunction that had been detected.
    We are proposing that manufacturers be required to store in non-
volatile memory random access memory (NVRAM) a minimum of four MIL-on 
DTCs that are, at present, commanding the MIL-on. These ``permanent'' 
DTCs would have to be stored in NVRAM at the end of every key cycle. By 
requiring these permanent DTCs to be stored in NVRAM, one would not be 
able to erase them simply by disconnecting the battery. Further, 
manufacturers would not be allowed to design their OBD systems such 
that these permanent DTCs could be erased by any generic or 
manufacturer-specific scan tool command. Instead, the permanent DTCs 
could be erased only via an OBD system self-clearing--i.e., upon 
evaluating the component or system for which the permanent DTC has been 
stored and detecting on sufficient drive cycles that the malfunction is 
no longer present, the OBD system would erase the fault memory as 
discussed in section II.A.2. Once this has occurred, the permanent DTC 
stored in NVRAM would be erased also.
    The permanent DTCs should help if states choose to implement OBD-
based I/M programs for heavy trucks. A truck with readiness status 
indicators for EGR and boost control set to ``incomplete'' and with 
permanent DTCs stored for both EGR and boost control would quite 
probably be a truck that should be rejected from inspection. The OBD 
system on such a truck has almost certainly had its fault memory 
cleared--via scan tool command or battery disconnect--which would set 
the readiness indicators to ``incomplete'' and erase all MIL-on DTCs 
but would still have permanent DTCs stored (only the OBD system itself 
can erase permanent DTCs). Likewise, a truck with the same readiness 
indicators set to ``incomplete'' and no permanent DTCs for those 
monitors should almost certainly be accepted for inspection since the 
lack of readiness is almost certainly due to circumstances outside the 
driver's control.
    We believe that the permanent DTCs also provide advantages to 
technicians attempting to repair a malfunction and prepare it for 
subsequent inspection or proof of correction. The permanent DTC would 
identify the specific monitor that would need to be exercised after 
repair and prior to inspection to be sure that the malfunction has been 
repaired. By combining this information with the vehicle manufacturer's 
service information, technicians could identify the exact conditions 
necessary to exercise the particular monitor. As such, technicians 
could more effectively verify that the specific monitor (that monitor 
having illuminated the MIL for which the repair has been done) has run 
and confirmed that the malfunction no longer exists and the repair has 
been made correctly. This should also reduce vehicle owner ``come-
backs'' for incomplete or ineffective repairs.

[[Page 3247]]

iv. Real Time Indication of Monitor Status
    We are also proposing provisions to make it easier for technicians 
to prepare a vehicle for an inspection following a repair. These 
provisions would require that the OBD system provide real time data 
that indicate whether the necessary conditions are present currently to 
set all of the readiness indicators to ``complete.'' These data would 
indicate whether a particular monitor may still have an opportunity to 
run on the current drive cycle or whether a condition has been 
encountered that has disabled the monitor for the rest of the drive 
cycle regardless of the driving conditions that might be encountered. 
While these data would not provide technicians with the exact 
conditions necessary to exercise the monitors (only service information 
would provide such information), the date in combination with the 
service information should assist technicians in verifying repairs and/
or preparing a vehicle for inspection. Technicians would be able to use 
this information to identify when specific monitors have indeed 
completed or to identify situations where they have overlooked one or 
more of the enable criteria and need to check the service information 
and try again.
v. Communicating Readiness Status to the Vehicle Operator
    As mentioned above, substantial feedback has been received from 
OBD-based I/M programs throughout the U.S. Much of this feedback 
pertains to the effect on vehicle owners caused by being rejected from 
I/M inspection due to ``incomplete'' readiness status indicators. To 
address this, some light-duty vehicle manufacturers requested that they 
be allowed to communicate the vehicle's readiness status to the vehicle 
owner directly without need of a scan tool. This would provide 
assurance to the vehicle owner that their vehicle is ready for 
inspection prior to taking the vehicle to the I/M station. We are 
proposing that heavy-duty engine manufacturers be allowed to do the 
same thing (this is a proposed option, not a proposed requirement). If 
a manufacturer chooses to implement this option, though, they would be 
required to do so in a standardized manner. On engines equipped with 
this option, the owner would be able to initiate a self-check of the 
readiness status, thereby greatly reducing the possibility of being 
rejected at a roadside inspection.
vi. Diagnostic Trouble Codes (DTC)
    Malfunctions are reported by the OBD system and displayed on a scan 
tool for service technicians in the form of diagnostic trouble codes 
(DTCs). We are proposing that manufacturers be required to report all 
emissions-related DTCs using a standardized format and to make them 
accessible to all service technicians, including the independent 
service industry. The reference document standards selected by the 
manufacturer would define many generic DTCs to be used by all 
manufacturers. In the rare circumstances that a manufacturer cannot 
find within the reference documents a suitable DTC, a unique 
``manufacturer-specific'' DTC could be used. However, such 
manufacturer-specific DTCs are not as easily interpreted by the 
independent service industry. Excessive use of manufacturer-specific 
DTCs may increase the time and cost for vehicle repairs. Thus, we are 
proposing to restrict the use of manufacturer-specific DTCs. If a 
generic DTC suitable for a given malfunction cannot be found, the 
manufacturer would be expected to pursue approval and addition of 
appropriate generic DTCs into the reference documents; the intent being 
to standardize as much information as possible.
    Additionally, we are proposing that the OBD system store DTCs that 
are as specific as possible to identify the nature of the malfunction. 
The intent being to provide service technicians with as detailed 
information as possible to diagnose and repair vehicles in an efficient 
manner. In other words, manufacturers should use separate DTCs for 
every monitor where the monitor and repair procedure, or likely cause 
of the failure, is different. Generally, a manufacturer would design an 
OBD monitor that detects different root causes (e.g., sensor shorted to 
ground or battery) for a malfunctioning component or system. We would 
expect manufacturers to store a specific DTC such as ``sensor circuit 
high input'' or ``sensor circuit low input'' rather than a general code 
such as ``sensor circuit malfunction.'' Further, we expect 
manufacturers to store different DTCs that distinguish circuit 
malfunctions from rationality and functional malfunctions since the 
root cause for each is different and, thus, the repair procedures may 
be different.
    We are also proposing specific provisions for storage of pending 
and MIL-on DTCs. These proposed provisions were discussed in section 
II.A.2.
    We are also proposing requirements that would help to distinguish 
between DTCs stored for malfunctions that are currently present and for 
malfunctions that are no longer present. These requirements would apply 
only to those engines using ISO 15765-4 as the communication protocol. 
As described in section II.A.2, the OBD system would generally 
extinguish the MIL if the malfunction responsible for the MIL 
illumination has not been detected (i.e., the monitor runs and 
determines that the malfunction no longer exists) on three subsequent 
sequential drive cycles. However, a manufacturer would not be allowed 
to erase the associated MIL-on DTC until 40 engine warm-up cycles have 
occurred without again detecting the malfunction. So even though the 
malfunction is no longer present and a MIL-on is not being commanded, 
the DTC would still remain (termed a ``history'' code in the ISO 
standard). Consequently, if another unrelated malfunction occurs and 
results in a MIL-on, a new DTC would be stored along with the history 
DTC. When trying to diagnose the OBD problem, technicians accessing DTC 
information may have trouble distinguishing which DTC is responsible 
for illuminating the MIL (i.e., which malfunction is present 
currently), and thus could have trouble determining what exactly must 
be repaired. Therefore, we are proposing this requirement for ISO 
engines to help distinguish between DTCs stored for malfunctions that 
are present and those that were present. Note that, for engines using 
SAE J1939 as the communication protocol, such a distinction is already 
provided for.
    Permanent DTCs would also need to be separately identified from the 
other types of DTCs. Additionally, as described above, manufacturers 
would be required to develop additional software routines to store and 
erase permanent DTCs in NVRAM and to prevent erasure from any battery 
disconnect or scan tool command.
vii. Data Stream/Freeze Frame/Test Results
    An important aspect of OBD is the ability of technicians to access 
critical information from the onboard computer to diagnose and repair 
emissions-related malfunctions. We believe that having access through 
the diagnostic connector to real-time electronic information regarding 
certain emissions critical components and systems would provide 
valuable assistance for repairing vehicles properly. The availability 
of real-time information would also provide assistance to technicians

[[Page 3248]]

responding to drivability complaints since the vehicle could be 
operated within the necessary operating conditions and the technician 
could see how various sensors and systems were acting. Similarly, fuel 
economy complaints, loss of performance complaints, intermittent 
problems, and others issues could also be addressed.
    We are proposing a number of data parameters that the OBD system 
would be required to report to a generic scan tool. These parameters, 
which would include information such as engine speed and exhaust gas 
sensor readings, would allow technicians to understand how the vehicle 
engine control system is functioning, either as the vehicle operates in 
a service bay or during actual driving. They would also help 
technicians diagnose and repair emission-related malfunctions by 
allowing them to watch instantaneous changes in the values while 
operating the vehicle.
    Some of the data parameters we are proposing are intended to assist 
us in performing in-use testing of heavy-duty engines for compliance 
with emissions standards. One of the parameters that manufacturers 
would be required to report is the real-time status of the 
NOX and PM ``not-to-exceed'' (NTE) control areas. The NTE 
standards define a wide range of engine operating points where a 
manufacturer must design the engine to be below a maximum emission 
level. In theory, whenever the engine is operated within the speed and 
load region defined as the NTE zone, emissions will be below the 
required standards. However, within the NTE zone, manufacturers are 
allowed, if justified on a case-by-case basis, to either modify the 
time frame in which the standard must be met, and in the second case to 
be exempted from the emission standards under specific conditions 
(e.g., an NTE deficiency). Manufacturers can request two types of 
modifications: first, a five percent limited testing region within 
which no more than five percent of in-use operation is expected to 
occur and, thus, no more than five percent of NTE emissions sampling 
within that region can be compared to the NTE standard for a given 
sampling event; and second, NTE deficiencies which are precisely 
defined exemption conditions where compliance cannot be met due to 
technical reasons or for engine protection. These regions and 
conditions can be defined by directly measured signals or, in some 
cases, by complicated modeled values calculated internally in the 
engine computer. When conducting emissions testing of these engines, 
knowing if the engine is inside the NTE zone--and subject to the NTE 
standards--or is outside of the NTE zone or, perhaps, in an NTE limited 
testing region or covered by an NTE deficiency is imperative. As our 
in-use testing program requirements are written currently, we must post 
process data to determine which data points were generated within a 
compliance zone and which were generated within an exempted zone. Such 
post processing, while possible, is inefficient, time consuming, and 
resource intensive. Having the NTE zone data broadcast in real-time 
over the engine's network would allow for a much more efficient use of 
our resources.
    The specific parameters we are proposing for inclusion in the data 
stream are, for gasoline engines: calculated load value, engine coolant 
temperature, engine speed, vehicle speed, time elapsed since engine 
start, absolute load, fuel level (if used to enable or disable any 
other monitors), barometric pressure (directly measured or estimated), 
engine control module system voltage, commanded equivalence ratio, 
number of stored MIL-on DTCs, catalyst temperature (if directly 
measured or estimated for purposes of enabling the catalyst 
monitor(s)), monitor status (i.e., disabled for the rest of this drive 
cycle, complete this drive cycle, or not complete this drive cycle) 
since last engine shut-off for each monitor used for readiness status, 
distance traveled/engine run time with a commanded MIL-on, distance 
traveled/engine run time since fault memory last cleared, number of 
warm-up cycles since fault memory last cleared, OBD requirements to 
which the engine is certified (e.g., California OBD, EPA OBD, non-OBD) 
and MIL status (i.e., commanded-on or commanded-off). And, for diesel 
engines: calculated load (engine torque as a percentage of maximum 
torque available at the current engine speed),\47\ driver's demand 
engine torque (as a percentage of maximum engine torque), actual engine 
torque (as a percentage of maximum engine torque), reference engine 
maximum torque, reference maximum engine torque as a function of engine 
speed (suspect parameter numbers (SPN) 539 through 543 defined in SAE 
J1939 within parameter group number (PGN) 65251 for engine 
configuration), engine coolant temperature, engine oil temperature (if 
used for emission control or any OBD monitors), engine speed, time 
elapsed since engine start, fuel level (if used to enable or disable 
any other diagnostics), vehicle speed (if used for emission control or 
any OBD monitors), barometric pressure (directly measured or 
estimated), engine control module system voltage, number of stored MIL-
on DTCs, monitor status (i.e., disabled for the rest of this drive 
cycle, complete this drive cycle, or not complete this drive cycle) 
since last engine shut-off for each monitor used for readiness status, 
distance traveled/engine run time with a commanded MIL-on, distance 
traveled/engine run time since fault memory last cleared, number of 
warm-up cycles since DTC memory last cleared, OBD requirements to which 
the engine is certified (e.g., EPA OBD parent rating, EPA OBD child 
rating, non-OBD), and MIL status (i.e., commanded-on or commanded-off). 
Also for diesel engines, as discussed above, separate NOX 
and PM NTE control area status (i.e., inside control area, outside 
control area, inside manufacturer-specific NTE carve-out area, or 
deficiency active area). Also, for all engines so equipped (and only 
those so equipped): absolute throttle position, relative throttle 
position, fuel control system status (e.g., open loop, closed loop), 
fuel trim, fuel pressure, ignition timing advance, fuel injection 
timing, intake air/manifold temperature, engine intercooler 
(aftercooler) temperature, manifold absolute pressure, air flow rate 
from mass air flow sensor, secondary air status (upstream, downstream, 
or atmosphere), ambient air temperature, commanded purge valve duty 
cycle/position, commanded EGR valve duty cycle/position, actual EGR 
valve duty cycle/position, EGR error between actual and commanded, PTO 
status (active or not active), redundant absolute throttle position 
(for electronic throttle or other systems that utilize two or more 
sensors), absolute pedal position, redundant absolute pedal position, 
commanded throttle motor position, fuel rate, boost pressure, 
commanded/target boost pressure, turbo inlet air temperature, fuel rail 
pressure, commanded fuel rail pressure, DPF inlet pressure, DPF inlet 
temperature, DPF outlet pressure, DPF outlet temperature, DPF delta 
pressure, exhaust pressure sensor output, exhaust gas temperature 
sensor output, injection control pressure, commanded injection control 
pressure, turbocharger/turbine speed,


[[Continued on page 3249]]


From the Federal Register Online via GPO Access [wais.access.gpo.gov]
]                         
 
[[pp. 3249-3298]] Control of Air Pollution From New Motor Vehicles and New Motor 
Vehicle Engines; Regulations Requiring Onboard Diagnostic Systems on 
2010 and Later Heavy-Duty Engines Used in Highway Applications Over 
14,000 Pounds; Revisions to Onboard Diagnostic Requirements f[[Page 3249]]

[[Continued from page 3248]]

[[Page 3249]]

variable geometry turbo position, commanded variable geometry turbo 
position, turbocharger compressor inlet temperature, turbocharger 
compressor inlet pressure, turbocharger turbine inlet temperature, 
turbocharger turbine outlet temperature, wastegate valve position, glow 
plug lamp status, oxygen sensor output, air/fuel ratio sensor output, 
NOX sensor output, and evaporative system vapor pressure.
---------------------------------------------------------------------------

    \47\ Note that, for purposes of the calculated load and torque 
parameters for diesel engines, manufacturers would be required to 
report the most accurate values that are calculated within the 
applicable electronic control unit (e.g., the engine control 
computer). ``Most accurate values,'' in this context, would be those 
of sufficient accuracy, resolution, and filtering that they could be 
used for the purpose of in-use emissions testing with the engine 
still in a vehicle (e.g., using portable emissions measurement 
equipment).
---------------------------------------------------------------------------

    We are also proposing requirements for storage of ``freeze frame'' 
information at the time a malfunction is detected and a DTC is stored. 
The freeze frame provides the operating conditions of the vehicle at 
the time of malfunction detection and the DTC associated with the data. 
The parameters we are proposing for inclusion in the freeze frame are a 
subset of the parameters listed above for the data stream. Note that 
storage of only one freeze frame would be required. Manufacturers may 
choose to store additional frames, provided that the required frame can 
be read using a scan tool meeting SAE J1978 specifications or designed 
to communicate with an SAE J1939 network.
    We are also proposing that the OBD system store the most recent 
monitoring results for most of the major monitors. Manufacturers would 
be required to store and make available to the scan tool certain test 
information--i.e., the minimum and maximum values that should occur 
during proper operation along with the actual test value--of the most 
recent monitoring event. ``Passing'' systems would store test results 
that are within the test limits, while ``failing'' systems would store 
test results that are outside the test limits. The storage of test 
results would assist technicians in diagnosing and repairing 
malfunctions and would help distinguish between components that are 
performing well below the malfunction thresholds from those that are 
passing the malfunction thresholds marginally.
viii. Identification Numbers
    We are also proposing that manufacturers be required to report two 
identification numbers related to the software and specific calibration 
values in the onboard computer. The first item, Calibration 
Identification Number (CAL ID), would identify the software version 
installed in the onboard computer. Software is often changed following 
production of the engine. These software changes often make changes to 
the emissions control system or the OBD system. We are proposing that 
these changes include a new CAL ID and that it be communicated via the 
diagnostic connector to the scan tool. The second item, Calibration 
Verification Number (CVN), would help to ensure that the current 
software has not been corrupted, modified inappropriately, or otherwise 
tampered with. Both CAL ID and CVN help ensure the integrity of the OBD 
system. The CVN proposal would require manufacturers to develop 
sophisticated software algorithms that would essentially be a self-
check calculation of all of the emissions-related software and 
calibration values in the onboard computer and would return the result 
of the calculation to a scan tool. If the calculated result did not 
equal the expected result for that CAL ID, one would know that the 
software had been corrupted or otherwise modified. The CVN result would 
have to be made available at all times to a generic scan tool.
    We are also proposing that the Vehicle Identification Number (VIN) 
be communicated via the diagnostic connector to a generic scan tool in 
a standardized format. The VIN would be a unique number assigned by the 
vehicle manufacturer to every vehicle built. The VIN is commonly used 
for purposes of ownership and registration to uniquely identify every 
vehicle. By requiring the VIN to be stored in the onboard computer and 
available electronically to a generic scan tool, the possibility of a 
fraudulent inspection (e.g., by plugging into a different vehicle than 
an inspection citation was issued originally to generate a proof of 
correction) would be minimized. Electronic access to this number would 
also simplify the inspection process and reduce transcription errors 
from manual data entry.
    We are proposing that the VIN be electronically stored in a control 
module on the vehicle, but not that it necessarily be stored in the 
engine control module. As long as the VIN is reported correctly and 
according to the selected reference document standards, we consider it 
irrelevant as to which control module (e.g., engine controller, 
instrument cluster controller) contains the information. Further, we 
are proposing that the ultimate responsibility would lie with the 
engine manufacturer to ensure that every vehicle manufactured with one 
of its engines satisfies this requirement. However, we would expect 
that the physical task of implementing this requirement would likely be 
passed from the engine manufacturer to the vehicle manufacturer via an 
additional build specification. Thus, analogous to how the engine 
manufacturer currently provides engine purchasers with detailed 
specifications regarding engine cooling requirements, additional sensor 
inputs, physical mounting specifications, weight limitations, etc., the 
engine manufacturer would likely include an additional specification 
dictating the need for the VIN to be made available electronically. It 
would be left to each engine manufacturer to determine the most 
effective method to achieve this, as long as the VIN requirement is 
met. Some manufacturers may find it most effective to provide the 
capability in the engine control module delivered with the engine 
coupled with a mechanism for the vehicle manufacturer to program the 
module with the VIN upon installation of the engine into an actual 
vehicle. Others may find it more effective to require the vehicle 
manufacturer to have the capability built into other modules installed 
on the vehicle such as instrument cluster modules, etc. We are aware of 
several current vehicles with engines from three different engine 
manufacturers that already have the VIN available through engine-
manufacturer specific scan tools; this indicates that such arrangements 
already exist in one form or another and that they are working.
5. In-Use Performance Ratio Tracking Requirements
    To separately report an in-use performance ratio for each 
applicable monitor as discussed in sections II.B through II.D, we are 
proposing that manufacturers be required to implement software 
algorithms to report a numerator and denominator in the standardized 
format specified below and in accordance with the specifications of the 
reference documents listed in section II.F.1.
    For the numerator, denominator, general denominator, and ignition 
cycle counter:
     Each number must have a minimum value of zero and a 
maximum value of 65,535 with a resolution of one.
     Each number must be reset to zero only when a non-volatile 
random access memory (NVRAM) reset occurs (e.g., reprogramming event) 
or, if the numbers are stored in keep-alive memory (KAM), when KAM is 
lost due to an interruption in electrical power to the control module 
(e.g., battery disconnect). Numbers may not be reset to zero under any 
other circumstances including when commanded to do so via a scan tool 
command to clear DTCs or reset KAM.
     If either the numerator or denominator for a specific 
component reaches the maximum value of 65,535 2, both 
numbers should be divided by two before either is incremented again to 
avoid overflow problems.

[[Page 3250]]

     If the ignition cycle counter reaches the maximum value of 
65,535 2, the ignition cycle counter should rollover and 
increment to zero on the next ignition cycle to avoid overflow 
problems.
     If the general denominator reaches the maximum value of 
65,535 2, the general denominator should rollover and 
increment to zero on the next drive cycle that meets the general 
denominator definition to avoid overflow problems.
     If an engine is not equipped with a component (e.g., 
oxygen sensor bank 2, secondary air system), the corresponding 
numerator and denominator for that specific component should always be 
reported as zero.
    For the in-use performance ratio:
     The ratio should have a minimum value of zero and a 
maximum value of 7.99527 with a resolution of 0.000122.
     A ratio for a specific component should be considered to 
be zero whenever the corresponding numerator is equal to zero and the 
corresponding denominator is not zero.
     A ratio for a specific component should be considered to 
be the maximum value of 7.99527 if the corresponding denominator is 
zero or if the actual value of the numerator divided by the denominator 
exceeds the maximum value of 7.99527.
    For engine run time tracking on all gasoline and diesel engines, 
manufacturers would be required to implement software algorithms to 
individually track and report in a standardized format the engine run 
time while being operated in the following conditions:
     Total engine run time
     Total idle run time (with ``idle'' defined as accelerator 
pedal released by driver, vehicle speed less than or equal to one mile 
per hour, and PTO not active);
     Total run time with PTO active.
    Each of the above engine run time counters would have the following 
numerical value specifications:
     Each numerical counter must be a four-byte value with a 
minimum value of zero at a resolution of one minute per bit.
     Each numerical counter must be reset to zero only when a 
nonvolatile memory reset occurs (e.g., a reprogramming event). 
Numerical counters cannot be reset to zero under any other 
circumstances including a scan tool (generic or enhanced) command to 
clear DTCs or reset KAM.
     When any of the individual numerical counters reaches its 
maximum value, all counters must be divided by two before any are 
incremented again. This is meant to avoid overflow problems.
6. Exceptions to Standardization Requirements
    For alternative-fueled engines derived from a diesel-cycle engine, 
we are proposing that the manufacturer be allowed to meet the 
standardized requirements discussed in this section that are applicable 
to diesel engines rather than meeting the requirements applicable to 
gasoline engines.

G. Implementation Schedule, In-Use Liability, and In-Use Enforcement

1. Implementation Schedule and In-Use Liability Provisions
    Table II.G-1 summarizes the proposed implementation schedule for 
the OBD monitoring requirements--i.e., the proposed certification 
requirements and in-use liabilities. More detail regarding the 
implementation schedule and liabilities can be found in the sections 
that follow.

 Table II.G-1.--OBD Certification Requirements and In-use Liability for
Diesel Fueled and Gasoline Fueled Engines over 14,000 Pounds: Monitoring
                              Requirements
------------------------------------------------------------------------
                                        Certification
    Model year        Applicability      requirement    In-use liability
------------------------------------------------------------------------
2010-2012.........  Parent rating     Full liability    Full liability
                     within 1          to thresholds     to 2x
                     compliant         according to      thresholds. \c\
                     engine family.    certification
                     \a\               demonstration
                                       procedures. \b\
                    Child ratings     Certification     Liability to
                     within the        documentation     monitor and
                     compliant         only (i.e., no    detect as noted
                     engine family.    certification     in
                                       demonstration);   certification
                                       no liability to   documentation.
                                       thresholds.
                    All other engine  None............  None.
                     families and
                     ratings.
2013-2015.........  Parent rating     Full liability    Full liability
                     from 2010-2012    to thresholds     to 2x
                     and parent        according to      thresholds.
                     rating within 1-  certification
                     2 additional      demonstration
                     engine families.  procedures.
                    Child ratings     Full liability    Full liability
                     from 2010-2012    to thresholds     to 2x
                     and parent        but               thresholds.
                     ratings from      certification
                     any remaining     documentation
                     engine families   only.
                     or OBD
                     groups.\d\
                    Additional        Certification     Liability to
                     engine ratings.   documentation     monitor and
                                       only; no          detect as noted
                                       liability to      in
                                       thresholds.       certification
                                                         demonstration.
 2016-2018........  One rating from   Full liability    Full liability
                     1-3 engine        to thresholds     to thresholds.
                     families and/or   according to
                     OBD groups.       certification
                                       demonstration
                                       procedures.
                    Remaining         Full liability    Full liability
                     ratings.          to thresholds     to 2x
                                       but               thresholds.
                                       certification
                                       documentation
                                       only.
2019+.............  One rating from   Full liability    Full liability
                     1-3 engine        to thresholds     to thresholds.
                     families and/or   according to
                     OBD groups.       certification
                                       demonstration
                                       procedures.
                    Remaining         Full liability    Full liability
                     ratings.          to thresholds     to thresholds.
                                       but
                                       certification
                                       documentation
                                       only.
------------------------------------------------------------------------
Notes: (a) Parent and child ratings are defined in section II.G; which
  rating(s) serves as the parent rating and which engine families must
  comply is not left to the manufacturer, as discussed in section II.G.
  (b) The certification demonstration procedures and the certification
  documentation requirements are discussed in section VIII.B. (c) Where
  in-use liability to thresholds and 2x thresholds is noted,
  manufacturer liability to monitor and detect as noted in their
  certification documentation is implied. (d) OBD groups are groupings
  of engine families that use similar OBD strategies and/or similar
  emissions control systems, as described in the text.

    For the 2010 through 2012 model years, manufacturers would be 
required to implement OBD on one engine family. All other 2010 through 
2012 engine families would not be subject to any OBD requirements 
unless otherwise required to do so (e.g., to demonstrate that SCR 
equipped vehicles will not be operated without urea). For 2013,

[[Page 3251]]

manufacturers would be required to implement OBD on all engine 
families.
    We are proposing this implementation schedule for several reasons. 
First, industry has made credible arguments that their resources are 
stretched to the limit developing and testing strategies for compliance 
with the 2007/2010 heavy-duty highway emissions standards. We do not 
want to jeopardize their success toward that goal by being too 
aggressive with our OBD program. Second, OBD is a complex and difficult 
regulation with which to comply. We believe that our implementation 
schedule would give industry the opportunity to introduce OBD systems 
on a limited number of engines giving them and us very valuable 
learning experience. Should mistakes or errors in regulatory 
interpretation occur, the ramifications would be limited to only a 
subset of the new vehicle fleet rather than the entire new vehicle 
fleet. Lastly, the proposed OBD requirements outlined above, and the 
production vehicle evaluation provisions discussed in Section VIII, 
reflect 10 to 20 years of learning by EPA, CARB, and industry 
(primarily the light-duty gasoline industry) as to what works and what 
does not work. This is, perhaps, especially true for those OBD elements 
that involve the interface between the OBD system and service and I/M 
inspection personnel. Gasoline manufacturers have had the ability to 
evolve their OBD systems along with this learning process. However, 
diesel engine manufacturers have not really been involved in this 
learning process and, as a result, 100 percent implementation in 2010 
would be analogous to implementing 10 to 20 years of OBD learning in 
one implementation step. We believe that implementing in two or three 
gradual steps rather than one big step will benefit everyone involved.
    Table II.G-1 makes reference to ``parent'' and ``child'' ratings. 
In general, engine manufacturers certify an engine family that consists 
of several ratings having slightly different horsepower and/or torque 
characteristics but no differences large enough to require a different 
engine family designation. For emissions certification, the parent 
rating--i.e., the rating for which emissions data are submitted to EPA 
for the purpose of demonstrating emissions compliance--is defined as 
the ``worst case'' rating. This worst case rating is the rating 
considered as having the worst emissions performance and, therefore, 
its compliance demonstrates that all other ratings within the family 
must comply. For OBD purposes, we wanted to limit the burden on 
industry--hence the proposal for only one compliant engine family in 
2010--yet maximize the impact of the OBD system. Therefore, for model 
years 2010 through 2012, we are defining the OBD parent rating as the 
rating having the highest weighted projected sales within the engine 
family having the highest weighted projected sales, with sales being 
weighted by the useful life of the engine rating. Table II.G-2 presents 
a hypothetical example for how this would work. Using this approach, 
the OBD compliant engine family in 2010 would be the engine family 
projected to produce the most in-use emissions (based on sales weighted 
by expected miles driven). Likewise, the fully liable parent OBD rating 
would be the rating within that family projected to produce the most 
in-use emissions.

                             Table II.G-2.--Hypothetical Example of How the OBD Parent and Child Ratings Would Be Determined
--------------------------------------------------------------------------------------------------------------------------------------------------------
                                                                                                                     OBD weighting--    OBD weighting--
                                                                                       Projected       Certified    engine rating \a\  engine family \b\
                  OBD group                          Engine family         Rating        sales        useful life       (billions)         (billions)

--------------------------------------------------------------------------------------------------------------------------------------------------------
I............................................  A                                 1          10,000         285,000              2.85              14.25
                                               ........................          2          40,000         285,000             11.4    .................
                                               B                                 1          10,000         435,000              4.35              21.60
                                               ........................          2          20,000         435,000              8.70   .................
                                               ........................          3          30,000         285,000              8.55   .................
II...........................................  C                                 1          20,000         110,000              2.20               7.70
                                               ........................          2          50,000         110,000              5.50   .................
--------------------------------------------------------------------------------------------------------------------------------------------------------
Notes: (a) For engine family A, rating 1, 10,000 x 285,000 / 1 billion = 2.85.
(b) For engine family A, 2.85 + 11.4 = 14.25.

    In the example shown in Table II.G-2, the compliant engine family 
in 2010 would be engine family B and the parent OBD rating within that 
family would be rating 2. The other OBD compliant ratings within engine 
family B would be dubbed the ``child'' ratings. For model years 2013 
through 2015, the parent ratings would be those ratings having the 
highest weighted projected sales within each of the one to three engine 
families having the highest weighted projected sales, with sales being 
weighted by the useful life of the engine rating. In the example shown 
in Table II.G-2, the parent ratings would be rating 2 of engine family 
A, rating 2 of engine family B, and rating 2 of engine family C (Note 
that this is only for illustration purposes since our proposal would 
not require that a manufacturer with only three engine families have 
three parent ratings and instead would require only one).
    The manufacturer would not need to submit test data demonstrating 
compliance with the emissions thresholds for the child ratings. We 
would fully expect these child ratings to use OBD calibrations--i.e., 
malfunction trigger points--that are identical or nearly so to those 
used on the parent rating. However, we would allow manufacturers to 
revise the calibrations on their child ratings where necessary so as to 
avoid unnecessary or inappropriate MIL illumination. Such revisions to 
OBD calibrations have been termed ``extrapolated'' OBD calibrations 
and/or systems. The revisions to the calibrations on child ratings and 
the rationale for them would need to be very clearly described in the 
certification documentation.
    For the 2013 and later model years, we are proposing that 
manufacturers certify one to three parent ratings. The actual number of 
parent ratings would depend upon the manufacturer's fleet and would be 
based on both the emissions control system architectures present in 
their fleet and the similarities/differences of the engine families in 
their fleet. For example, a manufacturer that uses a DPF with 
NOX adsorber on each of the engines would have only one 
system architecture. Another manufacturer that uses a DPF with 
NOX adsorber on some engines and a DPF with SCR on others 
would have at least two architectures. We would expect that 
manufacturers would group similar architectures and similar engine

[[Page 3252]]

families into so called ``OBD groups.'' These OBD groups would consist 
of a combination of engines, engine families, or engine ratings that 
use the same OBD strategies and similar calibrations. The manufacturer 
would be required to submit details regarding their OBD groups as part 
of their certification documentation that shows the engine families and 
engine ratings within each OBD group for the coming model year. While a 
manufacturer may end up with more than three OBD groups, we do not 
intend to require a parent rating for more than three OBD groups. 
Therefore, in the example shown in Table II.G-2, rather than submitting 
test data for the three parent ratings as suggested above, the OBD 
grouping would result in the parent ratings being rating 2 of engine 
family B and rating 2 of engine family C. These parents would represent 
OBD groups I and II, and the manufacturer's product line. For 2013 
through 2015, we intend to allow the 2010 parent to again act as a 
parent rating and, provided no significant changes had been made to the 
engine or its emissions control system, complete carryover would be 
possible. However, for model years 2016 and beyond, we would work 
closely with CARB staff and the manufacturer to determine the parent 
ratings so that the same ratings are not acting as the parents every 
year. In other words, our definitions for the OBD parent ratings as 
discussed here apply only during the years 2010 through 2012 and again 
for the years 2013 through 2015. We request comment on this approach.
    In addition to this gradual certification implementation schedule, 
we are proposing some relaxations for in-use liability during the 2010 
through 2018 model years. The first such relaxation is higher interim 
in-use compliance standards for those OBD monitors calibrated to 
specific emissions thresholds. For the 2010 through 2015 model years, 
an OBD monitor on an in-use engine would not be considered non-
compliant (i.e., subject to enforcement action) unless emissions 
exceeded twice the OBD threshold without detection of a malfunction. 
For example, for an EGR monitor on an engine with a NOX FEL 
of 0.2 g/bhp-hr and an OBD threshold of 0.5 g/bhp-hr (i.e., the 
NOX FEL+0.3), a manufacturer would not be subject to 
enforcement action unless emissions exceeded 1.0 g/bhp-hr 
NOX without a malfunction being detected. For the model 
years 2016 through 2018, parent ratings would be liable to the 
certification emissions thresholds, but child ratings and other ratings 
would remain liable to twice the certification thresholds. Beginning in 
the 2019 model year, all families and all ratings would be liable to 
the certification thresholds.
    The second in-use relaxation is a limitation in the number of 
engines that would be liable for in-use compliance with the OBD 
emissions thresholds. For 2010 through 2012, we are proposing that 
manufacturers be fully liable in-use to twice the thresholds for only 
the OBD parent rating. The child ratings within the compliant engine 
family would have liability for monitoring in the manner described in 
the certification documentation, but would not have liability for 
detecting a malfunction at the specified emissions thresholds. For 
example, a child rating's DPF monitor designed to operate under 
conditions X, Y, and Z and calibrated to detect a backpressure within 
the range A to B would be expected to do exactly that during in-use 
operation. However, if the tailpipe emissions of the child engine were 
to exceed the applicable OBD in-use thresholds (i.e., 2x the 
certification thresholds during 2010-2015), despite having a 
backpressure within range A to B under conditions X, Y, and Z, there 
would be no in-use OBD failure nor cause for enforcement action. In 
fact, we would expect the OBD monitor to determine that the DPF was 
functioning properly since its backpressure was in the acceptable 
range. For model years 2013 through 2015, this same in-use relaxation 
would apply to those engine families that do not lie within an engine 
family for which a parent rating has been certified. For 2016 and later 
model years, all engines would have some in-use liability to 
thresholds, either the certification thresholds or twice those 
thresholds.
    These in-use relaxations are meant to provide ample time for 
manufacturers to gain experience without an excessive level of risk for 
mistakes. They would also allow manufacturers to fine-tune their 
calibration techniques over a six to ten year period.
    We are also proposing some a specific implementation schedule for 
the standardization requirements discussed in section II.F. We 
initially intended to require that any compliant OBD engine family 
would be required to implement all of the standardization requirements. 
However, we became concerned that, during model years 2010 through 
2012, we could have a situation where OBD compliant engines from 
manufacturer A might be competing against non-OBD engines from 
manufacturer B for sales in the same truck. In such a case, the truck 
builder would be placed in a difficult position of needing to design 
their truck to accommodate OBD compliant engines--along with a 
standardized MIL, a specific diagnostic connector location 
specification, etc.--and non-OBD engines. After consideration of this 
almost certain outcome, we have decided to limit the standardization 
requirements that must be met during the 2010 through 2012 model years. 
Beginning in 2013, all engines will be OBD compliant and this would 
become a moot issue. Table II.G-3 shows the proposed implementation 
schedule for standardization requirements.

  Table II.G-3.--OBD Standardization Requirements for Diesel Fueled and
               Gasoline Fueled Engines Over 14,000 Pounds
------------------------------------------------------------------------
                                          Required           Waived
    Model year        Applicability    standardization   standardization
                                          features          features
------------------------------------------------------------------------
2010-2012.........  Parent and Child  Emissions         Standardized
                     ratings within    related           connector
                     1 compliant       (II.F.4) except   (II.F.2).
                     engine family.    for the           Dedicated
                     \a\               requirement to    (i.e.,
                                       make the data     regulated OBD-
                                       available in a    only) MIL.
                                       standardized      Communication
                                       format or in      protocols
                                       accordance with   (II.F.3).
                                       SAE J1979/1939    Emissions
                                       specifications)   related
                                       . MIL             functions
                                       activation and    (II.F.4) with
                                       deactivation.\b   respect to the
                                       \ Performance     requirement to
                                       tracking--calcu   make the data
                                       lation of         available in a
                                       numerators,       standardized
                                       denominators,     format or in
                                       ratios.           accordance with
                                                         SAE J1979/1939
                                                         specifications)
                    Other engine      None............  All.
                     families.
2013+.............  All engine        All.............  None.
                     families and
                     ratings.
------------------------------------------------------------------------
Notes: (a) Parent and child ratings are defined in section II.G; which
  rating serves as the parent rating and which engine families must
  comply is not left to the manufacturer, as discussed in section II.G.
  (b) There would be no requirement for a dedicated MIL and no
  requirement to use a specific MIL symbol, only that a MIL be used and
  that it use the proposed activation/deactivation logic.


[[Page 3253]]

2. In-Use Enforcement
    When conducting our in-use enforcement investigations into OBD 
systems, we intend to use all tools we have available to analyze the 
effectiveness and compliance of the system. These tools may include on-
vehicle emission testing systems such as the portable emissions 
measurement systems (PEMS). We would also use scan tools and data 
loggers to analyze the data stream information to compare real world 
operation to the documentation provided at certification.
    Importantly, we would not intend to pursue enforcement action 
against a manufacturer for not detecting a failure mode that could not 
have been reasonably predicted or otherwise detected using monitoring 
methods known at the time of certification. For example, we are 
proposing a challenging set of requirements for monitoring of DPF 
systems. As of today, engine manufacturers are reasonably confident in 
their ability to detect certain DPF failure modes at or near the 
proposed thresholds--e.g., a leaking DPF resulting from a cracked 
substrate--but are not confident in their ability to detect some other 
DPF failure modes--e.g., a leaking DPF resulting from a partially 
melted substrate. If a partially melted substrate indeed cannot be 
detected and this is known during the certification process, we cannot 
expect such a failure to be detected on an in-use vehicle.
    We also want to make it clear who would be the responsible party 
should we pursue any in-use enforcement action with respect to OBD. We 
are very familiar with the heavy-duty industry and its tendency toward 
separate engine and component suppliers. This contrasts with the light-
duty industry which tends toward a more vertically integrated 
structure. The non-vertically integrated nature of the heavy-duty 
industry can present unique difficulties for OBD implementation and for 
OBD enforcement. With the complexity of OBD systems, especially those 
meeting the requirements being proposed today, we would expect the 
interactions between the various parties involved--engine manufacturer, 
transmission manufacturer, vehicle manufacturer, etc.--to be further 
complicated. Nonetheless, in the end the vast majority of the proposed 
OBD requirements would apply directly to the engine and its associated 
emission controls, and the engine manufacturer would have complete 
responsibility to ensure that the OBD system performs properly in-use. 
Given the central role the engine and engine control unit would play in 
the OBD system, we are proposing that the party certifying the engine 
and OBD system (typically, the engine manufacturer) be the responsible 
party for in-use compliance and enforcement actions. In this role, the 
certifying party would be our sole point of contact for potential 
noncompliances identified during in-use or enforcement testing. We 
would leave it to the engine manufacturer to determine the ultimate 
party responsible for the potential noncompliance (e.g., the engine 
manufacturer, the vehicle manufacturer, or some other supplier). In 
cases where remedial action such as an engine recall would be required, 
the certifying party would take on the responsibility of arranging to 
bring the engines or OBD systems back into compliance. Given that 
heavy-duty engines are already subject to various emission requirements 
including engine emission standards, labels, and certification, engine 
manufacturers currently impose restrictions via signed agreements with 
engine purchasers to ensure that their engines do not deviate from 
their certified configuration when installed. We would expect the OBD 
system's installation to be part of such agreements in the future.

H. Proposed Changes to the Existing 8,500 to 14,000 Pound Diesel OBD 
Requirements

    We are also proposing changes to our OBD requirements for diesel 
engines used in heavy-duty vehicles under 14,000 pounds (see 40 CFR 
86.005-17 for engine-based requirements and 40 CFR 86.1806-05 for 
vehicle or chassis-based requirements). Table II.H-1 summarizes the 
proposed changes to under 14,000 pound heavy-duty diesel emissions 
thresholds at which point a component or system has failed to the point 
of requiring an illuminated MIL and a stored DTC. Table II.H-2 
summarizes the proposed changes for diesel engines used in heavy-duty 
applications under 14,000 pounds. The proposed changes are meant to 
maintain consistency with the diesel OBD requirements we are proposing 
for over 14,000 pound applications.

 Table II.H-1.--Proposed New, or Proposed Changes to Existing, Emissions Thresholds for Diesel Fueled CI Heavy-
                                    duty Vehicles Under 14,000 Pounds (g/mi)
----------------------------------------------------------------------------------------------------------------
      Component/monitor              MY             NMHC             CO              NOX               PM
----------------------------------------------------------------------------------------------------------------
NMHC catalyst system.........  2010-2012.....  2.5x.
                               2013+.........  2x.
NOX catalyst system..........  2007-2009.....  ..............  ..............  3x............
                               2010+.........  ..............  ..............  +0.3.
DPF system...................  2010-2012.....  2.5x..........  ..............  ..............  4x.
                               2013+.........  2x............  ..............  ..............  +0.04.
Air-fuel ratio sensors         2007-2009.....  2.5x..........  2.5x..........  3x............  4x.
 upstream.
                               2010-2012.....  2.5x..........  2.5x..........  +0.3..........  +0.02.
                               2013+.........  2x............  2x............  +0.3..........  +0.02.
Air-fuel ratio sensors         2007-2009.....  2.5x..........  ..............  3x............  4x.
 downstream.
                               2010-2012.....  2.5x..........  ..............  +0.3..........  4x.
                               2013+.........  2x............  ..............  +0.3..........  +0.04.
NOX sensors..................  2007-2009.....  ..............  ..............  4x............  5x.
                               2010-2012.....  ..............  ..............  +0.3..........  4x.
                               2013+.........  ..............  ..............  +0.3..........  +0.04.
``Other monitors'' with        2007-2009.....  2.5x..........  2.5x..........  3x............  4x.
 emissions thresholds.
                               2010-2012.....  2.5x..........  2.5x..........  +0.3..........  4x.
                               2013+.........  2x............  2x............  +0.3..........  +0.02.
----------------------------------------------------------------------------------------------------------------
Notes: MY=Model Year; 2.5x means a multiple of 2.5 times the applicable emissions standard; +0.3 means the
  standard plus 0.3; not all proposed monitors have emissions thresholds but instead rely on functionality and
  rationality checks as described in section II.D.4.


[[Page 3254]]


Table II.H-2.--Proposed New, or Proposed Changes to Existing, Emissions Thresholds for Diesel Fueled CI Engines Used in Heavy-duty Vehicles Under 14,000
                                                                    Pounds (g/bhp-hr)
--------------------------------------------------------------------------------------------------------------------------------------------------------
        Component/Monitor                  MY                Std/FEL              NMHC                CO                NOX                  PM
--------------------------------------------------------------------------------------------------------------------------------------------------------
NMHC catalyst system.............  2010-2012.........  All...............  2.5x.
                                   2013+.............  All...............  2x.
NOX catalyst system..............  2007-2009.........  >0.5 NOX..........  .................  .................  1.75x.
                                   2007-2009.........  < =0.5 NOX.........  .................  .................  +0.5.
                                   2010+.............  All...............  .................  .................  +0.3.
DPF system.......................  2010-2012.........  All...............  2.5x.............  .................  .................  0.05/+0.04.
                                   2013+.............  All...............  2x...............  .................  .................  0.05/+0.04.
Air-fuel ratio sensors upstream..  2007-2009.........  >0.5 NOX..........  2.5x.............  2.5x.............  1.75x............  0.05/+0.04.
                                   2007-2009.........  < =0.5 NOX.........  2.5x.............  2.5x.............  +0.5.............  0.05/+0.04.
                                   2010-2012.........  All...............  2.5x.............  2.5x.............  +0.3.............  0.03/+0.02.
                                   2013+.............  All...............  2x...............  2x...............  +0.3.............  0.03/+0.02.
Air-fuel ratio sensors downstream  2007-2009.........  >0.5 NOX..........  2.5x.............  .................  1.75x............  0.05/+0.04.
                                   2007-2009.........  < =0.5 NOX.........  2.5x.............  .................  +0.5.............  0.05/+0.04.
                                   2010-2012.........  All...............  2.5x.............  .................  +0.3.............  0.05/+0.04.
                                   2013+.............  All...............  2x...............  .................  +0.3.............  0.05/+0.04.
NOX sensors......................  2007-2009.........  >0.5 NOX..........  .................  .................  1.75x............  0.05/+0.04.
                                   2007-2009.........  < =0.5 NOX.........  .................  .................  +0.5.............  0.05/+0.04.
                                   2010+.............  All...............  .................  .................  +0.3.............  0.05/+0.04.
``Other monitors'' with emissions  2007-2009.........  >0.5 NOX..........  2.5x.............  2.5x.............  1.75x............  0.05/+0.04.
 thresholds.
                                   2007-2009.........  < =0.5 NOX.........  2.5x.............  2.5x.............  +0.5.............  0.05/+0.04.
                                   2010-2012.........  All...............  2.5x.............  2.5x.............  +0.3.............  0.03/+0.02.
                                   2013+.............  All...............  2x...............  2x...............  +0.3.............  0.03/+0.02.
--------------------------------------------------------------------------------------------------------------------------------------------------------
Notes: MY=Model Year; 2.5x means a multiple of 2.5 times the applicable emissions standard or family emissions limit (FEL); +0.3 means the standard or
  FEL plus 0.3; 0.05/+0.04 means an absolute level of 0.05 or an additive level of the standard or FEL plus 0.04, whichever level is higher; not all
  proposed monitors have emissions thresholds but instead rely on functionality and rationality checks as described in section II.D.4.

1. Selective Catalytic Reduction and Lean NOX Catalyst 
Monitoring
    We are proposing that the 8,500 to 14,000 pound SCR and lean 
NOX catalyst monitoring requirements mirror those discussed 
in section II.B.6. The current regulations require detection of a 
NOX catalyst malfunction before emissions exceed 1.5x the 
emissions standards. We no longer believe that such a tight threshold 
level is appropriate for diesel SCR and lean NOX catalyst 
systems. We believe that such a tight threshold could result in too 
many false failure indications. The required monitoring conditions with 
respect to performance tracking (discussed in section II.B.6.c) would 
not apply for under 14,000 pound heavy-duty applications since we do 
not have performance tracking requirements for under 14,000 pound 
applications. We are proposing this change for the 2007 model year.
2. NOX Adsorber System Monitoring
    We are proposing that the 8,500 to 14,000 pound NOX 
adsorber monitoring requirements mirror those discussed in section 
II.B.7. The current regulations require detection of a NOX 
adsorber malfunction before emissions exceed 1.5x the emissions 
standards. We no longer believe that such a tight threshold level is 
appropriate for diesel NOX adsorber systems. We believe that 
such a tight threshold could result in too many false failure 
indications. The required monitoring conditions with respect to 
performance tracking (discussed in section II.B.7.c) would not apply 
for under 14,000 pound heavy-duty applications since we do not have 
performance tracking requirements for under 14,000 pound applications. 
We are proposing this change for the 2007 model year.
3. Diesel Particulate Filter System Monitoring
    We are proposing that the 8,500 to 14,000 pound DPF monitoring 
requirements mirror those discussed in section II.B.8. Our current 
regulations require detection of a catastrophic failure only. The 
proposed monitoring requirements discussed in section II.B.8 would be 
far more comprehensive and protective of the environment than would a 
catastrophic failure monitor. The required monitoring conditions with 
respect to performance tracking (discussed in section II.B.8.c) would 
not apply for under 14,000 pound heavy-duty applications since we do 
not have performance tracking requirements for under 14,000 pound 
applications. We are proposing no changes to the DPF monitoring 
requirements in the 2007 to 2009 model years because there is not 
sufficient lead time for manufacturers to develop a new monitor. The 
new, more stringent monitoring requirements would begin in the 2010 
model year, with a further tightening of the DPF NMHC threshold in the 
2013 model year as is also proposed for over 14,000 pound applications.
4. NMHC Converting Catalyst Monitoring
    We are proposing that the 8,500 to 14,000 pound NMHC converting 
catalyst monitoring requirements mirror those discussed in section 
II.B.5. Our current regulations do not require the monitoring of NMHC 
catalysts on diesel applications. The proposed monitoring requirements 
discussed in section II.B.5 would be far more comprehensive and 
protective of the environment than the current lack of any requirement. 
The required monitoring conditions with respect to performance tracking 
(discussed in section II.B.8.c) would not apply for under 14,000 pound 
heavy-duty applications since we do not have performance tracking 
requirements for under 14,000 pound applications. We are not proposing 
this new threshold for the 2007 to 2009 model years because there is 
not sufficient lead time for manufacturers to develop a new monitor. 
The new, more stringent monitoring requirements would begin in the 2010 
model year, with a further tightening of the NMHC threshold in the 2013 
model year as is also proposed for over 14,000 pound applications.
5. Other Monitors
    We are also proposing changes to the emissions thresholds for all 
other diesel monitors in the 8,500 to 14,000 pound range (e.g., 
NOX sensors, air fuel ratio

[[Page 3255]]

sensors, etc.). These proposed changes are meant to maintain 
consistency with the proposed changes for over 14,000 pound 
applications. We believe that these proposed thresholds are far more 
appropriate for diesel applications than the thresholds we have in our 
current OBD requirements which are, generally, 1.5 times the applicable 
standards. None of the proposed thresholds represents a new threshold 
where none currently exists. Instead, they represent different 
thresholds that would require, in most cases, malfunction detection at 
different emissions levels than would be required by our current OBD 
requirements.
6. CARB OBDII Compliance Option and Deficiencies
    We are also proposing some changes to our deficiency provisions for 
vehicles and engines meant for vehicles under 14,000 pounds. We have 
included specific mention of air-fuel ratio sensors and NOX 
sensors where we had long referred only to oxygen sensors. We have also 
updated the referenced CARB OBDII document that can be used to satisfy 
the federal OBD requirements.\48\
---------------------------------------------------------------------------

    \48\ See 13 CCR 1968.2, released August 11, 2006, Docket 
ID EPA-HQ-OAR-2005-0047-0005.
---------------------------------------------------------------------------

I. How Do the Proposed Requirements Compare to California's?

    The California Air Resources Board (CARB) has its own OBD 
regulations for engines used in vehicles over 14,000 pounds GVWR.\49\ 
(13 CCR 1971.1) In August of 2004, EPA and CARB signed a memorandum of 
agreement to work together to develop a single, nationwide OBD program 
for engines used in vehicles over 14,000 pounds.\50\ We believe that, 
for the most part, we have been successful in doing so at least for the 
early years of implementation. Nonetheless, there are differences in 
some of the details contained within each regulation. These differences 
are summarized here and we request comment on all of these differences.
---------------------------------------------------------------------------

    \49\ 13 CCR 1971.1, Docket ID EPA-HQ-OAR-2005-0047-
0006.
    \50\ ``Memorandum of Agreement: On-road Heavy-duty Diagnostic 
Regulation Development,'' signed by Chet France, U.S. EPA, and Tom 
Cackette, California ARB, August 11, 2004, Docket ID EPA-
HQ-OAR-2005-0047-0002.
---------------------------------------------------------------------------

    The first difference is that the CARB regulation contains some more 
stringent thresholds beginning in the 2013 timeframe for some engines 
and 2016 for all engines. Specifically, CARB's PM threshold for diesel 
particulate filters (DPF) and exhaust gas sensors downstream of 
aftertreatment devices, and their NOX threshold for 
NOX aftertreatment devices and exhaust gas sensors 
downstream of aftertreatment devices, become more stringent in 2013 for 
some engines and 2016 for all. We are not proposing these more 
stringent thresholds--our proposed thresholds are shown in Table II.B-
1. At this time, EPA is not in a position to propose these more 
stringent OBD thresholds for the national program. The industry 
believes that CARB's more stringent NOX and PM thresholds 
for 2013 and 2016 are not technically feasible. EPA is reviewing these 
longer term OBD thresholds, but at this time we have not made a 
decision regarding the feasibility and the appropriateness of these 
longer term thresholds. Because these thresholds do not take effect 
until model year 2013 at the earliest, we do not believe it is 
necessary to make such a determination in this rulemaking. It would be 
our intention to monitor the progress made towards complying with the 
2010 thresholds contained in today's proposal and potentially revisit 
the appropriateness of more stringent OBD thresholds for model year 
2013 and later in the future. CARB has made commitments to review their 
HD OBD program every two years and they can consider making changes to 
their long-term program during this biennial review process. EPA's 
regulatory development process does not lend itself to making updates 
every two years because the Federal rulemaking process tends to be 
lengthier than CARB's. As mentioned above, we intend to monitor the 
CARB long-term thresholds during the coming years, and if we determine 
that more stringent thresholds are appropriate, we would consider 
changing our thresholds to include the more stringent thresholds 
through a notice and comment rulemaking process.
    CARB also has some slightly different certification demonstration 
requirements in the 2011 and 2012 model years. They are requiring 
demonstration testing of the child ratings from the 2010 model year 
certified engine family for 2011 and 2012 model year certification. As 
Table II.B-1 shows, we are not requiring such demonstration testing in 
the 2011 and 2012 model years provided the child ratings meet the 
requirements of certification carry-over. Further, CARB is requiring 
that one engine rating from one to three engine families undergo full 
certification demonstration testing in the 2013 model year and every 
model year thereafter. In contrast, EPA is requiring that one to three 
engine ratings be fully demonstrated in the 2013 model year and then 
carry-over through the 2015 model year (again, provided the engine 
ratings meet the requirements of certification carry-over). In 2016 and 
subsequent model years, EPA would require that one to three engine 
ratings be fully demonstrated on an ``as needed'' basis. In the same 
vein, our evaluation protocol associated with certification 
demonstration testing, as discussed in section VIII.C, requires less 
testing than is required in CARB's regulation.
    Our OBD requirements for over 14,000 pounds do not contain any 
provisions to monitor control strategies associated with idle emission 
control strategies because EPA does not have currently any regulatory 
requirements that specifically target idle emissions control 
strategies.\51\ We are not proposing a provision to charge fees 
associated with OBD deficiencies as CARB does. We are also not 
proposing provisions for ``retroactive deficiencies'' as CARB has. Our 
deficiency provisions along with our misbuild and other in-use 
enforcement programs accomplish the same thing. Deficiencies are 
discussed in section VIII.D.\52\
---------------------------------------------------------------------------

    \51\ Note that, by idle emission control strategies we mean 
strategies that, for example, shut down the engine after 10 minutes 
of constant idle. We do not mean strategies that control emissions 
during engine idles that occur at stop lights or in congested 
traffic.
    \52\ See also proposed Sec.  86.010-18(n).
---------------------------------------------------------------------------

    For diesel engines used in heavy-duty vehicles under 14,000 pounds, 
our proposed OBD requirements are in line with those recently proposed 
by CARB.\53\ Our proposed requirements are also in line--both the 
technical aspects and the implementation timing aspects--with our 
proposed requirements for over 14,000 pound diesel applications. We are 
also proposing diesel vehicle-based OBD requirements in line with the 
proposed diesel engine-based requirements. In contrast, CARB does not 
have diesel thresholds in terms of ``grams per mile'' specified in 
their regulation for the 8,500 to 14,000 pound range.
---------------------------------------------------------------------------

    \53\ See 13 CCR 1968.2, released August 11, 2006, Docket 
ID EPA-HQ-OAR-2005-0047-0005.
---------------------------------------------------------------------------

    Specifically for gasoline engines meant for applications over 
14,000 pounds, our proposal differs from CARB's in that we are not 
requiring detection of catalysts that are less than 50 percent 
effective at converting emissions.\54\ We are not requiring this 
because we are relying on the emissions threshold of 1.75 times the 
applicable standard as a means of defining a catalyst system 
malfunction. We are also proposing some differences with respect to 
misfire monitoring. Most notably, we are not proposing a provision 
analogous

[[Page 3256]]

to CARB's provision that allows the Executive Officer to approve 
misfire monitor disablement or alternative malfunction criteria on a 
case by case basis.\55\ In general, we prefer to avoid having 
regulatory provisions that are implemented on a case by case basis. For 
similar reasons, we are also not proposing a provision analogous to 
CARB's provision that allows the Executive Officer to revise the 
orifice for evaporative leak detection if the most reliable monitoring 
strategy cannot detect the required orifice.\56\
---------------------------------------------------------------------------

    \54\ See 13 CCR 1971.1(f)(6.2.1)(B) and compare to proposed 
Sec.  86.010-18(h)(6)(ii).
    \55\ See 13 CCR 1971.1(f)(2.3.4)(D) and compare to proposed 
Sec.  86.010-18(h)(2)(iii)(D).
    \56\ See 13 CCR 1971.1(f)(7.2.3) and compare to proposed Sec.  
86.010-18(h)(7)(ii)(B) and (C).
---------------------------------------------------------------------------

III. Are the Proposed Monitoring Requirements Feasible?

    Some of the OBD monitoring strategies discussed here would be 
intrusive monitors that would result in very brief emissions increases, 
or spikes, for the sake of determining if certain emissions control 
components/systems are working properly during the remaining 99 percent 
or more of the engine's operation. While these emissions spikes are 
brief, and their levels cannot be meaningfully predicted or estimated, 
we are concerned about strategies that might give little concern to 
emissions during such spikes in favor of an easier monitor. We request 
comment on this issue--should such strategies be allowed or should such 
strategies be prohibited? If a commenter has the latter opinion, then 
suggestions should be provided for how the monitoring requirements 
should be changed to allow for a non-intrusive monitor--i.e., one that 
could run during normal operation or operation ``on the cycle''--that 
may not provide the monitoring capability nor the control expected by 
the requirements we are proposing.

A. Feasibility of the Monitoring Requirements for Diesel/Compression-
Ignition Engines

1. Fuel System Monitoring
a. Fuel Pressure Monitoring
    Manufacturers control fuel pressure by using a closed-loop feedback 
algorithm that allows them to increase or decrease fuel pressure until 
the fuel pressure sensor indicates they have achieved the desired fuel 
pressure. For the common-rail OBD systems certified in the under 14,000 
pound category, the manufacturers are monitoring the actual fuel system 
pressure sensed by a fuel rail pressure sensor, comparing it to the 
target fuel system pressure stored in a software table or calculated by 
an algorithm inside the onboard computer, and indicating a malfunction 
if the magnitude of the difference between these two exceeds an 
acceptable level. The error limits are established by engine 
dynamometer emission tests to ensure that a malfunction would be 
detected before emissions exceed the applicable thresholds.
    In cases where no fuel pressure error can generate a large enough 
emission increase to exceed the applicable thresholds, manufacturers 
are required to set the malfunction trigger at their fuel pressure 
control limits (e.g., when they reach a point where they can no longer 
increase or decrease fuel pressure to achieve the desired fuel 
pressure). This monitoring requirement has been demonstrated as 
technically feasible given that several under 14,000 pound diesels 
already meet this requirement. Further, the nature of a closed-loop 
algorithm is that such a system is inherently capable of being 
monitored because it simply requires analysis of the same closed-loop 
feedback parameter being used by the system for control purposes.
    Another promising technology is a pressure sensing glow plug. The 
glow plug is an electronic device in the cylinder of most diesel 
engines used to facilitate combustion during cold engine starting 
conditions. Glow plugs are being developed that incorporate a pressure 
sensor capable of detecting the quality of combustion within the 
cylinder.\57\ Pressure-sensing glow plugs provide feedback to the 
engine-management system that controls the timing and quantity of fuel 
injected into the cylinder. This feedback allows the engine electronics 
to adjust the injection characteristics so the engine avoids fuel-
mixture combinations that generate high levels of NOX. In 
this sense, a feedback loop is available that works like the oxygen 
sensor in a gasoline engine exhaust system. By measuring the quality of 
combustion, a determination can also be made about the quality of the 
fuel injection event--the pressure of fuel delivered, quantity of fuel 
delivered, timing of fuel delivered.
---------------------------------------------------------------------------

    \57\ ``Spotlight on Technology: Smart glowplugs may make Clean 
Diesels cost-effective Pressure-sensing units could let designers 
cut NOX aftertreatment,'' Tony Lewin, Automotive News, 
February 6, 2006.
---------------------------------------------------------------------------

b. Fuel Injection Quantity Monitoring
    Absent combustion sensors and/or pressure sensing glow plugs 
mentioned above, there is currently no feedback sensor indicating that 
the proper quantity of fuel has been injected. Therefore, injection 
quantity monitoring will be more difficult than pressure monitoring. 
Nonetheless, a manufacturer has identified a strategy currently being 
used that verifies the injection quantity under very specific engine 
operating conditions and appears to be capable of determining that the 
system is accurately delivering the desired fuel quantity. This 
strategy entails intrusive operation of the fuel injection system 
during a deceleration event where fuel injection is normally shut off 
(e.g., coasting or braking from a higher vehicle speed down to a low 
speed or a stop). During the deceleration, fuel injection to a single 
cylinder is turned back on to deliver a very small amount of fuel. 
Typically, the amount of fuel would be smaller than, or perhaps 
comparable to, the amount of fuel injected during a pilot or pre-
injection. If the fuel injection system is working correctly, that 
known injected fuel quantity will generate a known increase in 
fluctuations (accelerations) of the crankshaft that can be measured by 
the crankshaft position sensor. If too little fuel is delivered, the 
measured crankshaft acceleration will be smaller than expected. If too 
much fuel is delivered, the measured crankshaft acceleration will be 
larger than expected. This process can even be used to ``balance'' out 
each cylinder or correct for system tolerances or deterioration by 
modifying the commanded injection quantity until it produces the 
desired crankshaft acceleration and applying a correction or adaptive 
term to that cylinder's future injections. Each cylinder can, in turn, 
be cycled through this process and a separate analysis can be made for 
the performance of the fuel injection system for each cylinder. Even if 
this procedure would require only one cylinder be tested per revolution 
(to eliminate any change in engine operation or output that would be 
noticeable to the driver) and require each cylinder to be tested on 
four separate revolutions, this process would only take two seconds for 
a six cylinder engine decelerating through 1500 rpm.
    The crankshaft position sensor is commonly used to identify the 
precise position of the piston relative to the intake and exhaust 
valves to allow for very accurate fuel injection timing control and, as 
such, there exists sufficient resolution and data sampling within the 
onboard computer to enable such measurement of crankshaft 
accelerations. Further, in addition to the current use of this strategy 
in an under 14,000 pound diesel application, a nearly identical 
crankshaft fluctuation technique has been used since 1997 on under 
14,000 pound diesel engines

[[Page 3257]]

during idle conditions to determine if individual cylinders are 
misfiring.
    Another technique that may be used to achieve the same monitoring 
capability is some variation on the current cylinder balance tests used 
by many manufacturers to improve idle quality. In such strategies, 
fueling to individual cylinders is increased, decreased, or shut off to 
determine if the cylinder is contributing an equal share to the output 
of the engine. This strategy again relies on changes in crankshaft/
engine speed to measure the individual cylinder's contribution relative 
to known good values and/or the other cylinders. Such an approach seems 
viable to determine whether the fuel injection quantity is correct for 
each cylinder, but it has the disadvantage of not necessarily being 
able to verify whether the system is able to deliver small amounts of 
fuel precisely (such as those commanded during a pilot injection).
    One other approach that has been mentioned but not investigated 
thoroughly is the use of a wide-range air-fuel (A/F) sensor in the 
exhaust to confirm fuel injection quantity. The A/F sensor output could 
be compared to the measured air going into the engine and calculated 
fuel quantity injected to see if the two agree. Differences in the 
comparison may allow for the identification of incorrect fuel injection 
quantity.
c. Fuel Injection Timing Monitoring
    In the same manner as described for quantity monitoring, we believe 
that fuel injection timing could be verified. By monitoring the 
crankshaft speed fluctuation and, most notably, the time at which such 
fluctuation begins, ends, or reaches a peak, the OBD system could 
compare the time to the commanded fuel injection timing point and 
verify that the crankcase fluctuation occurred within an acceptable 
time delay relative to the commanded fuel injection. If the system was 
working improperly and actual fuel injection was delayed relative to 
when it was commanded, the corresponding crankshaft speed fluctuation 
would also be delayed and would result in a longer than acceptable time 
period between commanded fuel injection timing and crankshaft speed 
fluctuation. A more detailed discussion of this possible monitoring 
method is presented in the technical support document contained in the 
docket.\58\
---------------------------------------------------------------------------

    \58\ Draft Technical Support Document, HDOBD NPRM, EPA420-D-06-
006, Docket ID EPA-HQ-OAR-2005-0047-0008.
---------------------------------------------------------------------------

    Another possible monitoring method that has been mentioned but not 
investigated thoroughly would be to look for an electrical feedback 
signal from the injector to the computer to confirm when the injection 
occurred. Such a technique would likely use an inductive signature to 
identify exactly when an injector opened or closed and verify that it 
was at the expected timing. We expect that further investigation would 
be needed to confirm that such a monitoring technique would be 
sufficient to verify fuel injection timing.
d. Fuel System Feedback Control Monitoring
    The conditions necessary for feedback control (i.e., the feedback 
enable criteria) are defined as part of the control strategy in the 
engine computer. The feedback enable criteria are typically based on 
minimum conditions necessary for reliable and stable feedback control. 
When the manufacturer is designing and calibrating the OBD system, the 
manufacturer would determine, for the range of in-use operating 
conditions, the time needed to satisfy these feedback enable criteria 
on a properly functioning engine. In-use, the OBD system would evaluate 
the time needed for these conditions to be satisfied following an 
engine start, compare that to normal behavior for the system, and 
indicate a malfunction when the time exceeds a specified value (i.e., 
the malfunction criterion). For example, fuel pressure feedback control 
may be calibrated to begin once fuel system pressure has reached a 
minimum specified value. In a properly functioning system, pressure 
builds in the system during engine cranking and shortly after starting 
and the pressure enable criterion are reached within a few seconds. 
However, in a malfunctioning system (e.g., due to a faulty low-pressure 
fuel pump), it may take a significantly longer time to reach the 
feedback enable pressure. A malfunction would be indicated when the 
actual time to reach feedback enable pressure exceeds the malfunction 
criterion.
    Malfunctions that cause open-loop or default operation can be 
readily detected as well. As discussed above, the feedback enable 
criteria are clearly defined in the computer and are based on what is 
necessary for reliable control. After feedback control has begun, the 
OBD system can detect these criteria and indicate a malfunction when 
they are no longer being satisfied. For example, one enable criterion 
could be a pressure sensor reading within a certain range where the 
upper pressure limit would be based on the maximum pressure that could 
be generated in a properly functioning system. A malfunction would be 
indicated if the pressure sensor reading exceeded the upper limit which 
would cause the fuel system to go open loop.
    The feedback control system adjusts the base fuel strategy such 
that actual engine operating characteristics meet driver demand. But, 
the feedback control system has limits on how much adjustment can be 
made based, presumably, on the ability to maintain acceptable control. 
Like the feedback enable criteria, these control limits are defined in 
the computer. The OBD system would track the actual adjustments made by 
the control system and continuously compare them with the control 
limits. A malfunction would be indicated if the limits were reached.
    2. Engine Misfire Monitoring
    Diesel engines certified to the under 14,000 pound OBD requirements 
have been monitoring for misfire since the 1998 model year. The 
monitoring requirements we are proposing for over 14,000 pound 
applications are identical to the existing requirements for under 
14,000 pound applications for those engines that do not use combustion 
sensors.\59\ Therefore, technological feasibility has been demonstrated 
for these applications.
---------------------------------------------------------------------------

    \59\ Technically, the EPA OBD diesel misfire monitoring 
requirement for under 14,000 pound applications is to detect a lack 
of combustion whereas the California OBDII diesel misfire monitoring 
requirement is identical to what we are proposing for over 14,000 
pounds. Since all manufacturers to date are designing to the OBDII 
requirements, this statement is, for practical purposes, true.
---------------------------------------------------------------------------

    For engines that use combustion sensors, the misfire monitoring 
requirements are more stringent since the requirement calls for 
detection of malfunctions causing emissions to exceed the emissions 
thresholds. Nonetheless, detection on these engines should be straight 
forward since the combustion sensors would provide a direct measurement 
of combustion. Therefore, lack of combustion (i.e., misfire) could be 
measured directly. The combustion sensors are intended to measure 
various characteristics of a combustion event for feedback control. 
Such feedback is needed for engines that require very precise air and 
fuel metering controls such as would be required for homogeneous charge 
compression ignition (HCCI) engine. Accordingly, the resolution of 
sensors having that capability is well beyond what would be needed to 
detect a complete lack of combustion.

[[Page 3258]]

3. Exhaust Gas Recirculation (EGR) Monitoring
a. EGR Low Flow/High Flow Monitoring
    Typically, the EGR control system determines a desired EGR flow 
rate based on the engine operating conditions such as engine speed and 
engine load. The desired EGR flow rates, and the corresponding EGR 
valve positions needed to achieve the desired flow rates, are 
established when the manufacturer designs and calibrates the EGR 
system. Once established, manufacturers store the desired EGR flow 
rate/valve position in a lookup table in the onboard computer. During 
operation, the onboard computer commands the EGR valve to the position 
necessary to achieve the desired flow--i.e., the commanded EGR flow. 
The onboard computer then calculates or directly measures both the 
fresh air charge (fresh air intake) and total intake charge. The 
difference between the total intake charge and fresh air intake is the 
actual EGR flow. The closed-loop control system continuously adjusts 
the EGR valve position until the actual EGR flow equals the desired EGR 
flow.
    Such closed-loop control strategies and their associated OBD 
monitoring strategies are used on many existing gasoline and diesel 
vehicles under 14,000 pounds. The OBD system evaluates the difference 
(i.e., error) between the look-up value--i.e., the desired flow rate--
and the final commanded value needed to achieve the desired flow rate. 
Typically, as the feedback parameter or learned offset increases, there 
is an attendant increase in emissions. A correlation can be made 
between feedback adjustment and emissions. When the error exceeds a 
specific threshold, a malfunction would be indicated. This type of 
monitoring strategy could be used to detect both high and low flow 
malfunctions.
    While the closed-loop control strategy described above is effective 
in measuring and controlling EGR flow, some manufacturers are currently 
investigating the use of a second control loop based on an air-fuel 
ratio (A/F) sensor (also known as wide-range oxygen sensors or linear 
oxygen sensors) to further improve EGR control and emissions. With this 
second control loop, the desired air-fuel ratio is calculated based on 
engine operating conditions (i.e., intake airflow, commanded EGR flow 
and commanded fuel). The calculated air-fuel ratio is compared to the 
air-fuel ratio from the A/F sensor and refinements can be made to the 
EGR and airflow rates--i.e., the control can be ``trimmed''--to achieve 
the desired rates. On systems that use the second control loop, flow 
rate malfunctions could also be detected using the feedback information 
from the A/F sensor and by applying a similar monitoring strategy as 
discussed above for the primary EGR control loop.
    We are also proposing that two leaking EGR valve failure modes be 
detected. One type is the failure of the valve to seal when in the 
closed position. For example, if the valve or seating surface is 
eroded, the valve could close and seat, yet still allow some flow 
across the valve. A flow check is necessary to detect a malfunctioning 
valve that closes properly but still leaks. EGR flow--total intake 
charge minus fresh air charge--could be calculated using the monitoring 
strategy described above for high and low flow malfunctions. With the 
valve closed, a malfunction would be indicated when flow exceeds 
unacceptable levels. Or, some cooled EGR systems will incorporate an 
EGR temperature sensor that could be used to detect a leaking EGR valve 
by reacting to the presence of hot exhaust gases when none should be 
present. A leaking valve can also be caused by failure of the valve to 
close/seat. For example, carbon deposits on the valve or seat could 
prevent the valve from closing fully. The flow check described above 
could detect failure of the valve to close/seat, but this approach 
would require a repair technician to further diagnose whether the 
problem is a sealing or seating problem. Such a failure of the valve to 
close/seat could be more specifically monitored by closing the valve 
and checking the zero position of the valve with a position sensor. If 
the valve position is out of the acceptable range for a closed valve, a 
malfunction would be indicated. This type of zero position sensor check 
is commonly used to verify the closed position of valves/actuators used 
in gasoline OBD systems (e.g. gasoline EGR valves, electronic throttle) 
and should be feasible for diesel EGR valves.
b. EGR Slow Response Monitoring
    While the flow rate monitor discussed above would evaluate the 
ability of the EGR system to achieve a commanded flow rate under 
relatively steady state conditions, the EGR slow response monitor would 
evaluate the ability of the EGR system to modulate (i.e., increase and 
decrease) EGR flow as engine operating conditions and, consequently, 
commanded EGR rates change. Specifically, as engine operating 
conditions and commanded EGR flow rates change, the monitor would 
evaluate the time it takes for the EGR control system to achieve the 
commanded change in EGR flow. This monitor could evaluate EGR response 
passively during transient engine operating conditions encountered 
during in-use operation. The monitor could also evaluate EGR response 
intrusively by commanding a change in EGR flow under a steady state 
engine operating condition and measuring the time it takes to achieve 
the new EGR flow rate. Similar passive and intrusive strategies have 
been developed for variable valve control and/or timing (VVT) 
monitoring on vehicles under 14,000 pounds.
c. EGR Feedback Control Monitoring
    Monitoring of EGR feedback control could be performed using 
analogous strategies to those discussed in Section III.A.1 for 
monitoring of fuel system feedback control.
d. EGR Cooling System Monitoring
    Some diesel engine manufacturers currently use exhaust gas 
temperature sensors as an input to their EGR control systems. On such 
systems--EGR temperature--which is measured downstream of the EGR 
cooler--could be used to monitor the effectiveness of the EGR cooler. 
For a given engine operating condition (e.g., a steady speed/load that 
generates a known exhaust mass flow and exhaust temperature to the EGR 
cooler), EGR temperature will increase as the performance of the EGR 
cooling system decreases. During the OBD calibration process, 
manufacturers could develop a correlation between increased EGR 
temperatures and cooling system performance (i.e., increased 
emissions). The EGR cooling system monitor would use such a correlation 
and indicate a malfunction when the EGR temperature increases to the 
level that would cause emissions to exceed the emissions thresholds.
    While we anticipate that most, if not all, manufacturers will use 
EGR temperature sensors to meet future emissions standards, EGR cooling 
system monitoring may be feasible without such a temperature sensor. 
The monitor could be done using the intake manifold temperature (IMT) 
sensor by looking at the change in IMT (i.e., ``delta'' IMT) with EGR 
turned on and EGR turned off (IMT would be higher with EGR turned on). 
If there is significant cooling capacity with a normally functioning 
EGR cooling system, there would likely be a significant difference in 
IMT with EGR turned on versus turned off. Delta IMT could be correlated 
to decreased EGR cooling system performance and increased emissions.

[[Page 3259]]

4. Turbo Boost Control System Monitoring
a. Turbo Underboost/Overboost Monitoring
    To monitor boost control systems, manufacturers are expected to 
look at the difference between the actual pressure sensor reading (or 
calculation thereof) and the desired/target boost pressure. If the 
error between the two is too large or persists for too long, a 
malfunction would be indicated. Manufacturers would need to calibrate 
the size of error and/or error duration to ensure robust malfunction 
detection occurs before the emissions thresholds are exceeded. Given 
that the purpose of a closed-loop control system with a feedback sensor 
is to measure continuously the difference between actual and desired 
boost pressure, the control system is already monitoring that 
difference and attempting to minimize it. As such, a monitoring 
requirement to indicate a malfunction when the difference gets large 
enough such that it can no longer achieve the desired boost is 
essentially an extension of the existing control strategy.
    To monitor for malfunction or deterioration of the boost pressure 
sensors, manufacturers could validate sensor readings against other 
sensors present on the vehicle or against ambient conditions. For 
example, at initial key-on before the engine is running, the boost 
pressure sensor should read ambient pressure. If the vehicle is 
equipped with a barometric pressure sensor, the two sensors could be 
compared and a malfunction indicated when the two readings differ 
beyond the specific tolerances. A more crude rationality check of the 
boost pressure sensor could be accomplished by verifying that the 
pressure reading is within reasonable atmospheric limits for the 
conditions the vehicle will be subjected to.
b. VGT Slow Response Monitoring
    The VGT slow response monitor would evaluate the ability of the VGT 
system to modulate (i.e., increase and decrease) boost pressure as 
engine operating conditions and, consequently, commanded boost pressure 
changes. Specifically, as engine operating conditions and commanded 
boost pressures change, the monitor would evaluate the time it takes 
for the VGT control system to achieve the commanded change in boost 
pressure. This monitor could evaluate VGT response passively during 
transient engine operating conditions encountered during in-use 
operation. The monitor could also evaluate VGT response intrusively by 
commanding a change in boost pressure under a steady state engine 
operating condition and measuring the time it takes to achieve the new 
boost pressure.
    Rationality monitoring of VGT position sensors could be 
accomplished by comparing the measured sensor value to expected values 
for the given engine speed and load conditions. For example, at high 
engine speeds and loads, the position sensor should indicate that the 
VGT position is opened more than would be expected at low engine speeds 
and loads. Such rationality checks would need to be two-sided (i.e., 
position sensors should be checked for appropriate readings at both 
high and low engine speed/load operating conditions.
c. Turbo Boost Feedback Control Monitoring
    Monitoring of boost pressure feedback control could be performed 
using analogous strategies to those discussed for fuel system feedback 
control monitoring in Section III.A.1.
d. Charge Air Undercooling Monitoring
    We expect that most engines will make use of a temperature sensor 
downstream of the charge air cooler to protect against overcooling 
conditions that could cause excessive condensation, and to prevent 
undercooling that could result in loss of performance. A comparison of 
the actual charge air temperature to the expected, or design, 
temperature would indicate any errors that might be occurring. 
Manufacturers could correlate that error to an emissions impact and, 
when the error reached a level such that emissions would exceed the 
emissions thresholds, a malfunction would be indicated.
5. Non-Methane Hydrocarbon (NMHC) Converting Catalyst Monitoring
a. NMHC Converting Catalyst Conversion Efficiency Monitoring
    Monitoring of the NMHC converting catalyst, or diesel oxidation 
catalyst (DOC), could be performed similar to three-way catalyst 
monitoring on gasoline engines. Three-way catalyst monitoring uses the 
concept that catalyst's oxygen storage capacity correlates well with 
its hydrocarbon conversion efficiency. Oxygen sensors located upstream 
and downstream of the catalyst can be used to determine when its oxygen 
storage capacity--and, hence, its conversion efficiency--has 
deteriorated below a predetermined level.
    Determining the oxygen storage capacity would require lean air-fuel 
(A/F) operation followed by rich A/F operation or vice-versa during the 
catalyst monitoring event. Since a diesel engine normally operates lean 
of stoichiometry, lean A/F operation would be normal operation. 
However, rich A/F operation would have to be commanded intrusively when 
the catalyst monitor is active. The rich A/F operation could be 
achieved by injecting some fuel late enough in the four stroke process 
(i.e., late injection) that the raw fuel would not combust in-cylinder. 
Rich A/F operation could also be achieved using an in-exhaust fuel 
injector upstream of the catalyst. During normal lean operation, the 
catalyst would become saturated with stored oxygen. As a result, both 
the front and rear oxygen sensors should be reading lean. When rich A/F 
operation initiates, the front oxygen sensor would switch immediately 
to a ``rich'' indication. For a short time, the rear oxygen sensor 
should continue to read ``lean'' until such time as the stored oxygen 
in the catalyst is consumed by the rich fuel mixture in the exhaust and 
the rear oxygen sensor would read ``rich.'' As the catalyst 
deteriorates, the delay time between the front and rear oxygen sensors 
switching from their normal lean state to a rich state would become 
progressively smaller because the deteriorated catalyst would have less 
oxygen storage capacity. Thus, by comparing the time difference between 
the responses of the front and rear oxygen sensors to the lean-to-rich 
or rich-to-lean A/F changes, the performance of the catalyst could be 
estimated. Although this discussion suggests the use of conventional 
oxygen sensors, these sensors could be substituted with A/F sensors 
which would also provide for additional engine control benefits such as 
EGR trimming and fuel trimming.
    If a malfunction of the catalyst cannot cause emissions to exceed 
the emissions thresholds, then only a functional monitor would be 
required. A functional monitor could be done using temperature sensors. 
A functioning oxidation catalyst would be expected to provide some 
level of exotherm when it oxidizes HC and CO. The temperature of the 
catalyst could be measured by placing one or more temperature sensors 
at or near the catalyst. However, depending on the nominal conversion 
efficiency of the catalyst and the duty cycle of the vehicle, the 
exotherm may be difficult to discern from the inlet exhaust 
temperatures. To add robustness to the monitor, the functional monitor 
would need to be conducted during predetermined

[[Page 3260]]

operating conditions where the amount of HC and CO entering the 
catalyst could be known. This may require an intrusive monitor that 
actively forces the fueling strategy richer (e.g., through late or post 
injection) than normal for a short period of time. If the measured 
exotherm does not exceed a predetermined amount that only a properly-
working catalyst could achieve, a malfunction would be indicated. As 
noted, such an approach would require a brief period of commanded rich 
operation that would result in a very brief HC and perhaps a PM 
emissions spike.
b. Other Aftertreatment Assistance Function Monitoring
    A functional monitor should be sufficient for monitoring the 
oxidation catalyst's ability to fulfill aftertreatment assistance 
functions such as generating an exotherm for DPF regeneration or 
providing a proper feedgas for SCR or NOX adsorbers. We 
would expect that manufacturers would use the exotherm approach 
mentioned above either to measure directly for the proper exotherm or 
to correlate indirectly for the proper feedgas. For catalysts upstream 
of a DPF, we expect that this monitoring would be conducted during an 
active or forced regeneration event.\60\ For catalysts downstream of 
the DPF, we expect that manufacturers would have to add fuel 
intrusively (either in-exhaust or through in-cylinder post-injection) 
to create a sufficient exotherm to distinguish malfunctioning from 
properly operating catalysts.
---------------------------------------------------------------------------

    \60\ An active or forced regeneration would be those 
regeneration events that are initiated via a driver selectable 
switch or activator and/or those initiated by computer software.
---------------------------------------------------------------------------

6. Selective Catalytic Reduction (SCR) and NOX Conversion 
Catalyst Monitoring
a. SCR and NOX Catalyst Conversion Efficiency Monitoring
    We would expect manufacturers to use NOX sensors to 
monitor a lean NOX catalyst. NOX sensors placed 
upstream and downstream of the lean NOX catalyst could be 
used to determine directly the NOX conversion efficiency. 
Manufacturers could potentially use a single NOX sensor 
placed downstream of the catalyst to measure catalyst-out 
NOX emissions. This would have to be done within a tightly 
controlled engine operation window where engine-out NOX 
emissions (i.e., NOX emissions at the lean NOX 
catalyst inlet) performance is relatively stable and could be estimated 
reliably. Within this engine operation window, catalyst-out 
measurements could be compared to the expected engine-out 
NOX emissions and a catalyst conversion efficiency could be 
calculated. Should the calculated conversion efficiency be insufficient 
to maintain emissions below the emissions thresholds, a malfunctioning 
or deteriorated lean NOX catalyst would be indicated. If 
both an upstream and downstream NOX sensor are used for 
monitoring, the upstream sensor could be used to improve the overall 
effectiveness of the catalyst by precisely controlling the air-fuel 
ratio in the exhaust to the levels where the catalyst is most 
effective.
    For monitoring the SCR catalyst, care must be taken to account for 
the cross sensitivity of NOX sensors to ammonia 
(NH3). Current NOX sensor technology tends to 
have such a cross-sensitivity to ammonia in that as much as 65 percent 
of ammonia can be read as NOX.\61\ However, urea SCR 
feedback control studies have shown that the NH3 
interference signal is discernable from the NOX signal and 
can, in effect, allow the design of a better feedback control loop than 
a NOX sensor that doesn't have any NH3 cross-
sensitivity. In one study, a signal conditioning method was developed 
that resulted in a linear output for both NH3 and 
NOX from the NOX sensor downstream of the 
catalyst.\62\ Monitoring of the catalyst can be done by using the same 
NOX sensors that are used for SCR control. When the SCR 
catalyst is functioning properly, the upstream sensor should read 
``high'' for high NOX levels while the downstream sensor 
should read ``low'' for low NOX and low ammonia levels. With 
a deteriorated SCR catalyst, the downstream sensor should read similar 
or higher values as the upstream sensor (i.e., high NOX and 
high ammonia levels) since the NOX reduction capability of 
the catalyst has diminished. Therefore, a malfunctioning SCR catalyst 
could be detected when the downstream sensor output is near to or 
greater than the upstream sensor output. A similar monitoring approach 
could be used if a manufacturer models upstream NOX 
emissions instead of using an upstream NOX sensor. In this 
case, the comparison would be made between the modeled upstream 
NOX value and the downstream sensor value.
---------------------------------------------------------------------------

    \61\ Schaer, C.M., Onder, C.H., Geering, H.P., and Elsener, M., 
``Control of a Urea SCR Catalytic Converter System for a Mobile 
Heavy-Duty Diesel Engine,'' SAE Paper 2003-01-0776 which may be 
obtained from Society of Automotive Engineers International, 400 
Commonwealth Dr., Warrendale, PA, 15096-0001.
    \62\ Ibid.
---------------------------------------------------------------------------

    Manufacturers have expressed concern over both the sensitivity and 
the durability of NOX sensors. They are concerned that 
NOX sensors will not have the necessary sensitivity to 
detect NOX at the low levels that will exist downstream of 
the NOX catalyst. They are also concerned that 
NOX sensors will not be durable enough to last the full 
useful life of big diesel trucks. We have researched NOX 
sensors--the current state of development and future expectations--and 
summarized our findings in the technical support document in the docket 
for this rule.\63\ Some of our findings are summarized here.
---------------------------------------------------------------------------

    \63\ Draft Technical Support Document, HDOBD NPRM, EPA420-D-06-
006, Docket ID EPA-HQ-OAR-2005-0047-0008.
---------------------------------------------------------------------------

    Regarding NOX sensor sensitivity, we expect that 2010 
and later model year engines will have average tailpipe NOX 
emissions in the 0 to 50 ppm range. Current NOX sensors have 
an accuracy of 10 ppm in the 0 to 100 ppm range. This means 
that current NOX sensors should be able to detect 
NOX emissions that exceed the standard by two to three times 
the 2010 limit.\64\ This should allow for compliance with our proposed 
threshold which is effectively 2.5 times the 2010 limit. Further, we 
expect that NOX sensors in the 0 to 100 ppm range with 
5 ppm accuracy will be available by the middle of 2006. 
Regarding durability, improvements are being made and a test program is 
currently underway with the intent of aging several NOX 
sensors placed at various exhaust system locations out to 6,000 hours 
(roughly equivalent to 360,000 miles). Results after 2,000 hours of 
aging are promising and results after 4,000 hours of aging are 
currently being analyzed.\65\
---------------------------------------------------------------------------

    \64\ Ibid.
    \65\ Ibid.
---------------------------------------------------------------------------

b. SCR and NOX Catalyst Active/Intrusive Reductant Injection 
System Monitoring
    If an active catalyst system is used--i.e., one that relies on 
injection of a reductant upstream of the catalyst to assist in 
emissions conversion--manufacturers would be required to monitor the 
mechanism for adding the fuel reductant. In the active catalyst system, 
a temperature sensor is expected to be placed near or at the catalyst 
to determine when the catalyst temperature is high enough to convert 
emissions. Because NOX catalyst systems, especially lean 
NOX catalyst systems, tend to have a narrow temperature 
range where they are most effective, adding reductant when the catalyst 
temperature is not sufficiently high would waste reductant. If fuel is

[[Page 3261]]

used as the reductant, this would adversely affect fuel economy without 
a corresponding reduction in emissions levels. Therefore, a temperature 
sensor is expected to be placed in the exhaust near or at the catalyst 
to help determine when reductant injection should occur. This same 
sensor could be used to determine if an exotherm resulted following 
reductant injection. The lack of an exotherm would indicate a 
malfunction of the reductant delivery system.
    Alternatively, any NOX sensors used to monitor 
conversion efficiency could be used to determine if reductant injection 
has occurred. NOX sensors are also oxygen sensors so they 
could be used to determine the air-fuel ratio in the exhaust stream 
which would allow for verification of reductant injection into the 
exhaust. Further, with a properly functioning injector, the downstream 
NOX sensor should see a change from high NOX 
levels to low NOX levels. In contrast, a lack of reductant 
injection would result in continuously high NOX levels at 
the downstream NOX sensor. Therefore, a malfunctioning 
injector could be indicated when the downstream NOX sensor 
continues to measure high NOX after an injection event has 
been commanded.
    Reductant level monitoring could also be conducted by using the 
existing NOX sensors that are used for control purposes. 
Specifically, the downstream NOX sensor can be used to 
determine if the reductant tank no longer has sufficient reductant 
available. Similar to the fuel reductant injection functionality 
monitor described above, when the reductant tank has a sufficient 
reductant quantity and the injection system is working properly, the 
downstream NOX sensor should see a change from high 
NOX levels to low NOX levels. If the 
NOX levels remain constant both before and after reductant 
injection, then the reductant was not properly delivered and either the 
injection system is malfunctioning or there is no longer sufficient 
reductant available in the reductant tank. Alternatively, reductant 
level monitoring could be conducted by using a dedicated ``float'' type 
level sensor similar to the ones used in fuel tanks. Some manufacturers 
may prefer using a dedicated reductant level sensor in the reductant 
tank to inform the vehicle operator of current reductant levels via a 
gauge on the instrument panel. If such a sensor is used by the 
manufacturer for operator convenience, it could also be used to monitor 
the reductant level in the tank.
    Monitoring the reductant itself--whether it be the wrong reductant 
or a poor quality reductant--could also be conducted using the 
NOX sensors used for control purposes. If an improper 
reductant is injected, the NOX catalyst system would not 
function properly. Therefore, NOX emissions downstream from 
the catalyst would remain high both before and after injection. The 
downstream NOX sensor would see the high NOX 
levels after injection and a malfunction would be indicated. If the 
reductant tank level sensor indicated sufficient levels for injection 
and decreasing levels following injections (which would mean the 
injection system was working), then the probable cause of the 
malfunction would be the reductant itself. For urea SCR systems, 
another possible means of monitoring the reductant itself would be to 
use a urea quality sensor in the urea tank. First generation sensors 
show promise at verifying that urea is indeed in the tank, rather than 
water or some other fluid, and that the urea concentration is within 
the needed range (i.e., not diluted with water or some other fluid). 
The sensor could also be used in place of a urea level sensor. By 2010, 
we would expect subsequent generation sensors to provide even better 
capability.\66\
---------------------------------------------------------------------------

    \66\ Crawford, John M., Mitsui Mining & Smelting Co., Ltd., 
presentation to EPA, October 2006, Docket ID EPA-HQ-OAR-
2005-0047-0007.
---------------------------------------------------------------------------

c. SCR and NOX Catalyst Feedback Control Monitoring
    Monitoring of feedback control could be performed using analogous 
strategies to those discussed for fuel system feedback control 
monitoring in Section III.A.1.
7. NOX Adsorber Monitoring
a. NOX Adsorber Capability Monitoring
    We expect that either NOX sensors or A/F sensors along 
with a temperature sensor will be used to provide the feedback 
necessary to control the NOX adsorber system. These same 
sensors could also be used to monitor the NOX adsorber 
system's capability. The use of NOX sensors placed upstream 
and downstream of the adsorber system would allow the system's 
NOX reduction performance to be continuously monitored. For 
example, the upstream NOX sensor on a properly functioning 
adsorber system operating with lean fuel mixtures, will read high 
NOX levels while the downstream NOX sensor should 
read low NOX levels. With a deteriorated NOX 
adsorber system, the upstream NOX levels will continue to be 
high while the downstream NOX levels will also be high. 
Therefore, a malfunction of the system can be detected by comparing the 
NOX levels measured by the downstream NOX sensor 
versus the upstream sensor.
    The possibility exists that an upstream NOX sensor will 
not be used for NOX adsorber control. Manufacturers may 
choose to model engine-out NOX levels--based on engine 
operating parameters such as engine speed, fuel injection quantity and 
timing, EGR flow rate--thereby eliminating the need for the upstream 
NOX sensor. In this case, we believe that monitoring of the 
system could be conducted using A/F sensors in place of NOX 
sensors.\67\ During lean engine operation with a properly operating 
NOX adsorber system, both the upstream and downstream A/F 
sensors would indicate lean mixtures. When the exhaust gas is 
intrusively commanded rich to regenerate the NOX adsorber, 
the upstream A/F sensor would quickly indicate a rich mixture while the 
downstream sensor should continue to see a lean mixture due to the 
chemical reaction of the reducing agents with NOX and oxygen 
stored on the adsorber. Once all of the stored NOX and 
oxygen has been released, the reducing agents in the exhaust would 
cause the downstream A/F sensor to indicate a rich reading. The more 
NOX that is stored in the adsorber, the longer the delay 
between the rich indications from the upstream and downstream sensors. 
Thus, the time differential between the rich indications from the 
upstream and downstream A/F sensors is a gauge of the NOX 
storage capacity of the adsorber. This delay could be correlated to an 
emissions increase and the monitor could be calibrated to indicate a 
malfunction upon detecting an unacceptably short delay. In fact, Honda 
currently uses a similar approach to monitor the NOX 
adsorber on a 2003 model year gasoline vehicle which demonstrates the 
viability of the approach in a shorter lived application. We have 
studied A/F sensors and their durability with respect to longer lived 
diesel applications and our results are summarized in a report placed 
in the docket to this rule.\68\
---------------------------------------------------------------------------

    \67\ Ingram, G.A. and Surnilla, G., ``On-Line Estimation of 
Sulfation Levels in a Lean NOX Trap,'' SAE Paper 2002-01-
0731 may be obtained from Society of Automotive Engineers 
International, 400 Commonwealth Dr., Warrendale, PA 15096-0001.
    \68\ Draft Technical Support Document, HDOBD NPRM, EPA420-D-06-
006, Docket ID EPA-HQ-OAR-2005-0047-0008.

---------------------------------------------------------------------------

[[Page 3262]]

b. NOX Adsorber Active/Intrusive Reductant Injection System 
Monitoring
    The injection system used to achieve NOX regeneration of 
the NOX adsorber could also be monitored with A/F sensors. 
When the control system injects extra fuel to achieve a rich mixture, 
the upstream A/F sensor would respond to the change in fueling and 
could measure directly whether or not the proper amount of fuel had 
been injected. If manufacturers employ a NOX adsorber system 
design that uses only a single A/F sensor downstream of the adsorber, 
that downstream sensor could be used to monitor the performance of the 
injection system. As discussed above, the downstream sensor would 
switch from a lean reading to a rich reading when the stored 
NOX has been completely released and reduced. If the sensor 
switches too quickly after rich fueling is initiated, then either too 
much fuel has been injected or the adsorber itself has poor storage 
capability. Conversely, if the sensor takes too long to switch after 
rich fueling is initiated, it may be an indication that the adsorber 
has very good storage capability. However, excessive switch times 
(i.e., times that exceed the maximum storage capability of the 
adsorber) could be indicative of an injection system malfunction (i.e., 
insufficient fuel has been injected) or a sensor malfunction (i.e., the 
sensor has a slow response).
c. NOX Adsorber Feedback Control Monitoring
    Monitoring of feedback control could be performed using analogous 
strategies to those discussed for fuel system feedback control 
monitoring in Section III.A.1.
8. Diesel Particulate Filter (DPF) Monitoring
a. PM Filtering Performance Monitoring
    The PM filtering performance monitor is perhaps the monitor for 
which we have the most concern with respect to feasibility. Part of 
this concern stems from the difficulty in detecting the very low PM 
emissions levels required for 2007/2010 engines (i.e., 0.01 g/bhp-hr). 
While we have made changes to our test procedures that will allow for 
more accurate measurement of PM in the test cell, it is still very 
difficult to do. With today's proposal, we are expecting manufacturers 
to detect failures in the filtering performance of only a few times the 
actual standards. Success at doing so presents a very difficult 
challenge to manufacturers. Our concerns, in part, have led us to 
propose a different 2013 and later emissions threshold for this monitor 
than that proposed by ARB. This was discussed in more detail in section 
I.D.2.
    We anticipate that manufacturers can meet the proposed PM filtering 
monitor requirements without adding hardware other than that used for 
control purposes. We believe that the same pressure and temperature 
sensors that are used to control DPF regeneration will be used for OBD 
monitoring. For control purposes, manufacturers generally use a 
differential or delta pressure sensor placed across the DPF and at 
least one temperature sensor located near the DPF. The differential 
pressure sensor is expected to be used on DPF systems to prevent damage 
that could be caused by delayed or incomplete regeneration. Such 
conditions could lead to excessive temperatures and melting of the DPF 
substrate. When the differential pressure exceeds a predetermined 
level, a regeneration event would be initiated to burn the trapped PM.
    However, engine manufacturers have told us that differential 
pressure alone does not provide a robust indication of trapped PM in 
the DPF. For example, most if not all DPFs in the 2010 timeframe will 
be catalyzed DPFs that are designed to regenerate passively during most 
operation. Sometimes, conditions will not permit the passive 
regeneration and an active regeneration would have to be initiated. 
Relying solely on the differential pressure sensor to determine when an 
active regeneration event was necessary would not be sufficient. A low 
differential pressure could mean a low PM load and could also mean a 
leaking DPF substrate. A high differential pressure could mean a high 
PM load and could also mean a melted substrate. In the latter case, the 
system may continually attempt to regenerate the DPF despite a low PM 
load which would both waste fuel and increase HC emissions.
    As a result, manufacturers will probably use some sort of soot-
loading model to predict the PM load on the DPF as part of their 
regeneration strategy. Without a robust prediction, a regeneration 
event could be initiated too early (i.e., when too little PM was 
present which would be a waste of fuel and would increase HC emissions) 
or too late (i.e., when too much PM has been allowed to build and the 
regeneration event could cause a meltdown of the substrate). The model 
would estimate the PM load by tracking the difference between the 
modeled engine-out PM (i.e., the emissions that are being loaded on the 
DPF) and regenerated PM (i.e., the PM that is being burned off the DPF 
due to passive and/or active regenerations).
    Given this, we believe that a comprehensive and accurate soot-
loading model is also necessary for successful monitoring of DPF 
filtering performance. The model would predict the PM load on the DPF 
based on fuel consumption and engine operating conditions and would 
predict passively regenerated PM based on temperatures. This predicted 
PM load would be compared to the measured PM load taken from the 
differential pressure sensors. Differences would correspond to either a 
leaking substrate (i.e., predicted load greater than measured load) or 
melting of the substrate faceplate (i.e., measured load greater than 
predicted load).
    Nonetheless, much development remains to be done and success is not 
guaranteed. Manufacturers have noted that a melted substrate through 
which a large channel has opened could have differential pressure 
characteristics identical to a good substrate despite allowing most of 
the engine-out PM to flow directly through. We agree that this is a 
difficult failure mode and have proposed language that would allow 
certification of DPF monitors that are unable to detect it. Possibly, a 
temperature sensor in the DPF could detect the extreme temperatures 
capable of causing such a severe substrate melting. Upon detecting such 
a temperature, a regeneration event could be initiated to burn off any 
trapped PM. Following that event, the soot model would expect a certain 
increase in differential pressure based on modeled engine-out PM and 
passive regeneration characteristics. Presumably, the measured 
differential pressure profile would not match the predicted profile 
because most PM would be flowing straight through the melted channel. 
This same approach, or perhaps a simple temperature sensor, should 
quite easily be able to detect a missing substrate.
    Lastly, manufacturers have noted their concern that small 
differences in substrate crack size or location may generate large 
differences in tailpipe emission levels. They have also noted their 
lack of confidence that they will be able to reliably detect all leaks 
that would result in emissions exceeding the proposed thresholds. 
Accordingly, the manufacturers have suggested pursuing an alternate 
malfunction criterion independent of emission level. They have 
suggested criteria such as a percent of exhaust flow leakage or a 
specific leak or hole size that must be detected. We believe that 
pursuit of such alternate thresholds would not be appropriate at this 
time. Manufacturers have not yet completed work on initial widespread

[[Page 3263]]

implementation of DPFs for the 2007 model year. We expect that during 
the year or two following that implementation, substantial refinement 
and optimization will occur based on field experiences and that 
correlation of sensor readings to emissions levels will be possible for 
at least some DPF failure modes by the 2010 model year.
b. DPF Regeneration Monitoring
    Pressure sensing, in combination with the soot model, could also be 
used to determine if regeneration is functioning correctly. After a 
regeneration event, the differential pressure should drop significantly 
since the trapped PM has been removed. If it does not drop to within 
the soot model's predicted range after the regeneration event, either 
the regeneration did not function correctly or the filter could have 
excessive ash loading. Ash loading is a normal byproduct of engine 
operation (the ash loading is largely a function of oil consumption by 
the engine and the ash content of the engine oil). The ash builds up in 
the DPF and does not burnout as does the PM but rather must be removed 
or blown out of the DPF. Manufacturers are working with us to determine 
the necessary maintenance intervals at which this ash removal will 
occur. The soot model would have to account for ash buildup in the DPF 
with miles or hours of operation. Future engine oils will have lower 
ash content and have tighter quality control such that more accurate 
predictions of ash loading will be possible. By including ash loading 
in the soot model, we believe that its effects could be accounted for 
in the predicted differential pressure following a regeneration event.
    As stated, manufacturers are projected to make use of temperature 
sensors for regeneration control. These same sensors could also be used 
to monitor active regeneration of the filter. If excess temperatures 
are seen by the temperature sensor during active regeneration, the 
regeneration process can be stopped or slowed down to protect the 
filter. If an active regeneration event is initiated and there a 
temperature rise commensurate with the amount of trapped PM is not 
detected, the regeneration system is not working and a malfunction 
would be indicated.
c. DPF NMHC Conversion Efficiency Monitoring
    Given the stringency of the 2010 standards, we believe that 
manufactures may rely somewhat on the DPF to convert some of the HC 
emissions. The proposed requirement requires monitoring this function 
only if the system serves this function. We believe that, provided the 
filtering performance and regeneration system monitors have not 
detected any malfunctions, the NMHC conversion is probably working 
fine. Given the level of the threshold, and the expectation that the 
DPF will serve to control NMHC only marginally, we do not anticipate 
this monitor needing emissions correlation work. Instead, we expect 
that, with the DPF temperature sensor, it should be possible to infer 
adequate NMHC conversion by verifying an exotherm. Nonetheless, if a 
manufacturer relies so heavily on the DPF for NMHC conversion that its 
ability to convert could be compromised to the point of emissions 
exceeding the threshold, a more robust monitor may be required by 
correlating exotherm levels to NMHC impacts.
d. DPF Regeneration Feedback Control Monitoring
    Monitoring of DPF regeneration feedback control could be performed 
using analogous strategies to those discussed for fuel system feedback 
control monitoring in Section III.A.1.
9. Exhaust Gas Sensor Monitoring
    The under 14,000 pound OBD regulations have required oxygen sensor 
monitoring since the 1996 model year. Vehicles have been certified 
during that time meeting the requirements. The technological 
feasibility of monitoring oxygen sensors has been demonstrated. 
Additionally, A/F sensor monitoring has been required, manufacturers 
have complied, and the feasibility has been similarly demonstrated.
    NOX sensors are a recent technology and, as such, they 
are still being developed and improved. However, we would expect that 
manufacturers would design their upstream NOX sensor 
monitors to be similar the A/F sensor monitors used in under 14,000 
pound applications. Monitoring of downstream sensors may require 
modifications to existing A/F sensor strategies and/or new strategies. 
Since NOX sensors are projected to be used only for control 
and monitoring of aftertreatment systems that reduce NOX 
emissions (e.g., SCR systems), the OBD system would have to distinguish 
between deterioration of the aftertreatment system and the 
NOX sensor itself. As the aftertreatment deteriorates, 
NOX emissions downstream of the aftertreatment device will 
increase and, assuming there is no such deterioration in the 
NOX sensor, the NOX sensor will read these 
increasing NOX levels. As discussed in sections III.A.6 and 
III.A.7, the increased NOX levels can be the basis for 
monitoring the performance of the aftertreatment system. However, if 
the NOX sensor does deteriorate with the aftertreatment 
device (i.e., its response rate slows with mileage/operating hours), 
the sensor may not properly read the increasing NOX levels 
from the deteriorating aftertreatment system, and the aftertreatment 
monitor might conclude that the aftertreatment system is functioning 
properly. Similarly, the performance or level of deterioration of the 
NOX aftertreatment device could affect the results of the 
NOX sensor monitor. Therefore to achieve robust monitoring 
of aftertreatment and sensors, the OBD system has to distinguish 
between deterioration of the aftertreatment system and deterioration of 
the NOX sensor. To properly monitor the NOX 
sensor, the sensor monitor has to run under conditions where the 
aftertreatment performance can be quantified and compensated for or 
eliminated in the monitoring results.
    For example, the effects of the SCR performance could be eliminated 
by monitoring the NOX sensor under a steady-state operating 
condition during which engine-out NOX emissions were stable. 
Under a relatively steady-state condition, reductant injection could be 
``frozen'' (i.e., the reductant injection quantity could be held 
constant) which would also freeze the conversion efficiency of the SCR 
system. With SCR performance held constant, engine-out NOX 
emissions could be intrusively increased by a known amount (e.g., by 
reducing EGR flow or changing fuel injection timing and allowing the 
engine-out NOX model to determine the increase in 
emissions). The resulting increase in emissions would pass through the 
SCR catalyst unconverted, and the sensor response to the known increase 
in NOX concentrations could be measured and evaluated. This 
strategy could be used to detect both response malfunctions (i.e., the 
sensor reads the correct NOX concentration levels but the 
sensor reading does not change fast enough to keep up with changing 
exhaust NOX concentrations) and rationality malfunctions 
(i.e., the sensor reads the wrong NOX level). Rationality 
malfunctions could be detected by making sure the sensor reading 
changes by the same amount as the intrusive change in emissions. 
Lastly, the sensor response to decreasing NOX concentrations 
could also be evaluated by measuring the response when the intrusive 
strategy is turned off and engine-out NOX emissions are 
returned to normal levels. By correlating sensor response rates and the 
resulting

[[Page 3264]]

emissions impacts, the malfunction criteria could then be determined.

B. Feasibility of the Monitoring Requirements for Gasoline/Spark-
Ignition Engines

1. Fuel System Monitoring
    For gasoline vehicles since the 1996 model year and gasoline 
engines since the 2005 model year, the under 14,000 pound OBD 
requirements have required fuel system monitoring identical to that 
being proposed. Over 100 million cars and light trucks have been built 
and sold in the U.S. to these fuel system monitoring requirements 
including some heavy-duty vehicles that use the exact same gasoline 
engines that are used in some over 14,000 pound applications. This 
clearly demonstrates the technological feasibility of the proposed 
requirements.
2. Engine Misfire Monitoring
    For gasoline vehicles since the 1996 model year and gasoline 
engines since the 2005 model year, the under 14,000 pound OBD 
requirements have required misfire monitoring identical to that being 
proposed. One of the most reliable methods for detecting misfire is the 
use of a crankshaft position sensor--which measures the fluctuations in 
engine angular velocity to determine the presence of misfire--along 
with a camshaft position sensor--which can be used to identify the 
misfiring cylinder. This method has been shown to be technologically 
feasible and should work equally well on over 14,000 pound 
applications.
3. Exhaust Gas Recirculation (EGR) Monitoring
    For vehicles since the 1996 model year and engines since the 2005 
model year, the under 14,000 pound OBD requirements have required EGR 
system monitoring identical to that being proposed. The general 
approach has been to detect EGR flow rate malfunctions by looking at 
the change in fuel trim or manifold pressure under conditions when the 
EGR system is active. This demonstrates the technological feasibility 
of the proposed requirements.
4. Cold Start Emission Reduction Strategy Monitoring
    We expect this monitoring to be done mainly via computer software. 
For example, if spark retard is used during cold starts, the commanded 
amount of spark retard would have to be monitored if the amount of 
spark retard can be restricted by external factors such as idle quality 
or driveability. This can be done with software algorithms that compare 
the actual overall commanded final ignition timing with the threshold 
timing that would result in emissions that exceed the emissions 
thresholds. Cold start strategies that always command a predetermined 
amount of ignition retard independent of all other factors and do not 
allow idle quality or other factors to override the desired ignition 
retard would not require monitoring of the commanded timing. Other 
methods that could be used to ensure that the actual timing has been 
reached include verifying other factors such as corresponding increases 
in mass air flow and idle speed indicative of retarded spark 
combustion. Both mass air flow and idle speed are used currently by the 
engine control system and the OBD system and, therefore, only minor 
software modifications should be required to analyze these signals 
while the cold start strategy is invoked.
5. Secondary Air System Monitoring
    A/F sensors would most likely be required to monitor effectively 
the secondary air system when it is normally active. These sensors are 
currently installed on many new cars and their implementation is 
projected to increase in the future as more stringent emission 
standards are phased in. A/F sensors are useful in determining air-fuel 
ratio over a broader range than conventional oxygen sensors and are 
especially valuable in engines that require very precise fuel control. 
They would be useful for secondary air system monitoring because of 
their ability to determine air-fuel ratio with high accuracy. This 
would enable a correlation between secondary airflow rates and 
emissions.
6. Catalytic Converter Monitoring
    A common method used for estimating catalyst efficiency is to 
measure the catalyst's oxygen storage capacity. This monitoring method 
has been used by all light-duty gasoline vehicles since the 1996 model 
year and most gasoline engines since the 2005 model year as a result of 
our under 14,000 OBD requirements. Generally, as the catalyst's oxygen 
storage capacity decreases, the conversion efficiencies of HC and 
NOX also decrease. With this strategy, a catalyst 
malfunction would be detected when its oxygen storage capacity has 
deteriorated to a predetermined level. Manufacturers determine this by 
using the information from an upstream oxygen sensor and a downstream 
or mid-bed oxygen sensor (this second sensor is also used for trimming 
the front sensor to maintain more precise fuel control). By comparing 
the level of oxygen measured by the second sensor with that measured by 
the upstream sensor, manufacturers can determine the catalyst's oxygen 
storage capacity and estimate its conversion efficiency. With a 
properly functioning catalyst, the second oxygen sensor signal will be 
fairly steady since the fluctuating oxygen concentration (due to fuel 
system cycling around stoichiometry) at the inlet of the catalyst is 
damped by the storage and release of oxygen in the catalyst. When a 
catalyst is deteriorated it is no longer capable of storing and 
releasing oxygen. This causes the frequency and peak-to-peak voltage of 
the second oxygen sensor to simulate the signal from the upstream 
oxygen sensor at which time a malfunction would be indicated.
7. Evaporative System Monitoring
    Our OBD requirements have required monitoring for evaporative 
system leaks for many years. The EPA OBD requirement has been the 
equivalent of a 0.040 inch hole, while the ARB requirement has gone as 
low as a 0.020 inch hole. These requirements have been met on 
applications such as incomplete trucks and engine dynamometer certified 
configurations equipped with similar and, in many cases, identical 
configurations as are used in over 14,000 pound applications. 
Manufacturers have successfully met these requirements by using engine 
vacuum to create a vacuum in both the fuel tank and evaporative system 
and then monitoring the system's ability to maintain that vacuum. The 
ramp down in vacuum (or ramp up in pressure) can then be correlated to 
leak size. In general, these systems require the addition of an 
evaporative system pressure sensor and a canister vent valve capable of 
closing the vent line.
    Manufacturers of over 14,000 pound applications have expressed 
concerns with their ability to detect evaporative system leaks on these 
larger vehicles. One such concern relates to the relatively larger fuel 
tank sizes on the larger applications. These tanks can be on the order 
of 50 to 80 gallons, which makes the impact of a small hole, on a 
percentage basis, less severe and less easily detected. Another concern 
is the relatively large number of fuel tank and evaporative system 
configurations on the larger applications. Confounding both of these 
concerns is that the engine manufacturers quite often have no idea what 
tanks and configurations will ultimately be matched with their engine 
in the final vehicle product.
    While we agree that these concerns are valid, they can also be said 
of the

[[Page 3265]]

under 14,000 pound applications (except perhaps the tank size concern). 
The over 14,000 pound gasoline applications are expected to use near 
identical, if not equivalent, evaporative system components and we are 
not aware of any reason why the existing monitoring techniques would 
not continue to work on over 14,000 pound applications. Nonetheless, we 
do not want false failures in the field. By limiting the monitoring 
requirement to leaks of 0.150 inch or larger, we believe that 
manufacturers would be able to employ a single monitoring strategy to 
all possible tank sizes and configurations without much concern for 
false failures. Nonetheless, it may be necessary for manufacturers to 
impose tighter restrictions on their engine purchasers than is done 
currently with regards to tank specifications and evaporative system 
components.
8. Exhaust Gas Sensor Monitoring
    Our light-duty OBD requirements since the 1996 model year and our 
8,500 to 14,000 pound OBD requirements since the 2005 model year have 
required oxygen sensor monitoring similar to the requirements being 
proposed. Years of compliance with those requirements demonstrates the 
technological feasibility of the proposed requirements. Additionally, 
A/F sensor monitoring has been required and demonstrated on these 
vehicles for many years.

C. Feasibility of the Monitoring Requirements for Other Diesel and 
Gasoline Systems

1. Variable Valve Timing and/or Control (VVT) System Monitoring
    VVT systems are already in general use in many under 14,000 pound 
applications. Further, under the California OBD II requirements, 
vehicles equipped with VVT systems have been monitoring those systems 
for proper function since the 1996 model year. More recently, 
manufacturers have employed monitoring strategies to detect VVT system 
malfunctions that detect not only proper function but also exceedances 
of emissions thresholds. Such strategies include the use of the crank 
angle sensor and camshaft position sensor to confirm that the valve 
opening and closing occurs within an allowable tolerance of the 
commanded crank angle. By calculating the difference between the 
commanded valve opening crank angle and the achieved valve opening 
crank angle, a diagnostic algorithm can differentiate between a 
malfunctioning system with too large of an error and a properly 
functioning system with very little to no error. By calibrating the 
size of this error (or integrating it over time), manufacturers can 
design the system to indicate a malfunction prior to the required 
emissions thresholds. In the same manner, system response can be 
measured by monitoring the length of time necessary to achieve the 
commanded valve timing. To ensure adequate resolution between properly 
functioning systems and malfunctioning systems, most manufacturers 
perform this type of monitor only when a sufficiently large ``step 
change'' in commanded valve timing occurs.
2. Engine Cooling System Monitoring
    The existing OBD requirements have required identical ECT sensor 
and thermostat monitoring for several years. While the technical 
feasibility of the proposed requirements has been demonstrated on 
lighter applications which tend to be produced through a vertically 
integrated manufacturing process, the manufacturers of big diesel 
engines have expressed concerns that monitoring of the cooling system 
on over 14,000 pound applications would create unique and possibly 
insurmountable challenges. Generally, the cooling system is divided 
into two cooling circuits connected by the thermostat. The two circuits 
are the engine circuit and the radiator circuit. Since the big diesel 
engine industry tends to be horizontally integrated, the manufacturers 
contend that they do not know what types of devices will be added to 
the cooling system when the vehicle is manufactured or the vehicle is 
put into service. They are concerned that the unknown devices can add/
remove unknown quantities of heat to/from the system which would 
prevent them from predicting reliably the proper system behavior (e.g., 
warm up). Without the ability to predict system behavior reliably, they 
fear that they cannot know when the system is malfunctioning (e.g., not 
warming up as expected).
    The industry's concerns regarding unknown devices added on the 
radiator circuit of the system seem unwarranted. A properly functioning 
thermostat does not allow flow through the radiator during warm-up. 
Devices added to the radiator circuit could only affect coolant 
temperature when there is significant coolant flow through the radiator 
(i.e., after the engine is warmed-up and the thermostat is open, 
allowing coolant to flow through the radiator).
    We agree that unknown devices added on the engine circuit (e.g., 
passenger compartment heaters) can affect the warm-up rate of the 
system. Manufacturers of under 14,000 pound applications have 
demonstrated robust thermostat monitoring with high capacity passenger 
heaters in the cooling system. To do so, they have to know the maximum 
rate of heat loss due to the heater. Manufacturers of over 14,000 pound 
applications have control over this by providing limits on such devices 
in the build specifications that they provide to the vehicle 
manufacturers. In some cases, an engine manufacturer might need 
multiple build specifications with corresponding thermostat monitoring 
calibrations to accommodate the ranges of heater capacities that are 
needed when a given engine is used in a range of vehicle applications 
(e.g., a local delivery truck having a passenger compartment for two 
people and a small capacity heater versus a bus having a passenger 
compartment for 20 people and a large capacity heater). The vehicle 
manufacturer would then select the appropriate calibration for the 
engine when installing it in the vehicle. Nonetheless, engine 
manufacturers have requested limited enable conditions for the 
thermostat monitor (e.g., to disable the thermostat monitor below 50 
degrees F). This would help to minimize their resource needs to 
calibrate the thermostat monitor. While this may be directionally 
favorable to manufacturers, it would result in disabled thermostat 
monitoring during cold ambient conditions which occur in much of the 
country and, in some areas, during a large portion of the year. In such 
regions, a vehicle could experience a thermostat malfunction with no 
indication to the vehicle operator. Since many other OBD monitors will 
operate only after reaching a certain engine coolant temperature, a 
malfunctioning thermostat without any indication could effectively 
result in disablement of the OBD system.
3. Crankcase Ventilation System Monitoring
    Crankcase ventilation system monitoring requirements have been met 
for years by manufacturers of under 14,000 pound gasoline applications. 
Therefore, the technological feasibility has been demonstrated for 
gasoline applications.
    Effectively, diesel engine manufacturers would be required to meet 
design requirements for the entire system in lieu of actually 
monitoring any of the hoses for disconnection. Specifically, the 
proposed requirement would allow for an exemption for any portion of 
the system that is resistant to deterioration or accidental 
disconnection and not subject to disconnection during any of the

[[Page 3266]]

manufacturer's repair procedures for non-crankcase ventilation system 
repair work. These safeguards would be expected to eliminate the 
chances of disconnected or improperly connected hoses while still 
allowing manufacturers to meet the requirements without adding any 
additional hardware meant solely for the purpose of meeting the 
monitoring requirements.
4. Comprehensive Component Monitoring
    Both ARB and EPA OBD requirements have for year contained 
requirements to monitor computer input and output components. While 
these monitors are sometimes tricky and are not easy as many 
incorrectly assume, the many years of successful implementation and 
compliance with the existing requirements demonstrates their 
feasibility. The proposed requirements are equivalent to the under 
14,000 pound requirements.

IV. What Are the Service Information Availability Requirements?

A. What Is the Important Background Information for the Proposed 
Service Information Provisions?

    Section 202(m)(5) of the CAA directs EPA to promulgate regulations 
requiring OEMs to provide to:

any person engaged in the repairing or servicing of motor vehicles 
or motor vehicle engines, and the Administrator for use by any such 
persons, * * * any and all information needed to make use of the 
[vehicle's] emission control diagnostic system * * * and such other 
information including instructions for making emission-related 
diagnoses and repairs.

    Such requirements are subject to the requirements of section 208(c) 
regarding protection of trade secrets; however, no such information may 
be withheld under section 208(c) if that information is provided 
(directly or indirectly) by the manufacturer to its franchised dealers 
or other persons engaged in the repair, diagnosing or servicing of 
motor vehicles.
    On June 27, 2003 EPA published a final rulemaking (68 FR 38428) 
which set forth the Agency's service information regulations for light- 
and heavy-duty vehicles and engines below 14,000 pounds GVWR. These 
regulations, in part, required each-covered Original Equipment 
Manufacturer (OEM) to do the following: (1) OEMs must make full text 
emissions-related service information available via the World Wide Web. 
(2) OEMs must provide equipment and tool companies with information 
that allows them to develop pass-through reprogramming tools. (3) OEMs 
must make available enhanced diagnostic information to equipment and 
tool manufacturers and to make available OEM-specific diagnostic tools 
for sale. These requirements were finalized to ensure that aftermarket 
service and repair facilities have access to the same emission-related 
service information, in the same or similar manner, as that provided by 
OEMs to their franchised dealerships.
    As EPA moves forward proposing OBD requirements for the heavy-duty 
over 14,000 pounds sector, EPA is similarly moving forward with 
proposals to require the availability of service information to heavy-
duty aftermarket service providers as required by section 202(m) of the 
Clean Air Act.
    All of the following proposed provisions regarding the availability 
of service information for the heavy-duty industry are based on our 
extensive experience and regulatory history with the light-duty service 
industry. However, as discussed below, EPA understands that there may 
be significant differences between the light-duty service industry and 
the heavy-duty service industry. EPA welcomes comment on all of the 
proposed provisions and their need and/or applicability to the heavy-
duty service industry.
B. How Do the Below 14,000 Pound and Above 14,000 Pounds Aftermarket 
Service Industry Compare?
    As we consider proposing the availability of service information 
for the heavy-duty sector above 14,000 pounds, EPA recognizes that 
differences do exist between the industries that service vehicles above 
and below 14,000 pounds. On the below 14,000 pound side, estimates 
indicate that independent technicians perform up to 80% of all vehicle 
service and repairs once a vehicle exceeds the manufacturer warranty 
period.\69\ On the above 14,000 pound side, the 1997 U.S. Census Bureau 
Vehicle Inventory and Use Survey, estimated that 25 percent of the 
general maintenance and over 30 percent of the major overhaul on heavy-
duty vehicles was performed by the independent sector. According to the 
Census Bureau, these values represent a 16.7 percent increase in 
general maintenance and a 6.2 percent increase in major overhaul from 
1992. Trucks and Parts Service Magazine provides the following 
information on the breakdown of the independent repair industry for 
vehicles above 14,000 pounds (not including any fuel injection shops):
---------------------------------------------------------------------------

    \69\ Motor and Equipment Manufacturers Association, Automotive 
Industry Status Report, 1999.

U.S. independent machine shops for above 14,000 pounds--5,820
U.S. independent engine service shops for above 14,000 pounds--12,170
U.S. independent transmission repair shops for above 14,000 pounds--
11,420
Technicians, independent repair shops for above 14,000 pounds--133,700
Technicians, truck parts distributors for vehicles above 14,000 
pounds--41,600

    Thus, the increase in business and the large number of independent 
aftermarket shops make it necessary that repair information is readily 
available for the aftermarket trucking industry.
    On the light-duty side, vehicle manufacturers are entirely 
integrated in that they are responsible for the design and production 
of the entire vehicle from the chassis to the body. In comparison, the 
heavy-duty industry is mostly non-integrated. In other words, different 
manufacturers separately produce the engine, the chassis, and the 
transmission of a vehicle. This non-integration speaks to the fact that 
a completed vehicle is typically produced in response to the customized 
needs of owners/operators. In addition, the lack of integration 
indicates that a given engine will ultimately be part of many different 
engine, transmission, and chassis configurations. In addition, heavy-
duty manufacturers have stated that diagnostic tool designs differ 
significantly from tools produced for light-duty vehicles as a result 
of this non-integration.
    EPA requests comment and also additional data on the current state 
of the heavy-duty aftermarket industry.

C. What Provisions Are Being Proposed for Service Information 
Availability?

1. What Information Is Proposed To Be Made Available by OEMs?
    Today's action proposes a provision that requires OEMs to make 
available to any person engaged in the repairing or servicing of heavy-
duty motor vehicles or motor vehicle engines above 14,000 pounds all 
information necessary to make use of the OBD systems and any 
information for making emission-related repairs, including any 
emissions-related information that is provided by the OEM to franchised 
dealers beginning with MY2010. We are proposing that this information 
includes, but is not limited to, the following:
    (1) Manuals, technical service bulletins (TSBs), diagrams, and 
charts (the provisions for training materials,

[[Page 3267]]

including videos and other media are discussed in Sections II.C.3 and 
II.C.4 below.
    (2) A general description of the operation of each monitor, 
including a description of the parameter that is being monitored.
    (3) A listing of all typical OBD diagnostic trouble codes 
associated with each monitor.
    (4) A description of the typical enabling conditions for each 
monitor to execute during vehicle operation, including, but not limited 
to, minimum and maximum intake air and engine coolant temperature, 
vehicle speed range, and time after engine startup. A listing and 
description of all existing monitor-specific drive cycle information 
for those vehicles that perform misfire, fuel system, and comprehensive 
component monitoring.
    (5) A listing of each monitor sequence, execution frequency and 
typical duration.
    (6) A listing of typical malfunction thresholds for each monitor.
    (7) For OBD par