[Code of Federal Regulations]
[Title 40, Volume 18]
[Revised as of July 1, 2003]
From the U.S. Government Printing Office via GPO Access
[CITE: 40CFR92.129]

[Page 460-462]
 
                   TITLE 40--PROTECTION OF ENVIRONMENT
 
         CHAPTER I--ENVIRONMENTAL PROTECTION AGENCY (CONTINUED)
 
PART 92--CONTROL OF AIR POLLUTION FROM LOCOMOTIVES AND LOCOMOTIVE ENGINES--
Table of Contents
 
                       Subpart B--Test Procedures
 
Sec. 92.129  Exhaust sample analysis.

    (a) The analyzer response may be read by automatic data collection 
(ADC) equipment such as computers, data loggers, etc. If ADC equipment 
is used the following is required:

[[Page 461]]

    (1) The response complies with Sec. 92.130.
    (2) The response required in paragraph (a)(1) of this section may be 
stored on long-term computer storage devices such as computer tapes, 
storage discs, or they may be printed in a listing for storage. In 
either case a chart recorder is not required and records from a chart 
recorder, if they exist, need not be stored.
    (3) If the data from ADC equipment is used as permanent records, the 
ADC equipment and the analyzer values as interpreted by the ADC 
equipment are subject to the calibration specifications in Secs. 92.118 
through 92.122, as if the ADC equipment were part of the analyzer.
    (b) Data records from any one or a combination of analyzers may be 
stored as chart recorder records.
    (c) Software zero and span.
    (1) The use of ``software'' zero and span is permitted. The process 
of software zero and span refers to the technique of initially adjusting 
the analyzer zero and span responses to the calibration curve values, 
but for subsequent zero and span checks the analyzer response is simply 
recorded without adjusting the analyzer gain. The observed analyzer 
response recorded from the subsequent check is mathematically corrected 
back to the calibration curve values for zero and span. The same 
mathematical correction is then applied to the analyzer's response to a 
sample of exhaust gas in order to compute the true sample concentration.
    (2) The maximum amount of software zero and span mathematical 
correction is 10 percent of full scale chart deflection.
    (3) Software zero and span may be used to switch between ranges 
without adjusting the gain of the analyzer.
    (4) The software zero and span technique may not be used to mask 
analyzer drift. The observed chart deflection before and after a given 
time period or event shall be used for computing the drift. Software 
zero and span may be used after the drift has been computed to 
mathematically adjust any span drift so that the ``after'' span check 
may be transformed into the ``before'' span check for the next mode.
    (d) For sample analysis perform the following sequence:
    (1) Warm-up and stabilize the analyzers; clean and/or replace filter 
elements, conditioning columns (if used), etc., as necessary.
    (2) Leak check portions of the sampling system that operate at 
negative gauge pressures when sampling, and allow heated sample lines, 
filters, pumps, etc., to stabilize at operating temperature.
    (3) Optional: Perform a hang-up check for the HFID sampling system:
    (i) Zero the analyzer using zero air introduced at the analyzer 
port.
    (ii) Flow zero air through the overflow sampling system, where an 
overflow system is used. Check the analyzer response.
    (iii) If the overflow zero response exceeds the analyzer zero 
response by 2 percent or more of the HFID full-scale deflection, hang-up 
is indicated and corrective action must be taken.
    (iv) The complete system hang-up check specified in paragraph (f) of 
this section is recommended as a periodic check.
    (4) Obtain a stable zero reading.
    (5) Zero and span each range to be used on each analyzer used prior 
to the beginning of the test sequence. The span gases shall have a 
concentration between 75 and 100 percent of full scale chart deflection. 
The flow rates and system pressures shall be approximately the same as 
those encountered during sampling. The HFID analyzer shall be zeroed and 
spanned through the overflow sampling system, where an overflow system 
is used.
    (6) Re-check zero response. If this zero response differs from the 
zero response recorded in paragraph (d)(5) of this section by more than 
1 percent of full scale, then paragraphs (d) (4), (5), and (6) of this 
section should be repeated.
    (7) If a chart recorder is used, identify and record the most recent 
zero and span response as the pre-analysis values.
    (8) If ADC equipment is used, electronically record the most recent 
zero and span response as the pre-analysis values.

[[Page 462]]

    (9) Measure (or collect a sample of) the emissions continuously 
during each mode of the test cycle. Indicate the start of the test, the 
range(s) used, and the end of the test on the recording medium (chart 
paper or ADC equipment). Maintain approximately the same flow rates and 
system pressures used in paragraph (d)(5) of this section.
    (10)(i) Collect background HC, CO, CO2, and 
NOX in a sample bag (optional).
    (ii) Measure the concentration of CO2 in the dilution air 
and the diluted exhaust for particulate measurements.
    (11) Perform a post-analysis zero and span check for each range used 
at the conditions specified in paragraph (d)(5) of this section. Record 
these responses as the post-analysis values.
    (12) Neither the zero drift nor the span drift between the pre-
analysis and post-analysis checks on any range used may exceed 3 percent 
for HC, or 2 percent for NOX, CO, and CO2, of full 
scale chart deflection, or the test is void. (If the HC drift is greater 
than 3 percent of full-scale chart deflection, hydrocarbon hang-up is 
likely.)
    (13) Determine HC background levels (if necessary) by introducing 
the background sample into the overflow sample system.
    (14) Determine background levels of NOX, CO, or 
CO2 (if necessary).
    (e) HC hang-up. If HC hang-up is indicated, the following sequence 
may be performed:
    (1) Fill a clean sample bag with background air.
    (2) Zero and span the HFID at the analyzer ports.
    (3) Analyze the background air sample bag through the analyzer 
ports.
    (4) Analyze the background air through the entire sample probe 
system.
    (5) If the difference between the readings obtained is 2 percent or 
more of the HFID full scale deflection:
    (i) Clean the sample probe and the sample line;
    (ii) Reassemble the sample system;
    (iii) Heat to specified temperature; and
    (iv) Repeat the procedure in this paragraph (e).