Skip to main navigation

2004 I/M Emissions Test Failure Rate Report

Table of Contents

• Executive Summary
• I. Introduction
• II. OBD Inspections
• II.1. NYTEST-Based OBD II Inspection Program for NYMA
• II.2. NYVIP-Based OBD II Inspection Program for UPSTATE
• II.3. Taxi and Limousine Commission (TLC) OBD II Inspection Program
• III. Transient Tests
• IV. Idle Tests
• V. Conclusions
• Appendix A
• Appendix B

Mobile Emissions and Air Quality Section
Bureau of Air Quality and Research
Division of Air Resources
New York State Department of Environmental Conservation
625 Broadway

Albany, New York 12233

(518) 402-8502

<>

In 1998, in compliance with the 1990 Federal Clean Air Act Amendments (CAAA), the New York State (NYS)
Department of Environmental Conservation (DEC) and the Department of Motor Vehicles (DMV) jointly
implemented a high enhanced Inspection and Maintenance (I/M) program for the nine (9) counties of the
New York Metropolitan Area (NYMA) and a low enhanced I/M program for the remaining fifty-three (53) counties
of the Upstate Region (UPSTATE), as outlined in New York State's I/M State Implementation Plan (SIP).

In 2004, while continuing to operate the existing I/M programs, NYS began the process of implementing a
statewide On-Board Diagnostics (OBD II) inspection program. In addition, on December 8, 2003, the first
two inspection lanes at the Taxi and Limousine Commission (TLC) Woodside inspection facility started both
safety and OBD II inspections.

Failure rates, based on data collected during year 2004 from all three types of emissions tests (OBD II,
transient, and idle) and from all three area programs (NYMA, UPSTATE, and TLC) are summarized, compared,
and analyzed. The objectives are to verify the accuracy of the program software used in the field, to report
the failure rates based on failure criteria or pollutant species (in addition to the overall failure rates),
to evaluate the program efficiency based on failure rates, and to provide recommendations for program improvement.

As 2004 is the first year that the OBD II inspection program was implemented in NYS, the failure rates obtained
and reported in this report appeared to be high due to a combination of software issues and some inspections not
being conducted "properly." We expect the software issues to be resolved in year 2005. It is important
for the testing facilities (and vehicle owners) to be fully aware of the fact that vehicles must complete certain
driving cycles for certain monitors to be ready for the OBD inspections. It is not a "true"
failure for a vehicle to fail the OBD on readiness alone. In 2004, 60.18% of the UPSTATE failing OBD inspection
vehicles failed on readiness alone. Another problem common to all OBD II inspection programs is the elevated count
of MIL Off KOEO (Malfunction Indicator Light is off when the ignition is in the key on/engine off position). It
is easy for technicians to miss the momentary illumination of the MIL thus counting it as a failure. Additional
technician training emphasis on this problem and program experience should reduce false failures from this category.
With the adoption of these remedies, it is expected that in future years the reported OBD failure rates will be
lower than those observed in 2004.

The tailpipe exhaust tests, transient and idle, are being greatly reduced in number by the implementation of the
OBD II inspection program in the NYMA as the new I/M program requirement. The light duty vehicles and trucks still
receiving either tailpipe test will be a fleet with older model years (1981 to 1995), hence the reported overall failure
rates are expected to rise in the future. However, a comparison of year 2004 with year 2003 reported failure rates for
the same model year fleet has shown a decrease, most likely due to the I/M program effects of repair, both required and
anticipatory, and removal of vehicles from the fleet through attrition.

I. Introduction

In 1998, in compliance with the 1990 Federal Clean Air Act Amendments (CAAA), the New York State (NYS) Department of
Environmental Conservation (DEC) and the Department of Motor Vehicles (DMV) jointly implemented a high enhanced Inspection
and Maintenance (I/M) program for the nine (9) counties of the New York Metropolitan Area (NYMA) and a low enhanced I/M
program for the remaining fifty-three (53) counties of the Upstate Region (UPSTATE), as outlined in New York State's I/M
State Implementation Plan (SIP). Figure 1 displays the geographic distributions of counties
in the NYS.

All gasoline-powered vehicles between the ages of 2 and 25 years old registered to operate in the State of New York
are required to have the following inspections/emissions tests, on an annual basis as well as at the time of a change
in vehicle ownership:

• a safety inspection;
• emissions tests for the high enhanced I/M program, which include a tailpipe emissions test, a visual anti-tampering
  check of the various emissions control device systems, and a gas cap integrity test;
• emissions tests for the low enhanced I/M program, which include a visual anti-tampering check of the various emissions
  control device systems, and a gas cap presence check.

The tailpipe emissions test, commenced in 1998 for the high enhanced I/M program, consists of two test components:
a loaded-mode transient emissions test (NYTEST), and a single-speed idle emissions test (IDLE). In 2004, while continuing
to operate the existing I/M programs, NYS began the process of implementing a statewide On-Board Diagnostics (OBD II)
inspection program. The OBD inspection was instituted in the NYMA region to replace, in large part, the existing transient
and idle tests. For the UPSTATE counties, the OBD inspection was added to the anti-tampering check and gas cap check
which substantially increased the requirement for vehicle maintenance. In addition, on December 8, 2003, the first two
inspection lanes at the Taxi and Limousine Commission (TLC) Woodside inspection facility started both safety and OBD inspections.

Failure rates, based on data collected during the year 2004 from all three types of emissions tests (OBD II, transient,
and idle) and from all three area programs (NYMA, UPSTATE, and TLC) are summarized, compared, and analyzed. The objectives
are to verify the accuracy of the program software used in the field, to report the failure rates based on failure criteria
or pollutant species (in addition to the overall failure rates), to evaluate the program efficiency based on failure rates,
and to provide recommendations for program improvement.

For the OBD inspections, failure rates are determined by a set of predefining OBD failure criteria. That is, the vehicle
fails the OBD inspection if one or more of the OBD failure criteria are met. For the tailpipe emissions test (either NYTEST
or IDLE), failure rates are determined by a set of predefined cutpoints. That is, the vehicle fails the tailpipe emissions
test if the reported emissions value(s), for one or more pollutant species, is higher than a set of predefined cutpoints.

The reported failure rate is based on counts of reported failures, with those failing the test denoted as "F,"
obtained from our I/M program database. These reported OBD inspection results or emissions tests (pass or fail) were determined
by the software used by the testing stations at the time of testing and were transmitted to the I/M program data manager after
the tests were concluded. The evaluated failure rates are determined by using DEC's computer programs to re-evaluate the
reported pass/fail results based on each component of the OBD inspection results or with emissions values for each pollutant
species for the tailpipe tests.

II. OBD Inspections

OBD is a monitoring and fault detection/notification process of the powertrain control module (PCM) related to the vehicle's
emission control system and powertrain operation. OBD was originally adopted by the US Environmental Protection Agency (USEPA
or EPA) in 1989 for 1994+ model year vehicles, with OBD phase-in beginning in California in 1994.

OBD inspections are applicable to all non-diesel, non-electric, model year 1996+ light-duty vehicles (LDVs) and light-duty
trucks (LDTs). Based on the EPA's guidance document, NYS applies the following five (5) failure criteria to vehicles
undergoing an OBD II inspection:

• the vehicle is unable to communicate with the OBD inspection equipment;
• the vehicle's Malfunction Indicator Light (MIL) does not illuminate when the ignition is in the key on/engine off (KO/EO)
  position;
• the vehicle's MIL remains illuminated when the ignition is in the key on/engine running (KO/ER) position;
• the vehicle PCM has commanded the MIL and diagnostic trouble codes (DTCs) are stored in memory;
• the vehicle fails the readiness evaluation (with 3 or more monitors not ready for 1996-2000 model years, and 2 or more
  monitors not ready for 2001+ model years).

In 2004, while continuing to operate the existing I/M programs, NYS began the process of implementing a statewide OBD II
inspection program. The NYS OBD II inspection program consisted of three components:

1. A NYTEST-based optional program within NYMA;
2. A statewide New York Vehicle Inspection Program (NYVIP) which was implemented initially in UPSTATE; and,
3. A centralized testing program administered by the TLC in New York City for the medallion taxicab fleet.

Note that results from both NYMA and UPSTATE OBD II inspection program components described above were limited in that
they were based on a partial year's worth of data collection. Since the TLC OBD II inspection program commenced on December 8, 2003, the TLC program failure rate analysis included both
the December 2003 and year 2004 data. In addition, NYS regulations allow for the two newest model years to be exempt from
OBD inspections. However, the TLC has included all taxicabs, including the two newest model years, in its OBD II inspection program.

The overall reported OBDII inspection failure rates are shown in Table 1. Each of the three area components (NYMA, UPSTATE,
and TLC) is discussed in greater detail in the following three sections.

II.1. NYTEST-Based OBD II Inspection Program for NYMA

In 2004, the NYTEST-based OBD II inspection program for NYMA became operational. The test was based on upgraded software
from two of the NYTEST equipment providers (ESP and SPX). The optional OBD upgrade software was approved for ESP on
July 2, 2004 and for SPX on October 18, 2004, respectively. However, many vehicles were OBD inspected using beta versions of
the OBD upgrade software from either ESP or SPX. By December 31, 2004, a total of 726 NYTEST testing facilities within NYMA
had performed OBD inspections for a total of 140,538 LDVs and LDTs, which represented 6.17% of the total NYMA vehicle population.

Detailed NYMA OBD inspection results are presented in Appendix A. Note that vehicles may fail
multiple OBD failure criteria (e.g., with MIL commanded on and monitors not ready). Therefore, the sum of vehicles failed by
each criterion may not be the same as the number of vehicles failing OBD inspections.

Appendix A Table 1.a.i shows the OBD inspection results for LDVs by vehicle model year. The overall
LDV fleet failure rate was 9.34%. As expected, 1996 vehicles, the oldest model year tested, had the highest failure rate of
14.39%, and 2002 vehicles, the newest model year tested, had the lowest failure rate of 5.89%. Note that the failure rate for
model year 2001 (8.98%) is slightly higher than that from model year 2000 (8.21%), due to more vehicles failing readiness criteria.

The readiness statistics for LDVs are summarized in Appendix A Table 1.a.ii. As can be seen, the top
five (5) "not ready" monitors were: catalyst (7,922 or 8.92%), evaporative system (6,828 or 7.69%), O₂
sensor (5,358 or 6.03%), EGR (3,240 or 3.65%), and heated O₂ sensor (2,965 or 3.34%). For pre-1999
vehicles, more vehicles had catalyst monitors not ready. For 1999+ model years, more vehicles had evaporative system monitors
not ready.

Appendix A Table 1.b.i shows the OBD inspection results for LDTs by vehicle model year. The overall
LDT fleet failure rate was 10.40%. As with the LDVs, 1996 vehicles, the oldest model year tested, had the highest failure rate
of 18.56%, and 2002 vehicles, the newest model year tested, had the lowest failure rate of 7.07%. It is also noted that a large
number of model year 2001 LDTs had their "monitors not ready" resulting in an overall failure rate of 10.79%, which is
higher than the failure rates obtained from the two previous model years (1999 and 2000).

The readiness statistics for LDTs are summarized in Appendix A Table 1.b.ii. As can be seen, the top
five (5) "not ready" monitors reported for LDTs, the same as those for the LDVs but in a slightly different order, were:
catalyst, (5,481 or 10.60%), evaporative system (4,886 or 9.45%), O₂ sensor (3,156 or 6.10%), heated
O₂ sensor (2,505 or 4.84%), and EGR (1,940 or 3.75%). Similar to LDVs, for pre-1999 vehicles, more
vehicles had catalyst monitors not ready, while for 1999+ model years, more vehicles had evaporative system monitors not ready.

Appendix A Table 1.c summarizes the statistics based on a comparison of OBD inspections and gas cap check
results by model year for LDVs and LDTs, as required by the EPA annual I/M program report, defined in 40 CFR Part 51.366
(7-1-98 Edition). It is noted that, for unknown reasons, some vehicles bypassed the required gas cap integrity (or pressure)
check test. Thus, the total number counts from Appendix A Table 1.c are less than those from
Appendix A Table 1.a.i and Appendix A Table 1.b.i. Between the OBD inspection and
gas cap check results, it is noted that only 0.47% (or 368) of the LDVs and 0.52% (or 232) of the LDTs displayed "inconsistent"
results (i.e., passing one and failing the other).

The "reported" failure rates were compared with the "evaluated" results. As shown in Table 2,
good agreements were noted between the reported and the evaluated: 99.027% for the LDVs and 99.814% for the LDTs.

Note that 112 OBD inspected vehicles (97 LDVs and 15 LDTs) had "no communications" (i.e., no communications between the
OBD scanner and the OBD system of the vehicle) and were reported as "stopped" tests (denoted as "S" in database)
by the use of earlier ESP/SPX software versions. In the final DEC-certified software versions, however, all "no communications"
were reported as fails (denoted as "F" in database).

The discrepancies between the reported "pass" and the evaluated "fail" were explored. We examined the sample
that consisted of 405 LDVs (the 0.456% from Table 2) which, based on the OBD II failure criteria should have failed the OBD inspection
but were reported as passing the test. It is determined that these "discrepancies" were linked to the use of earlier ESP/SPX
software versions.

Appendix A Table 1.d summarizes the top five (5) most frequently stored DTCs obtained from this data set by
vehicle model year and by vehicle type. The seven (7) most frequently stored DTCs for LDVs and LDTs are presented in
Table 3.

Note that three (3) DTCs appeared on the top seven lists for both LDVs and LDTs: P0420 (catalyst efficiency below a threshold),
P0171 (air/fuel system too lean), and P0401 (insufficient EGR flow). Additional DTCs for LDVs were: P0430 (catalyst efficiency below
a threshold), P0440 (evaporative system), P0141 (O₂ sensor heater circuit), and P0325 (knock sensor).
Additional DTCs for LDTs were: P0120 (throttle/pedal position sensor), P0133 (O2 sensor circuit slow response), P0174 (air/fuel system
too lean), and P0455 (large leak in evaporative system)

The month-to-month variations in the evaluated OBD failure rates were also examined and results are summarized in Table 4. Data from the first three months (April to June) can be ignored due to the issue of the use of "earlier" software versions as well as their relatively smaller sample sizes. For the remaining six months of 2004, the failure rates varied between 10.72% (December) and 8.68% (September) for LDVs, and between 11.40% (December) and 9.27% (September) for LDTs. Based on the fact that this was the first year of the OBD II inspection program implementation and the differences being relatively small, these month-to-month variations should be considered as insignificant.

An examination of the failure rate differences due to vendors (equipment and software), as displayed in Table 5, showed that testing facilities using SPX software had higher failure rates than ESP testing facilities for both LDVs and LDTs.

An examination of the failure rate differences due to the county location of the testing facilities (Table 6) indicated that high (> 10%) OBD failure rates for both LDVs and LDTs were observed in four of the NYMA counties: New York, Queens, Kings, and Bronx. Richmond and Nassau counties had the lowest OBD failure rates for LDVs, while Richmond and Westchester had the lowest OBD failure rates for LDTs. (Note that the percentages from LDVs do not add up to 100% due to the fact that a few OBD inspection stations are located in Orange, Putnam counties and the Port Authority of New York and New Jersey testing facility is located in New Jersey.)

II.2. NYVIP-Based OBD II Inspection Program for UPSTATE

On March 24, 2003, NYS DMV released a Request for Proposal for a Program Manager for an OBD II-based I/M program named New York Vehicle Inspection Program (NYVIP), designed to be implemented in several phases, starting with a program roll-out in the fifty-three (53) counties in UPSTATE, and eventually expanding to NYMA. The NYVIP contract was awarded to SGS TestCom, Inc. on January 9, 2004.

Limited OBD inspections, using alpha and beta versions of NYVIP software, commenced in January 2004 and continued to August with approximately 30 to 40 UPSTATE stations participating. The final NYVIP software was installed on September 17, 2004 and the UPSTATE program was deemed fully operational by December 1, 2004. As of December 31, 2004, 5,726 UPSTATE inspection stations had submitted official NYVIP inspection records from a total of 601,341 registered vehicles. Among them, 155,293 LDVs and 127,895 LDTs had OBD inspections.

Detailed summary UPSTATE data are presented in Appendix A.

Appendix A Table 2.a.i shows the OBD inspection results for LDVs by vehicle model year. The overall LDV fleet failure rate was 8.35%. As expected, 1996 vehicles, the oldest model year tested, had the highest failure rate of 15.47%, and 2002 vehicles, the newest model year tested, had the lowest failure rate of 4.20%.

The LDV readiness statistics are summarized in Appendix A Table 2.a.ii. As can be seen, the top five (5) "not ready" monitors for LDVs were: catalyst (19,680 or 12.67%), evaporative system (18,776 or 12.09%), O₂ sensor (13,362 or 8.60%), EGR (9,188 or 5.92%), and heated O₂ sensor (8,374 or 5.39%). For pre-1998 model years, more vehicles had catalyst monitors not ready, and for 1998+ model years, more vehicles had evaporative system monitors not ready.

Appendix A Table 2.b.i shows the OBD inspection results for LDTs by vehicle model year. The overall LDT fleet failure rate was 10.00%. As with the LDVs, 1996 vehicles, the oldest model year tested, had the highest failure rate of 18.28%, and 2002 vehicles, the newest model year tested, had the lowest failure rate of 6.47%. It is noted that a large number of model year 2001 LDTs had "monitors not ready," resulting in an overall failure rate of 10.09%, higher than the previous two model years (1999, 2000).

The LDT readiness statistics are summarized in Appendix A Table 2.b.ii. As can be seen, the top five (5) "not ready" monitors reported for LDTs, the same as those from LDVs but in a slightly different order, were: evaporative system (17,447 or 13.64%), catalyst (17,350 or 13.57%), O₂ sensor (10,105 or 7.90%), heated O₂ sensor (9,011 or 7.05%), and EGR (6,869 or 5.37%). Except for model years 1996 and 1997, evaporative system was the top "not ready" monitor.

Appendix A Table 2.c summarizes the statistics on comparisons of UPSTATE OBD inspections vs. gas cap check results by vehicle type and vehicle model year. Note that for the UPSTATE I/M program, the gas cap check is a visual inspection for the presence of a gas cap only (not a gas cap pressure check test as required for the NYMA program). Thus the gas cap failure rates were expected to be lower than those observed in NYMA. It is noted that 0.70% (or 995) of the LDVs and 0.64% (or 747) of the LDTs displayed "inconsistent" results (i.e., passing one and failing the other).

The "reported" failure rates were compared with the "evaluated" results. As shown in Table 7, good agreements were noted between the reported and the evaluated: 99.39% for the LDVs and 99.83% for the LDTs.

Since there is only one software version used in the NYVIP program, it is expected that the discrepancies should be smaller than those found in NYMA (where multiple software versions from two different vendors were used). Possible reasons were explored for the differences between the reported and evaluated results based on a sub-sample consisting of the 793 LDVs (the 0.51% from Table 7) which, based on the OBD failure criteria, should have failed the OBD inspection. It is concluded that the discrepancies were the result of the inconsistency between the readiness criteria defined by NYVIP software and our algorithm for certain groups of vehicle makes/models, such as Subaru, Mitsubishi, Volvo, Nissan, and Chevrolet. Hence, the accuracy of the reported OBD failure rates is deemed acceptable.

Appendix A Table 2.d summarizes the top five (5) most frequently stored DTCs obtained from this data set by vehicle type and vehicle model year. The seven (7) most frequently stored DTCs for LDVs and LDTs are presented in Table 8. Note that four (4) DTCs appeared on the top seven lists for both LDVs and LDTs: P0420 (catalyst efficiency below a threshold), P0401 (insufficient EGR flow), P0141 (O₂ sensor heater circuit), and P0442 (small leak in evaporative system). Additional DTCs for LDVs were: P0440 (evaporative system), P0300 (random/multiple misfire detected), and P0133 (O₂ sensor circuit slow response). Additional DTCs for LDTs were: P0171/P0174 (air/fuel system too lean), and P0455 (large leak in evaporative system).

The month-to-month variations in the reported OBD failure rates were also examined and results are summarized in Table 9. Data from the first month (September) can be ignored due to their relatively smaller sample sizes. For the remaining three months of 2004, the failure rates varied between 9.17% (December) and 7.18% (October) for LDVs, and between 10.83% (December) and 8.58% (October) for LDTs. Based on the fact that this was the first year of this OBD II inspection program implementation, these month-to-month variations are considered as insignificant.

Table 10 displays the failure rate differences shown in the county location of the testing facilities. It should be noted that these failure rates by county were based only on three and a half months of data. Statistics are provided here for reference only, as many counties had OBD inspection data sample of less than 1,000 vehicles.

II.3. TLC OBD II Inspection Program

According to the New York City Taxi and Livery Fact Book, taxicabs are defined in New York City law as vehicles seating fewer than nine passengers (in addition to the driver) and permitted to pick up street hails. They are the only vehicles allowed by law to pick up street hails or pick up at taxi or airport stands. Uniquely to New York, regulations prohibit taxicabs from prearranging service through two-way radios. The number of taxicabs is set by local law and rarely changed. Taxicab vehicle licenses ("medallions" in common parlance) are transferable and currently fetch well over $200,000 in the market.

The entity TLC was created in 1971 by City Council legislation to regulate and improve taxi and livery service in New York City and establish overall transportation policy governing these services. As a city agency regulating a private industry, TLC licenses 40,000 taxicab drivers and the owners of the 12,187 taxicabs.

Under terms of a Consent Order filed on September 6, 1977, TLC is required to provide emissions tests for all taxi cabs on a three (3) times per year basis. Since 1998, all taxicabs have been included in the high enhanced I/M program.

In March 2003, TLC selected a contractor, SysTec International, and commenced working toward an OBD II inspection program for all their medallion taxi fleets. On December 8, 2003, the first two inspection lanes at the TLC's Woodside inspection facility started both safety and OBD inspections. By August 2004, all six lanes were operational. From the OBD inspection data provided by SysTec, 13,913 inspected taxicabs, including those less than two years old, were included in the OBD II inspection program at four (4) month intervals with no provision for waivers. The vast majority of the TLC medallion taxicab fleet were LDVs (Ford Crown Victoria), with a small fraction of LDTs (sport utility vans or passenger vans, such as Honda Odyssey and Isuzu Oasis).

Detailed TLC OBD inspection results are also presented in Appendix A.

Appendix A Table 3.a.i shows the OBD inspection results for LDVs by vehicle model year. The overall reported LDV fleet failure rate was 19.61%. It is noted that the majority of the taxicabs were from 2000 and later model years. Except for the two newest model years (2003 and 2004), the reported failure rates were a lot higher than those observed from privately-owned NYMA and UPSTATE vehicles.

It is noted that the number of vehicles reported failing for readiness across all model years was very low. In addition, vehicles that were unable to communicate with the OBD scanner were reported as "N" ("not tested"), when they should have been coded as an "F," according to EPA's guidance described previously in this report. Therefore, the reported LDV sample count (29,810 as reported in Appendix A Table 3.a.i) is lower. It should be 30,380 (i.e., adding the 570 LDVs coded as "no communication"). The failure rates were also "under-reported." Therefore, the reported TLC OBD inspection results were found to be inconsistent with the NYS OBD II inspection regulations.

Appendix A Table 3.a.ii shows the readiness statistics for LDVs. As can be seen, the number of vehicles with "monitors not ready" was in fact very high. Those monitors included: evaporative system (24,739 or 81.43%), catalyst (12,571 or 41.38%), both O₂ sensor and heated O₂ sensor (7,270 or 23.93%), and EGR (4360 or 14.35%).

The evaluated OBD II inspection results for LDVs are presented in Appendix A Table 3.a.iii. As can be seen, the overall evaluated failure rate was 46.72%, compared with the reported 19.61%.

Appendix A Table 3.b.i shows the OBD II inspection results for LDTs by vehicle model year. The reported overall LDT fleet failure rate was 12.75%. It is noted that the sample sizes for three of the model year LDTs were less than 100 (1996, 2000, and 2001). Note also that there were no vehicles reported failing for the readiness criterion. There were 28 "no communication" LDTs not included in the reported sample count of 1,961.

Appendix A Table 3.b.ii shows the readiness statistics for LDTs. As can be seen, the top five "not ready" monitors reported for LDTs were: catalyst, (593 or 29.81%), O2 sensor (437 or 21.97%), evaporative system (412 or 20.71%), EGR (357 or 17.95%), and heated O2 sensor (24 or 1.21%).

The evaluated OBD II inspection results are presented in Appendix A Table 3.b.iii. As can be seen, the overall evaluated failure rate was 24.74%, compared with the reported failure rate of 12.75%.

Based on the large differences between the reported and the evaluated shown in Appendix A Table 3.a.iii (LDVs) and Appendix A Table 3.b.iii (LDTs), it is evident that the failure rates obtained from an OBD inspection program are "sensitive" to two factors: consistency and validation of the software, and, more importantly, the manner in which the inspection was conducted. It was noted that most taxicab owners were probably not aware of the "special driving cycle" the vehicles were required to complete prior to their OBD inspections. Because their function is to provide transportation services and their areas of operation are limited to the stop-and-go driving pattern of the streets of New York City, many taxicabs failed the readiness criteria by having the vehicles' catalyst, evaporative system, and/or O₂ sensor monitors "not ready" for the OBD inspections.

Table 11 shows the discrepancies between the reported and the evaluated. As discussed previously, the large discrepancies are attributed to the exclusion of the readiness evaluation criteria in TLC's software, with most of the differences falling in the category where those reported as "passed" should have been reported as "failed".

Appendix A Table 3.d summarizes the top three (3) most frequently stored DTCs obtained from this data set by vehicle type and vehicle model years. The seven (7) most frequently stored DTCs for both LDVs and LDTs are presented in Table 12. Two DTC's appeared in both LDVs and LDTs lists: P0420, and P0401.

The month-to-month variations based on the evaluated OBD failure rates were examined and the results are summarized in Table 13. It is noted that the failure rates decreased steadily, ranging from 61.22% in December 2003 and remained around 42% after July of 2004 for the LDVs, for example. This seems to demonstrate the importance of the practice and experience in conducting OBD II inspections.

III. Transient Tests

The NYMA transient test data included in this report covers the period from January 1, 2004 through December 31, 2004. A total of 3,538,632 vehicles were emissions-tested during this time period, 2,346,188 (66.30%) were LDVs, 1,184,466 (33.47%) were LDTs, and 7,978 (0.23%) were HDVs. From the light-duty fleet, 3,015,251 vehicles (85.40%) were transient-tested, while 515,403 (14.60%) were idle-tested.

A vehicle fails the transient emissions test when at least one of the reported emission values of HC, CO, or NOx, measured in grams per mile (g/mi), exceeds the appropriate composite cutpoint value for the entire driving cycle. However, such a vehicle will still pass the test if the reported emission values for the failed pollutant species do not exceed their respective phase 2 cutpoints for the last 146 seconds of the driving cycle.

It is important to remember that the test-based failure rates are not the sums of the failure rates from each of the three pollutant species, since the test-based failure rates may contain multiple species failures.

There were no cutpoint changes made during 2004. The final IM240 transient test cutpoints were implemented on April 1, 2003. The evaluated transient test failure rates by model year, for each vehicle type, for each pollutant species, as well as the evaluated test-based failures are provided in Appendix B Table 1, with Figure 2 graphically displaying the same information. The evaluated failure rates are summarized in Table 14.

Year 2004 LDV failure rates, for test-based as well as for all three species, for each of the pre-1996 model years, were lower than those from year 2003, as shown in Figure 3. For example, the LDV failure rates in year 2004 ranged from 23.09% for model year 1981 vehicles to 4.53% for model year 1995 vehicles, while in 2003 the rates ranged from 29.12% for model year 1981 vehicles to 4.54% for model year 1995 vehicles. Note that all vehicles under the same model year were one year older in 2004 with higher accumulated mileage. Since the vehicles were emissions tested under the same set of cutpoints, the probability of failing the tailpipe test in 2004 should be slightly higher for the same set of vehicles. These observed lower failure rates in 2004 can be attributed to two factors:

- Some vehicles, especially those failing the emissions tests in 2003 or those with marginal emissions test results, were no longer operating in NYMA in 2004. For example, there were 4,064 model year 1981 LDVs in 2003 database, the sample size has reduced to 2,993 in 2004.

- Vehicles remained in the year 2004 fleet, having been in I/M programs for multiple years, tend to be "cleaner" through maintenance and repair work performed either prior to the year 2004 emissions test or after failing the year 2003 emissions test.

The same phenomenon was observed for the LDTs, as shown in Figure 4. Between model years 1981 and 1995, the LDT failure rates from year 2004 were lower than those from 2003.

Relatively small differences for LDVs were noted when we compared the evaluated and the reported failure rates as shown in Table 15. As in the past, the discrepancies between the evaluated and the reported for the LDTs were due to the fact that vehicles' "sales class" information was "incorrectly reported".

Table 16 provides the evaluated failure rates for LDVs and LDTs by month, for each pollutant species, as well as the test-based results. The test-based failure rates for LDVs are found to be slightly higher during the warm weather months (May to August) primarily due to slightly higher NOx failure rates. Overall, the month-month changes in failure rates are moderate.

A summary of transient test failure rates by equipment vendor is presented in Table 17. Among the three vendors, SPX had the lowest test-based failure rate for LDVs, while SUN had the lowest test-based failure rate for LDTs.

Finally, a summary of transient test failure rates by county is given in Table 18. The highest failure rates for LDVs are seen in Bronx, Queens, and Rockland counties. The highest failure rates for LDTs are Suffolk, Rockland, and Queens counties.

IV. Idle Tests

In NYMA, Idle tests are required for all of the following gasoline-powered vehicles:

• LDVs and LDTs that are pre-1981 model years but not older than 25 years;
• LDVs and LDTs that are 1981+ model years and older than 2 years old but equipped with four-wheel drive or non-disengageable traction control; or,
• HDVs, defined by a gross vehicle weight rating of 8,501 pounds or more.

For an idle test, the HC emissions are measured in parts-per-million (ppm) and CO in percent (%). A vehicle fails the idle emissions test when at least one of the reported emission values (for HC or CO) exceeds the appropriate cutpoint value. As with the transient test, the idle failure rates are not the sums of the failure rates for each pollutant species because the test-based idle failure rates can consist of multiple species failures.

Appendix B Table 2 provides the evaluated failure rates by model year, for each vehicle type (LDVs, LDTs, and HDVs), and for each pollutant species (HC and CO), as well as the test-based failure rates. Figure 5 graphically displays these same results. The evaluated failure rates are summarized in Table 19.

It is noted that the percentage of vehicles having an idle test (instead of a transient test) increased, from 8.70% for model year 2000 to 14.56% for model year 2002 for LDVs, and from 33.38% for model year 2000 to 49.29% for model year 2002 for LDTs, for example. This could be the result from an upsurge in the popularity of all wheel drive vehicles, especially for LDTs.

Table 20 compares the reported and evaluated failure rates. As can be seen, the discrepancies between the reported and the evaluated were small, less than 1% for all three vehicle types.

Table 21 provides the evaluated failure rates for LDVs, LDTs, and HDVs by month, for each pollutant species, as well as on a test basis. No significant month to month variations are noted.

A summary of idle test failure rates by vendor is presented in Table 22. Among the three vendors, SPX had the lowest failure rates, for all pollutant species as well as test-based, for both LDVs and LDTs, while SUN had the lowest failure rates for HDVs.

Finally, a summary of the idle test failure rates by county are given in Table 23. The highest failure rates were observed in Kings, Bronx, Queens counties for LDVs, in Suffolk, Kings, and Queens counties for LDTs, and, in Bronx, Suffolk, and Westchester counties for HDVs.

V. Conclusions

As 2004 is the first year that the OBD II inspection program was implemented in NYS (NYMA, UPSTATE, and TLC), the failure rates obtained and reported in this report appeared to be high due to a combination of software issues and some inspections not being conducted "properly." We expect the software issues to be resolved in year 2005 for both the NYVIP and TLC.

It is important for the testing facilities (and vehicle owners) to be fully aware of the fact that vehicles must complete certain driving cycles for certain monitors (such as catalyst, evaporative system, and O₂ sensor) to be ready for the OBD inspections. It is not a "true" failure for a vehicle to fail the OBD on readiness criteria alone. It is evident that the failure rates reported from the TLC data doubled after adding in the readiness failures. In fact, 60.18% of the UPSTATE failing OBD inspection vehicles failed on readiness alone.

Another problem common to all OBD II inspection programs in NYS is the elevated count of "MIL Off KOEO". It is easy for technicians to miss the momentary illumination of the MIL thus counting it as a failure. Additional technician training emphasis on this problem and program experience can and probably will reduce false failures from this category. With the adoption of these remedies, it is expected that in future years the reported OBD failure rates will be lower than those observed in 2004.

The tailpipe exhaust tests, transient and idle, are being greatly reduced in number by the implementation of the OBD II inspection program in the NYMA as the new I/M program requirement. Light duty vehicles and trucks, model year 1996 and newer, are now subjected to the OBD II inspections and thus will no longer require either of the tailpipe exhaust tests. The light duty vehicles and trucks still receiving either tailpipe test will be a fleet with older model years (1981 to 1995). It is expected that both the transient and idle reported failure rates will rise in the future. And this will be borne out if the restricted older light duty fleet is compared to the entire light duty fleet of previous years. However, a comparison of year 2004 with year 2003 reported failure rates for the same model year fleet has shown a decrease, most likely due to the I/M program effects of repair, both required and anticipatory, and removal of vehicles from the fleet through attrition.

Appendix A

* Vehicles failed on any one or combination of the following criteria: no communication, MIL off KOEO, MIL on KOER, Mil commanded on with DTC, and monitors not ready (3 or more for 1996-2000 model years, 2 or more for model years 2001+).

** Communication failure was reported as a "S" (stopped test) in the beta ESP/SPX software versions.

* A/C monitor is not listed in this table since it is not supported.

* Vehicles failed on any one or combination of the following criteria: no communication, MIL off KOEO, MIL on KOER, Mil commanded on with DTC, and monitors not ready (3 or more for 1996-2000 model years, 2 or more for model years 2001+).

** Communication failure was reported as a "S" (stopped test) in beta ESP/SPX software versions.

* A/C monitor is not listed in this table since it is not supported.

* Vehicles failed on any one or combination of the following criteria: no communication, MIL off KOEO, MIL on KOER, Mil commanded on with DTC, and monitors not ready (3 or more for 1996-2000 model years, 2 or more for model years 2001+).

* A/C monitor is not listed in this table since it is not supported.

* Vehicles failed on any one or combination of the following criteria: no communication, MIL off KOEO, MIL on KOER, Mil commanded on with DTC, and monitors not ready (3 or more for 1996-2000 model years, 2 or more for model years 2001+).

* A/C monitor is not listed in this table since it is not supported.

* Vehicles failed on any one or combination of the following criteria: MIL off KOEO, MIL on KOER, Mil commanded on with DTC, and "monitors not ready." Vehicles under "no communication" were reported as "N" ("Not Tested"), hence were not included in the sample count. In addition, it is uncertain how the monitor readiness criteria were determined by TLC's software.

* A/C monitor is not listed in this table since it is not supported.

** Based on the evaluated sample count.

*Evaluated counts of test failures include no communication.

* Vehicles failed on any one or combination of the following criteria: MIL off KOEO, MIL on KOER, Mil commanded on with DTC, and "monitors not ready." Vehicles under "no communication" were reported as "N" ("Not Tested"), hence were not included in the sample count. In addition, it is uncertain how the monitor readiness criteria were determined by TLC's software.