02-7960. General Motors Corporation; Notice of Appeal of Denial of Petition for Determination of Inconsequential Noncompliance

General Motors Corporation (GM), of Warren, Michigan, has appealed a decision by the National Highway Traffic Safety Administration (NHTSA) that denied its application for a decision that its noncompliances with Federal Motor Vehicle Safety Standard (FMVSS) No. 108, “Lamps, Reflective Devices, and Associated Equipment,” be deemed inconsequential to motor vehicle safety.

Notice of receipt of the petition was published in the Federal Register on August 14, 2000, (65 FR 49632). On July 23, 2001, NHTSA published a notice in the Federal Register denying GM's petition, stating that the petitioner had not met its burden of persuasion that the noncompliance is inconsequential to motor vehicle safety.

This notice of receipt of GM's appeal is published in accordance with NHTSA regulations (49 CFR 556.7 and 556.8) and does not represent any agency decision or other exercise of judgment concerning the merits of the appeal.

GM manufactured 201,472 Buick Century and Buick Regal models between October 1998 and June 1999, some of whose headlamps do not meet the photometric requirements in FMVSS No. 108 for test points above the horizontal (intended for overhead sign illumination). To evaluate the noncompliance, GM randomly collected 10 pairs of lamps from production and photometrically tested them. Additionally, GM tested the same 10 pairs of lamps using accurately-rated bulbs. These are bulbs that have their filaments positioned within strict tolerances. In large-scale bulb production, the filament positions vary slightly and, therefore, can produce varying photometric output. The photometric output of a lamp using an accurately-rated bulb is intended to closely represent the output that was intended in its design, and not that which would occur in a mass-produced headlamp as sold on motor vehicles.

The test results indicated that five test points (production bulbs) and three test points (accurately-rated bulbs), respectively, failed to meet the minimum candela requirements. The test results also indicated that the amount of light below the minimum required was generally less than 10 percent at all noncomplying test points. However, seven failures at certain test points that were greater than 16 percent below the minimum, with the maximum variation being 24.4 percent (at 1.5 degrees up) with a production bulb. Transport Canada conducted tests on headlamps used on the same types of vehicles, and found that all the test points in question met the requirements. GM believes that these results show the noncomplying results were related to manufacturing variations and were present in only a portion of the lamps.

GM supported its application for inconsequential noncompliance with the following statements:

The test points at issue are all above the horizon and are intended to measure illumination of overhead signs. They do not represent areas of the beam that illuminate the road surface, and the headlamps still fulfill applicable Federal Motor Vehicle Safety Standard 108 requirements regarding road illumination.

For years the rule of thumb has been that a 25 percent difference in light intensity is not significant to most people for certain lighting conditions.

GM has not received any complaints from owners of the subject vehicles about their ability to see overhead signs.

GM is not aware of any accidents, injuries, owner complaints or field reports related to this condition for these vehicles.

GM also cited a number of inconsequentiality applications that the agency has granted in the past as support for granting its application. Those cited were submitted by GM [59 FR 65428; December 19, 1994], Subaru of America, [56 FR 59971; November 26, 1991], and Hella, Inc. [55 FR 37602; September 12, 1990]. GM also cited a University of Michigan Transportation Research Institute (UMTRI) report entitled “Just Noticeable Differences for Low-Beam Headlamp Intensities” (UMTRI-97-4, February 1997)

In the only public comment received, Advocates stated its “strongest opposition to NHTSA granting a finding of inconsequential noncompliance for the GM headlamps which are the subject of this notice.” Advocates first pointed out that it believes GM's purported lack of complaints about inadequate headlamp illumination has “no merit whatever.” It believes that it is unlikely that drivers would attribute their driving errors or crashes to a faulty beam. Further, it believes it unlikely that an investigating officer at a crash scene would consider the characteristics of the beam pattern as the causal factor. It goes on to say that crashes may have occurred as a result of the noncompliance of which GM is not aware.

Advocates also discussed the importance of overhead lighting. It stated that:

It is especially crucial for adequate levels of lighting to fall on the surfaces of high-mounted retroreflectorized traffic control devices that advise of vehicle maneuvers, speed limit changes, warnings of hazardous conditions, and destination information to ensure driver confidence and safety in executing the moment-to-moment driving task.

Advocates referred to the amendment of FMVSS No. 108 on January 12, 1993 [58 FR 3856] that added minimum photometric requirements for headlamps for illumination of overhead signs. Advocates reiterated the agency's rationale for this rulemaking, namely that some manufacturers were introducing headlamps in the 1980s and 1990s that widely departed from the traditional U.S. beam pattern. These headlamps were providing inadequate light above the horizontal to illuminate overhead signs.

After review of its application the agency disagreed with GM that the noncompliances were inconsequential to motor vehicle safety. As Advocates correctly noted in its comment, the sole purpose of the 1993 final rule was to establish photometric minima above the horizon so that headlamps would sufficiently illuminate overhead signs. Without any test point minima specified, some manufacturers were designing headlamps that provided very little light above the horizon. Because States were choosing retroreflectorized overhead signs rather than the more expensive self-illuminated ones, the agency determined that it should address the increasing need for illumination of overhead reflectorized signs.

In setting these minima, the agency expected the industry to design its headlamps to ensure that production variability would not result in noncompliances. GM's own compliance tests showed failures that were as much as 24.4 percent below the required minima. Each of the ten headlamps GM tested had noncomplying test points, with all but two having failures that were greater than 14.1 percent below the minimum requirement. This testing indicated that there may be a serious flaw in the design and/or production of these lamps.

Although GM stated that Transport Canada tested and found all lamps to be compliant, the company did not provide any substantiating data, or even the number of headlamps tested by Transport Canada. The agency contacted Transport Canada and obtained the test data on the subject vehicles. Initially, there were four failures at the relevant test points. The failures were resolved by reaiming the headlamps one-quarter degree, an adjustment allowed by the standard. After reaiming, Transport Canada found the lamps to be in compliance at the four test points where they had previously failed. Although these four lamps were found to be in compliance, the need to reaim certain points and the marginal compliance at others shows that the design of the lamps was marginal.

A January 1991 study conducted by UMTRI (UMTRI-91-3) recommended certain minimum intensity levels for test points above the horizontal that are intended to illuminate signs. UMTRI divided its recommendations for minima between three types of retroreflectorized signs: enclosed lens, encapsulated lens, and microprismatic, each respectively more reflective than the previous. The first two are most relevant, as microprismatic signs comprised only about three percent of the current signs at that time. UMTRI concluded that, for a test point 1.5 degrees up, the minimum intensities for the enclosed and encapsulated lens signs were 700 and 250 candela (cd), respectively. The standard currently requires a minimum of 200 cd. In setting this level, the agency expected manufacturers to factor in a certain level of design variability to assure compliance. GM's poorest performing lamp provided about 150 cd at this test point. The agency finds this unacceptable. As Advocates pointed out in its comments, there are many critical maneuvers that must be undertaken in low light situations, and to not provide sufficient light to illuminate signs is a detriment to motor vehicle safety.

GM cited a number of the agency's previous grants of inconsequentiality applications that were based upon our conclusion that a change in luminous intensity of approximately 25 percent must occur before the human eye can discern a difference. GM also cited an UMTRI report [UMTRI-97-4; February 1997] to support its position.

The agency determined that these actions and the 1997 UMTRI report did not support GM's conclusion. The previous actions and the UMTRI report all dealt with an observer's ability to see a headlamp or a signal light, not the ability to see the light reflected back from headlamp-illuminated signs or other reflectors. The inconsequential applications that GM cited all involved signal lighting with deficiencies in photometric requirements. In all cases, the agency was confident that the noncompliant signal lights would still be visible to nearby drivers. Because signal lighting is not intended to provide roadway illumination to the driver, a less than 25 percent reduction in light output at any particular test point is less critical.

Regarding the UMTRI study on just-noticeable differences for lower-beam headlamps, the research and findings are mostly analogous to those of the signal lighting research. UMTRI's study was designed to evaluate the just-noticeable differences for glare intensities of oncoming headlamps. Like the signal light research, it was performed from the point of view of a driver observing differences in headlamp intensities. The agency was not persuaded by GM's contentions about the meaning of this research. In its report, UMTRI states:

The applications of (just noticeable differences) derived from judgments about the subjective brightnesses of lamps viewed directly seems less of a leap in the case of signal lamp functions, and of those aspects of headlamps that involve direct viewing (primarily discomfort glare), than in the case of headlamp functions that involve the illumination of objects. The primary reason for caution in extending the current results to illuminated objects is that the range of luminances of such objects (e.g., a pedestrian at 100 meters illuminated by headlamps at night) will be much lower than the luminances of the headlamps themselves. The [research] can therefore be used more confidently to justify applying the 25 percent limit for inconsequential noncompliance to a photometric test point that specifies a maximum for glare protection than to one that specifies a minimum for seeing light. Further work on the effects of changes in lamp intensity on the visibility of illuminated objects is desirable to clarify more completely the issue of inconsequential noncompliance for headlamps.

In its appeal, GM offers this new information to support its petition:

GM recently obtained and tested twenty-one pairs of headlamps from used 1999 Regal and Century vehicles built between August 1998 and March 1999. The 42 headlamps all exceed the minimum photometric requirements of FMVSS 108. This was true for the sign illumination test points as well as all other test points. [GM stated that t]he weathering of the lenses over the past two to three years accounts for this change in performance.

Because overhead sign illumination is affected by the output of both headlamps, GM asked two independent lighting research experts to analyze overhead sign illumination based on the test results of the ten pairs of headlamps. Their report shows that the combined sum of the illumination from any combination of two of those headlamps exceeds twice the minimum illumination from each headlamp required by FMVSS 108. The system light output, therefore, exceeds the implicit functional requirement of the standard.

This evidence, which [GM describes] in greater detail below, indicates that customers driving these vehicles are and have been experiencing no less than the amount of overhead sign illumination that FMVSS 108 requires. On this basis, the noncompliance is inconsequential and [thus, GM requested] reconsideration of NHTSA's decision.

Photometric Test Data From Field Headlamps

GM collected 42 headlamps from twenty-one vehicles and all photometric test points were measured. Each bulb appeared to be the original bulb for the headlamp assembly and the bulbs were not disturbed before testing. Visual aim was used because of the condition with the operation of the VHAD that lead to a recall campaign (NHTSA No. 99V356000, GM No. 99093).

The vehicles were produced between August 18, 1998 and February 15, 1999. Three of the vehicles were owned by GM employees and eighteen were selected at random at auto auctions in Detroit and Flint, Michigan. All 42 headlamps exceeded the minimum photometric requirements for the sign illumination test points found in FMVSS 108 (as summarized below).

Test point	Requirement (Candela)	Average (Candela)	Range (Candela)
Left Headlamp:
0.5U, 1R-3R	500	674	501-1214
4U-8L	64	114	88-148
4U-8R	64	91	64-125
2U-4L	135	159	136-198
Right Headlamp:
0.5U, 1R-3R	500	895	577-2679
4U-8L	64	82	64-107
4U-8R	64	135	109-196
2U-4L	135	308	274-346

[GM's] hypothesis was that these results were caused by weathering of the lens coating, which increases light scatter. Weathering is caused by exposure to temperature changes, precipitation, and contact with dust, stones, and other environmental factors. This is a well-known phenomenon that occurs in lamps that meet fully the haze requirement in S5.1.2, as these lamps do. To test our hypothesis, the lenses of four of the tested lamps were removed and replaced with a new, unused lens. The photometric results with the original and new lenses were:

Test point	Requirement Average	Average with new lenses (Candela)	Average with original lenses (Candela)	Percent change
Left Headlamp:
0.5U, 1R-3R	500	577	632	8.7
4U-8L	64	87	117	25.6
4U-8R	64	72	122	40.9
2U-4L	135	126	183	31.1
Right Headlamp:
0.5U, 1R-3R	500	957	864	−10.7
4U-8L	64	74	90	17.7
4U-8R	64	128	154	16.9
2U-4L	135	263	289	9.0

Using the averages, the results for the original lenses exceeded those for the new lenses for all but one test point.

In the group of 42 lamps, [GM] also compared the performance of the lamps from the ten newest and eleven oldest vehicles. No significant difference was observed.

Because of weathering, the headlamps on these vehicles now meet the photometric requirements that some of the new headlamps did not meet. The noncompliance of the new, unused lamps is, therefore inconsequential.

Combined Light Output From Left and Right Low-beam Headlamps

The test point values for each headlamp were set by NHTSA to achieve a certain overall level of sign illumination. 58 FR 3856, 3858 (Jan. 12, 1993). At least two headlamps are required by the standard. To assess the impact of the noncompliance on the illumination of overhead signs, one should examine the light output of both headlamps. [GM] asked two well-known researchers in the field of vehicle lighting to do so.

Their analysis was based on the 1999 photometric data from an independent test laboratory for ten pairs of headlamps with production bulbs. The combined light output from a left and a right headlamp was calculated for three different scenarios:

Worst case: The worst performing left lamp was paired with the worst performing right lamp. For each test point, the worst case headlamps were selected separately.

Best case: As above, but using the best performing left and right headlamps.

Average case: The mean values were paired for the left and right headlamps.

The result, even in the worst case scenario, is illumination of overhead signs that is greater than twice the minimum photometric requirements for a single headlamp. When pairing the worst performing left and right headlamps, the combined light exceeded twice the requirement by 20% for 4U-8R, 6% for 4U-8L, 45% for 2U-4L, 26% for 1.5U-1R to 3R, and 11% for 0.5U-1R to 3R. The points at which left and right lamps failed were consistently different, so the margin by which each exceeded the points at which they passed offset the failures when the results are combined.

Consistent with FMVSS 108, these vehicles could have been equipped with left and right headlamps that each precisely met (but did not exceed) the overhead sign illumination test point requirements. While some of these vehicles were equipped with lamps that did not meet some of the individual test points (and exceed others), the overhead sign illumination from these vehicles is no less than what is lawful. Indeed, the requirements are exceeded by six to forty-five percent for the worst case.

In denying the petition, NTHSA noted that it expected manufacturers to account for design variability. GM's design and performance requirements do account for expected variability to assure compliance. In this instance, variability exceeded reasonable expectations and a noncompliance occurred. When the light that can reach overhead signs from both headlamps on these vehicles is considered, the performance not only meets the implied requirement, but meets it with a margin. This demonstrates that the noncompliance is inconsequential.

Interested persons are invited to submit written data, views, and arguments on the application appealing NHTSA's decision described above. Comments should refer to the docket number and be submitted to: U.S. Department of Transportation, Docket Management, Room PL-401, 400 Seventh Street, SW., Washington, DC 20590. It is requested that two copies be submitted.

All comments received before the close of business on the closing date indicated below will be considered. The application and supporting materials, and all comments received after the closing date, will also be filed and will be considered to the extent possible. When the application is granted or denied, the notice will be published in the Federal Register pursuant to the authority indicated below. Comment closing date: May 2, 2002.

(49 U.S.C. 301118, 301120; delegations of authority at 49 CFR 1.50 and 501.8)

Issued on: March 28, 2002.

Stephen R. Kratzke,

Associate Administrator for Safety Performance Standards.