III. TIRE PRESSURE SURVEY AND TEST RESULTS

In February 2001, the agency conducted a tire pressure study to determine the extent to which passenger vehicle operators are aware of the recommended air pressure for their tires, if they monitor air pressure, and to what extent the actual tire pressure differs from that recommended tire pressure by the vehicle manufacturer on the placard. The most useful information for this analysis is the snap shot in time that tells us where the actual tire pressure of the fleet is in comparison to the vehicle manufacturer's recommended tire pressure. Although this was not a nationally representative survey, it is being treated as such in this analysis.

The field data collection was conducted through the infrastructure of 24 locations of the National Automotive Sampling System Crashworthiness Data System (NASS CDS). Data were collected on 11,530 vehicles that were inspected at a sample of 336 gas stations. There were 6,442 passenger cars, 1,874 sport utility vehicles (SUVs), 1,376 vans, and 1,838 light conventional trucks. Data can be separated by passenger cars with P-metric tires; trucks, SUVs and vans with P-metric tires; and trucks, SUVs, and vans with either LT-type or high flotation tires. For this analysis we only compare the passenger car tire pressures and the light truck tire pressures, without separating the light trucks by type of tire. Complete data were collected on 5,967 passenger cars and 3,950 light trucks for a total of 9,917 vehicles.[1]

The average placard pressure for passenger cars was about 30 psi, while the average placard pressure for light trucks was about 35 psi, although the light trucks have a much wider range of manufacturer recommended placard pressure. Because of the wide range of placard pressure for light trucks, it was determined that it would be best to propose a percentage reduction from the placard than a straight psi reduction.

The issue addressed is how often drivers would get a warning from a low tire pressure monitoring system. Three scenarios were examined for a direct measurement system, as shown in Table III-1. These assume the driver would be warned anytime one or more tires fell 20%, 25%, or 30% below the placard recommended pressure.

For Alternative 1, an average of 38 percent of the passenger car and light truck drivers in the tire pressure survey would get a warning with a direct measurement system that activated at 20 percent or more below the placard pressure.

For Alternative 3, a 25 percent differential for one to four tires is essentially the same as 25 percent or more below placard with the hybrid system, since you would have two tires with direct measurement systems that the other two wheel speeds could be compared to. Thus, an average of 27.5 percent of the passenger car and light truck drivers would get a warning with the hybrid system.

Table III-2 (a) shows, for example, the distribution of tire pressure when at least one tire is 20 percent or more below placard in terms of whether one, two, three, or all four tires were at least 20 percent below placard. Tables III-2 (b) and (c) show similar results for 25 percent and 30 percent below placard.

Table III-2 (d) provides an example of the grouping of pressure in tires on the same vehicle. This table says: given that all four tires are 25 percent or more below placard (which occurs in 26 percent of passenger cars and 29 percent of light trucks), what is the distribution of pressure in the four tires. This was developed by comparing the tire with the maximum pressure to the tire with the minimum pressure divided by the maximum pressure [(max-min)/max]. This provides information to the question of how often an indirect system cannot detect that all four tires are 25 percent below placard. Assuming an indirect system using a 25 percent differential in tire pressure, for passenger cars, 70 percent of the vehicles with four tires more than 25 percent below placard probably would not have gotten a warning [or 18.2 percent of all passenger cars (26 percent * 70 percent)], since the differential in tire pressure was not above 25 percent. For light trucks with four tires more than 25 percent below placard, 84 percent would not have gotten a warning [or 24.4 percent of all light trucks (29 percent * 84 percent)], since the differential in tire pressure was not above 25 percent.

At the time the survey was done, there were 207 million vehicles on the road. An estimated 4.7 million vehicles, which had all four tires 25 percent or more below placard, would not be identified as having low tire pressure by an indirect system that could not determine when there was never a difference of more than 25 percent in pressure between any of the tires. The agency does not know how often tires lose pressure at different rates such that a 25 percent differential in tire pressure could have occurred to one tire sometime before the survey took place and the owner could have been warned before all four tires were 25 percent below placard. However, we don't believe this will take place in a significant number of cases. [2]

Table III-3 (a) provides an analysis of what percent of the drivers would get a warning with an indirect measurement system that compares relative wheel speed of the four wheels. An assumption was made that if wheel speed were measured in all four wheels (an upgrade for some vehicles), then a comparison of wheel speed could be made for all situations except when all four tires lose air at about the same rate. Thus, we label it as (one to three tires). For analytical purposes we used from our tire pressure survey (maximum tire pressure minus the minimum tire pressure) divided by the maximum tire pressure to get an average reduction. The maximum tire pressure was used as the denominator since supposedly we are starting at placard tire pressure and decreasing tire pressure from there. Since the indirect systems use a relative measurement, it cannot tell whether the tire pressure is over placard or under placard. For the benefit analyses done in this assessment, cases were not considered in which there were a relative differential in tire pressure of 25 percent or more, yet none of the tires were below placard. Thus, for example, if placard pressure was 30 psi, and the four tire pressures were 30, 30, 30, and 60 psi, this case was not included in the benefit calculations.

For Alternative 2, an average of 24 percent of the passenger car and light truck drivers in the tire pressure survey would get a warning with an indirect measurement system that activated at 25 percent or more differential in wheel speed for one to three tires.

The current indirect measurement systems (which can determine relative differential in tire pressure of about 30%), give a warning about 10 percent of the time. The current indirect measurement systems that work on relative wheel speed cannot pick up when all four tires have lost air at about the same rate. In addition, the current systems do not always provide a warning when two tires are low in pressure. The agency analyzed the various algorithms used by the manufacturers to determine the percent of cases in which they would give a warning.

In summary, based on the tire pressure survey the agency conducted:

Alternative 1: a direct measurement system would result in 38 percent of the light vehicles operators being notified of low tire pressure.

Alternative 2: a hybrid (direct and indirect) or an improved ABS-based measurement system comparing one to three tires would result in 24 percent of the light vehicles operators being notified of low tire pressure. [Note that low tire pressure is defined as a differential in tire pressure in this case, which is different for each system.]

Alternative 3: a hybrid (direct and indirect) ABS-based measurement system comparing one to four tires would result in 27.5 percent of the light vehicles operators being notified of low tire pressure.

Alternative 4: an indirect ABS-based measurement system was analyzed to determine how often a warning would be given with the current indirect systems. It was assumed that a warning would be given when one or three tires were 30 percent or more different than the other tires or when two tires in particular cases (e.g. on the diagonal tires as the current systems are designed) were 30 percent or more different than the other two tires on the diagonal. The results are that 11 percent of the passenger car operators and 8 percent of the light truck operators would be notified of low tire pressure [see Table III-3 (b)].

Table III-1
Percent of Vehicles That Would Get a Warning
Assuming a Direct Measurement System
  Passenger Cars Light Trucks
20% or more Below Placard 36% 40%
25% or more Below Placard 26% 29%
30% or more Below Placard 20% 20%

Table III-2 (a)
Distribution of the Number of Tires on Vehicles
That Have One or More Tires that are
20% or more Below Placard
Number of Tires 20% or more Below Placard Passenger Cars Percent Light Trucks Percent
1 994 46.5% 574 36.7%
2 548 25.7 440 28.1
3 275 12.9 223 14.3
4 319 14.9 327 20.9
Total 2,136 100% 1,564 100%

Table III-2 (b)
Distribution of the Number of Tires on Vehicles
That Have One or More Tires that are
25% or more Below Placard
Number of Tires 25% or more Below Placard Passenger Cars Percent Light Trucks Percent
1 880 55.9% 542 47.2%
2 399 25.3 313 27.3
3 139 8.8 145 12.6
4 157 10.0 148 12.9
Total 1,575 100% 1,148 100%

Table III-2 (c)
Distribution of the Number of Tires on Vehicles
That Have One or More Tires that are
30% or more Below Placard
Number of Tires 30% or more Below Placard Passenger Cars Percent Light Trucks Percent
1 793 66.1% 454 57.6%
2 266 22.2 199 25.2
3 88 7.4 72 9.1
4 52 4.3 64 8.1
Total 1,199 100% 789 100%

Table III-2 (d)
Distribution of Tire Pressure in Vehicles with
All Four Tires at least 25 Percent Below Placard
  Passenger Cars Light Trucks
< 25% Difference 70% 84%
> 25% Difference 30% 16%
     
% of all Vehicles in the Survey that a 25% Indirect System Could Not Identify that had all 4 tires 25% below placard
(Calculation)
1.8%
(26%*10.0%*70%)
3.1%
(29%*12.9%*84%)
Number of Vehicles on the Road in 2000 128 million 79 million
Number of Vehicles Not Identified by 25% Indirect System 2.3 million 2.4 million

Table III-3 (a)
Percent of Vehicles That Would Get a Warning
Assuming an Indirect Measurement System
(one to three tires)
  Passenger Cars Light Trucks
25% Differential 27% 21%
30% Differential 22% 16%

Table III-3 (b)
Percent of Vehicles That Would Get a Warning
Assuming the Current Indirect Measurement System
(one or three tires in all cases, two tires in some cases)
  Passenger Cars Light Trucks
30% Differential 11% 8%

Table III-4
Percent of Vehicles That Would Get a Warning by Alternative
  Passenger Cars Light Trucks
Alt. 1 (20% Below Placard, 1-4 tires) 36% 40%
Alt. 2 (25% Below Placard, 1-3 tires) 27% 21%
Alt. 3 (25% Below Placard, 1-4 tires) 26% 29%
Alt. 4 (30% Below Placard, 1 tire) 11% 8%

TPMS Test Results

The agency tested six direct measurement systems to determine both the level at which they provided driver information and the accuracy of the systems. The warning level thresholds were determined by dynamic testing at GVWR at 60 mph by slowly leaking out air to a minimum of 14 psi. Some of the systems provide two levels of driver information, an advisory and a warning level. System F was a prototype with much lower thresholds for advisory and warning than the other systems. If System F is not considered, the typical advisory level is given at 20 percent under placard pressure, while the warning level averaged 36 percent below the placard. The static accuracy tests showed that those systems that displayed tire pressure readings were accurate to within 1 to 2 psi.

Table III-5
Direct measurement systems
Driver information provided at (%) below placard
System E F G H I J
Advisory N.A. -42% N.A. -20% N.A. -19%
Warning -20% -68% -33% -53% -35% -41%

The agency tested four indirect measurement systems to determine when they provided driver information. The warning thresholds were determined by slowly leaking out air to a minimum of 14 psi, while driving at 60 mph under a lightly loaded vehicle weight condition (LLVW) and at gross vehicle weight rating (GVWR). Table III-6 provides these results. The agency believes that the difference in the warning levels between the front and rear axle are due to variability in the system.

Table III-6
Indirect measurement systems
Driver warning provided at (%) below placard
Load Axle System A System B System C System D Ave. of 3
LLVW Front -21% No Warning -40% -28% -30%
LLVW Rear -16% No Warning -37% -38% -30%
GVWR Front -16% No Warning -18% -31% -24%
GVWR Rear -9% No Warning -20% n/a -14%

Vehicle Stopping Distance Tests

One of the potential safety benefits the agency is examining is the impact of low tire pressure on vehicle stopping distance. In the PEA we presented two sets of data from different sources Goodyear Tire and Rubber Company and NHTSA's Vehicle Research and Test Center (VRTC). In a comment to the docket Goodyear presented the results of additional testing. The information provided by these sources do not lead to the same conclusions.

Table III-7 shows data provided by Goodyear on an ABS vehicle. These wet stopping distance data indicate:

  1. Stopping distance generally increases with lower tire pressure. The only exception was on concrete at 25 mph.

  2. With fairly deep water on the road, (0.050 inches is equivalent to 1 inch of rain in an hour) lowering inflation to 17 psi and increasing speed to 45 mph increases the potential for hydroplaning and much longer stopping distances.

  3. Except for 25 mph on macadam, the difference between 25 and 29 psi is relatively small.

Goodyear provided test data to the agency on Mu values to calculate dry stopping distances. This information is used in the benefits chapter later in this assessment.

Table III-7
Braking Distance (in feet) provided by Goodyear
Wet Stopping Distance (0.050 water depth)
Surface Speed 17 psi 25 psi 29 psi 35 psi
Macadam 25 mph 32.4 30.8 29 27.4
Macadam 45 mph 107.6 101 100.8 98.6
Concrete 25 mph 47.4 48.2 48.2 48
Concrete 45 mph 182.6 167.2 167.4 163.6

Table III-8 shows test data from NHTSA - VRTC on stopping distance. Tests were performed using a MY 2000 Grand Prix with ABS. Shown is the average stopping distance based on five tests per psi level. The concrete can be described as a fairly rough surface that has not been worn down like a typical road. The asphalt was built to Ohio highway specifications, but again has not been worn down by traffic, so it is like a new asphalt road. A wet road consists of wetting down the surface by making two passes with a water truck, thus it has a much lower water depth than was used in the Goodyear tests.

Table III-8
Braking Distance (in feet) from NHTSA testing
Stopping Distance from 60 mph
Surface 15 psi 20 psi 25 psi 30 psi 35 psi
Wet Concrete 148.8 147.5 145.9 144.3 146.5
Dry Concrete 142.0 143.0 140.5 140.4 139.8
Wet Asphalt 158.5 158.6 162.6 161.2 158.0
Dry Asphalt 144.0 143.9 146.5 148.2 144.0

These stopping distances indicate:

  1. There is generally an increase in stopping distance as tire inflation decreases from the 30 psi placard on this vehicle on both wet and dry concrete.

  2. On wet and dry asphalt, the opposite generally occurs, stopping distance decreases as tire inflation decreases from the 30 psi placard.

  3. There is very little difference between the wet and dry stopping distance on the concrete pad (about 4 feet at 30 psi), indicating the water depth was not enough to make a noticeable difference on the rough concrete pad. There is a larger difference between the wet and dry stopping distance on the asphalt pad (13 feet at 30 psi).

  4. No hydroplaning occurred in the NHTSA tests, even though they were conducted at higher speed (60 mph vs. 45 mph in the Goodyear tests) and at lower tire pressure (15 psi vs. 17 psi in the Goodyear tests). Again, this suggests that the water depth in the VRTC tests was not nearly as deep as in the Goodyear testing.

In general, these data suggest that the road surface and depth of water on the road have a large influence over stopping distance. Given a specific road condition, one can compare the difference in stopping distance when the tire inflation level is varied. The Goodyear test results imply that tire inflation can have a significant impact on stopping distance, while the NHTSA testing implies these impacts would be minor or nonexistent on dry surfaces and wet surfaces with very little water depth.

In a comment to the docket (8572-160) Goodyear presented an extensive series of test data. These tests included two vehicles having tires with full tread depth and half tread depth on vehicles with ABS and on tires with full tread depth without ABS and on a dry, 0.02 inch wet and 0.05 inch wet macadam surface at three different psi levels. The full tread depth on the Integrity tire used on the Dodge Caravan was 10/32 inch and the half tread depth was 5/32 inch. The full tread depth on the Wrangler tire used on the Ford Ranger was 13/32 inch and the half tread depth was 6.5/32 inch. The stopping distance in feet is the average of six stops for most of the scenarios. The stopping distance is from 45 mph to 5 mph. Goodyear found that collecting the data at 5 mph reduced the variability in the results as compared to a full stop. Tables III-9 (a), (b), and (c) summarize these results.

Table III-9 (a)
Goodyear data Second Test Series
Dry Macadam Surface
(Stopping Distance in Feet)
2001 Dodge Grand Caravan Sport 20 psi 28 psi 35 psi
Full Depth Tread with ABS 75.5 76.2 75.8
Depth Tread with ABS 69.9 68.1 66.3
       
Full Depth Tread without ABS 98.3 95.9 91.6
       
1997 Ford Ranger      
Full Depth Tread with ABS 80.8 78.2 77.6
Depth Tread with ABS 79.0 74.8 71.4
       
Full Depth Tread without ABS 97.8 96.5 94.1

Table III-9 (b)
Goodyear data - Second Test Series
0.02 Inch Wet Macadam Surface
(Stopping Distance in Feet)
2001 Dodge Grand Caravan Sport 20 psi 28 psi 35 psi
Full Depth Tread with ABS 79.8 78.5 77.1
Depth Tread with ABS 84.7 73.7 81.4
       
Full Depth Tread without ABS 111.1 110.2 108.6
       
1997 Ford Ranger      
Full Depth Tread with ABS 83.8 81.5 79.8
Depth Tread with ABS 91.5 89.4 84.6
       
Full Depth Tread without ABS 131.9 126.0 118.4

Table III-9 (c)
Goodyear data Second Test Series
0.05 Inch Wet Macadam Surface
(Stopping Distance in Feet)
2001 Dodge Grand Caravan Sport 20 psi 28 psi 35 psi
Full Depth Tread with ABS 80.0 81.1 82.7
Depth Tread with ABS 103.7 99.7 92.2
       
Full Depth Tread without ABS 118.0 112.2 111.7
       
1997 Ford Ranger      
Full Depth Tread with ABS 89.7 86.0 81.5
Depth Tread with ABS 125.7 118.5 104.5
       
Full Depth Tread without ABS 142.9 134.8 125.7

These data indicate that stopping distance is longer with lower psi for every case except for two cases with the full depth tread with ABS on the Dodge Caravan. Full depth tread tires had shorter stopping distance than depth tread tires on wet surfaces, but not dry surfaces, and vehicles with ABS had shorter stopping distances than those vehicles without ABS.

The value of Mu is dependent on surface material (concrete, asphalt, etc.), surface condition (wet vs. dry), inflation pressure, and initial velocity. The following tables presents data provided by The Goodyear Tire and Rubber Company in response to the NPRM. NHTSA developed a model that predicts Mu based on Vi and inflation pressure. Separate models were developed for Mu at both peak (the maximum level of Mu achieved while the tire still rotates under braking conditions) and slide (the level of Mu achieved when tires cease to rotate while braking (i.e., skid)). These models are used in the benefits section when estimating stopping distance.

COEFFICIENT OF FRICTION DATA - m
Macadam Surface
215/70R15 Integrity - 1080 lbs. Load
  0.020" Wet   0.050" Wet   DRY
  20 mph   20 mph   20 mph
  35 psi 28 psi 20 psi   35 psi 28 psi 20 psi   35 psi 28 psi 20 psi
Peak 0.864 0.846 0.818   0.830 0.795 0.796   0.980 0.992 0.966
Slide 0.566 0.546 0.528   0.553 0.512 0.497   0.716 0.671 0.648
                       
                       
  0.020" Wet   0.050" Wet   DRY
  40 mph   40 mph   40 mph
  35 psi 28 psi 20 psi   35 psi 28 psi 20 psi   35 psi 28 psi 20 psi
Peak 0.827 0.808 0.786   0.740 0.687 0.690   0.940 0.926 0.921
Slide 0.474 0.454 0.448   0.444 0.416 0.397   0.696 0.696 0.682
                       
                       
  0.020" Wet   0.050" Wet   DRY
  60 mph   60 mph   60 mph
  35 psi 28 psi 20 psi   35 psi 28 psi 20 psi   35 psi 28 psi 20 psi
Peak 0.832 0.831 0.802   0.564 0.484 0.488   0.930 0.910 0.923
Slide 0.368 0.373 0.348   0.280 0.220 0.148   0.730 0.737 0.766

NHTSA - TIRE COEFFICIENT OF FRICTION DATA - m
Macadam Surface
P235/75R15 Wrangler RT/S - 1490 lbs. Load
                       
  0.020" Wet   0.050" Wet   DRY
  20 mph   20 mph   20 mph
  35 psi 28 psi 20 psi   35 psi 28 psi 20 psi   35 psi 28 psi 20 psi
Peak 0.924 0.913 0.864   0.878 0.844 0.790   0.942 0.961 0.904
Slide 0.600 0.562 0.522   0.548 0.502 0.491   0.690 0.606 0.644
                       
                       
  0.020" Wet   0.050" Wet   DRY
  40 mph   40 mph   40 mph
  35 psi 28 psi 20 psi   35 psi 28 psi 20 psi   35 psi 28 psi 20 psi
Peak 0.888 0.848 0.808   0.800 0.752 0.708   0.916 0.882 0.834
Slide 0.466 0.465 0.440   0.422 0.382 0.347   0.618 0.631 0.620
                       
                       
  0.020" Wet   0.050" Wet   DRY
  60 mph   60 mph   60 mph
  35 psi 28 psi 20 psi   35 psi 28 psi 20 psi   35 psi 28 psi 20 psi
Peak 0.840 0.806 0.770   0.602 0.626 0.555   0.882 0.860 0.814
Slide 0.364 0.346 0.314   0.266 0.212 0.133   0.672 0.700 0.704



[1] The Rubbers Manufacturers Association (Docket 8572-116) argued the tire pressure survey measured tires when they were hot. Thus, NHTSA's under-inflation estimates are conservative. The agency considered this point, but also notes that the survey was done in February when tires lose more pressure because of the ambient temperature and considered these to be unquantifiable offsetting conditions.

[2] Tires tend to lose air on a consistent basis per month. However, different tires lose air at different rates. Changes in temperature also affect tires differently. Logically, the largest difference between tires will occur at the longest time from being filled up. When all four tires are more than 25 percent below placard, there has been a long time since the last fill up. It is not likely that the gap in pressure between the tires was more than 25 percent and then less than 25 percent at a later time. It is more likely that the gap will widen over time.