AlexCisneros

Picked this up on an autoX forum




SPORTS CAR CLUB OF AMERICA, INC
Technical Services Participant Bulletin
DATE: November 16, 2004
FROM: National Staff
TO: All Participants
SUBJECT: Airbag Advisory
It has been brought to the attention of SCCA Technical Services that the use
of full-face or closed-face helmets while driving vehicles with active airbag
restraint systems may result in injuries in the event of a crash that deploys
the airbag. Because of the location of the steering wheel relative to a
driver’s position, the airbag axis is on a level with the driver’s chin. In a
crash with airbag deployment, contact with the chin area of a full-face
helmet can be so powerful “that the risk of fractures to the jaw cannot be
ruled out“ (Hubert Gramling, FIA Institute, FT3/AF, 18.5.1999). This applies
to vehicle that may be used in Solo, RallyCross, High Performance Car
Control Clinics, etc.
Therefore, it is highly recommended that full-face helmets not be used in
vehicles with functional airbag systems. Potentially more restrictive language
is currently being considered for 2005, which could appear in an early 2005
issue of FasTrack. If you have any questions, please contact the SCCA at
(800) 770-2055.

AlexCisneros

...anyone know how to temp. disable the airbag?

ZoomZoomH

iTrader: (1)

wow, aren't i glad i have a SA2000 open face helmet *being cheap does have its rewards!*

clyde

Below is the referenced report as it was posted to another forum. Split into multiple parts for posting here.

Hubert Gramling
FT3/AF
18.5.1999

4th status report for the project
´ Driver safety system for Formula 1 ´

Working step: HANS sensitivity study
Airbag overall function and determination of potential

1 Abstract

A comparative investigation of the driver safety systems “airbag” and “HANS” was carried out. With both systems, head and neck loads were reduced from near-fatal values to values well below the injury threshold. Both systems performed equivalently. For this reason and given the high costs of development and testing, there is no justification for pursuing airbag development any further. The following tests were carried out:

Baseline tests:
It was shown that the locking button of a Bell Vortex E.C. helmet prevents the visor from opening in a frontal impact even when it is subjected to high opening forces.
Doubling the shoulder belts resulted in a reduction in the forward movement of the head by 50 mm, with a slight reduction in chest and pelvis acceleration. This reduces the impact between head and steering wheel and the resultant load. The reason for this effect is not so much the stiffening of the belt material but the fact that each of the four shoulder belts was individually tightened, thus achieving a higher belt pre-tensioning. The potential of belt-tensioners therefore needs to be investigated.
Belt slack results in a slight increase in chest and pelvis values. Neck and head loads were reduced due to the changed interaction with the steering wheel, but as tests by John Melvin show, this need not be the case.

Airbag tests:
The overall functioning of the Knoll airbag was demonstrated.
An optimization of the airbag was carried out in order to assess its protective potential. This resulted in very good head and neck values being achieved.

HANS tests:
HANS showed itself relatively insensitive to a change in neck length.

In the frontal impact with belt slack and in the angled impact, an increase in head load is observed compared with the frontal impact without belt slack. The increase remains below 10 % if the tether attachment points at the helmet are positioned 30 – 60 mm above the points of attachment to HANS.
In the angled impact, the neck loads with HANS match the values in the frontal impact. Furthermore head rotation, which occurred in the baseline test due to the angled impact with the raised side panel of the cockpit, was substantially reduced with HANS.
In the side impact, HANS had no influence apart from a slight reduction in head rotation.
In the rear-end impact, HANS produced a reduction in neck loads because the HANS collar prevents friction between the helmet and the head restraint and thus prevents the associated neck compression and rearward rotation of the head.
Some helmet models require reinforcement in the area where the HANS tethers are attached to the helmet, since in a test with a model ‘Bieffe Formula 1 – GP Compotech. K/C‘ helmet the attachment slots broke away. A model 'Arai GP3' helmet came through 10 frontal, 4 angled, one side and one rear-end impact without noticeable damage.
A frontal impact with both airbag and HANS produced the lowest forward movement of the head observed in any of the tests, with very low head and neck values. This proved to be due to the indirect belt-tensioning by the combination of airbag and HANS.


2 Content

Page
1 Abstract .................................................. ................................. 1
2 Content .................................................. .................................................. .... 2
3 Introduction .................................................. ................................................ 2
4 Baseline tests, frontal impact .................................................. ................ 3
5 Airbag tests, frontal impact .................................................. ............. 10
6 HANS tests, frontal impact .................................................. ............. 18
7 Angled, side and rear-end impact .................................................. .......... 25
8 Airbag/HANS comparison .................................................. .......................... 30



3 Introduction

In the last status report, a sheet metal HANS component had been built and successfully tested. Also, the basic functioning of an airbag in a Formula 1 environment was demonstrated.

Since the last status report of 20.1.1998, a HANS device made of composite material has been constructed and its sensitivity to changes in neck length and belt slack, and to angled impacts, has been ascertained. Additionally, side and rear impacts were studied. As far as the airbag is concerned, overall functioning of the Knoll airbag and its basic characteristics were demonstrated. Thus it is possible to compare its advantages and disadvantages with the HANS concept and thus to reach a decision about further development of the airbag.


4 Baseline tests, frontal impact

After 20 tests, cracks appeared in the monocoque. These had to be repaired and the steel cradle reinforced. Due to the now increased mass, preliminary tests had to be carried out to adapt the acceleration characteristics. Given the uncertainty factor, it was not possible to use these preliminary tests in the test program but it was nevertheless possible to use them to explore a range of interesting issues.




Fig. 1: Sled acceleration before and after reinforcement of the steel cradle
Since the first preliminary test 98CF00 showed different acceleration characteristics compared with the characteristics before the modification (fig. 1), a second preliminary test 98CF12 was necessary in which the pressure and chamber volume of the Bendix system were changed. This test produced the same acceleration pattern as 98CF00. The changed acceleration characteristics are due to the changed vibration behavior of the now stiffer steel cradle. Modifying the acceleration characteristics would have required the production of a new metering pin, which would have exceeded the planned time frame for the tests. Since the deviation was not dramatic, it was decided to establish a basis for comparison between the previous and the new tests by repeating one baseline test and one HANS test.
Comparison between the repeat baseline test 98CF13 and test 97CF15 shows a substantial increase in pelvis, chest and neck values (table 1). This is surprising since the peak sled acceleration is 20 % lower.

The higher loads are due to the fact that with the new acceleration characteristics there is initially less relative speed developed between the dummy and chassis and thus less distance is covered. The dummy now takes 6 ms longer to attain peak chest and pelvis acceleration. The higher sled acceleration shortly before these peaks means that an approximately 10 % greater change in the speed of the sled and thus in the speed of the dummy has taken place by the time the peaks are reached, resulting in higher loads. The observed increased load however is not an inherent characteristic of late-peak sled accelerations. The critical factor is rather the precise behavior of the individual accelerations over time.




Fig. 2: Visor deformation Fig. 3: Helmet twisting and damage

In test 98CF00, the visor locking system of a Bell Vortex E.C. helmet supplied at short notice by Ian Brown of the FIA was studied. While the previously used Arai GP 3 helmets have a locking system consisting of a hemispherical nipple 3 mm in diameter, which engages in a groove in the visor, the Bell helmet has a cylindrical nipple 11 mm in diameter. It was found that an impact took place between the chin area and the steering wheel boss in which two attachment bolts of the steering wheel produced deep scratches in the helmet. Due to the downward rotation of the head, the two bolts caught on the lower edge of the visor. Although large opening forces were thereby exerted on the visor, the visor was only deformed, not opened. This deformation is shown in Fig. 2. The only effect of the forces was to produce extreme rearward twisting of the helmet on the head. Fig. 3 shows the rebound phase, in which the twisting of the helmet and the scratch traces produced by the bolts are visible.
Test Limit 97CF15 98CF13 98CF00 97CF18 98CF12 98CF17
HIC36 1000 1681 1400 1527 1570 1228 1205
Head acceln. 80 G 120 G 116 G 121 G 118 G 88 G 117 G
Head movement 408 mm 430 mm ? 379 mm 380 mm 385 mm
Rebound speed 27.8 km/h 25.0 km/h ? 26.5 km/h 22.7 km/h 22.2 km/h
Chest acceln. 60 G 54 G 63 G 67 G 54 G 61 G 66 G
Pelvis acceln. 60 G 60 G 71 G 75 G 59 G 68 G 74 G
Neck shear 3.1 kN 2.18 kN 2.37 kN 1.97 kN 2.32 kN 2.25 kN 2.15 kN
Neck tension 3.3 kN 3.65 kN 4.09 kN 3.47 kN 3.73 kN 3.66 kN 3.96 kN
Res. neck load 3.3 kN 4.14 kN 4.73 kN 3.99 kN 4.34 kN 4.26 kN 4.48 kN
Chest deflection 51 mm 37 mm 39 mm 38 mm 37 mm 34 mm 37 mm
Note old new old belt y-belts 2 belts belt slack
crash pulse crash pulse horizontal

Table 1: Baseline test results


At the presentation of the last status report in the presence of FIA representatives Prof. Sid Watkins and Peter Wright, the latter pointed out that Mika Salo has tested and approved a double shoulder belt system in which one belt runs horizontally from the shoulder and one runs downwards. This type of belt is referred to henceforth as a y-belt. Since HANS tests showed that the interaction of the shoulder belts with the belt-bearing surfaces behind the shoulder has a substantial influence on the performance of HANS, it was clear that HANS must be tested with shoulder belts that run downwards or with a y-belt. The latter was chosen, since this was regarded as the more critical case. In the interests of comparability, a baseline test 97CF18 with y-belt was carried out. The results are shown in Table 1. Head, chest and pelvis values remained the same, neck forces were slightly higher and forward movement of the head was significantly lower. It was therefore decided that it would be interesting to study double horizontal shoulder belts. This was done in preliminary test 98CF12, which due to the changed acceleration pattern must be compared with the repeat baseline test 98CF13. All loads were reduced by the double belt, the most marked improvements being in forward movement of the head and rebound speed. The improvement in head and neck values is chiefly due to the changed impact of the head on the steering wheel.

There are two reasons for the improvements: firstly the greater stiffness in the shoulder belt area and secondly the greater pre-tensioning, since double belt adjusters had been produced. A normal adjuster was taken, the mounting for the sewn-in belt webbing was sawn off and this was then welded to a second adjuster. The tongue of the belt buckle was attached to this adjuster via a short length of belt webbing. The two belts running to the shoulder were independently adjustable. Tightening the first belt produces the normal pre-tensioning, which is increased by tightening the second belt. The tensioning effect is not doubled, since tightening the second belt relieves some strain on the first belt.

In order to clarify which of the two reasons is the primary one, the chest and pelvis accelerations in tests 98CF13 (baseline test) and 98CF12 (double belt) were plotted in figures 4 and 5. Test 98CF00 was also plotted for comparison purposes. Since here the aim was only to test the Bell helmet, the belt from test 97CF06 had been used again. The difference between new and used material lies in the greater stiffness of the used belt since after being subjected to load the original stretch properties are not fully restored. This comparison test differs from the baseline test only as regards the belt stiffness. It yielded higher chest and pelvis values, with similar head and substantially reduced neck values. Again, the modified interaction with the steering wheel was responsible for this, since studies with stiffer belt material in an environment without steering wheel published by John Melvin in SAE 942482 showed a significant improvement in head load without any significant effect on neck load.

Before the comparison was carried out, it was first necessary to precisely synchronize the timing of the tests since the t0 signal of the sled can vary by about 1 ms. A calculation carried out later shows that this level of accuracy is not adequate. To synchronize the time factor, the time lag between the three sled accelerations was ascertained and the chest and pelvis accelerations were adjusted accordingly. The first 20 ms are then almost perfectly congruent, as is to be expected since at this stage the belt forces are still small and so do not influence movement. The only differences are attributable to the positional tolerances of the dummies. After 20 ms, the two accelerations with double belt start to exceed the other two cases. Compared with the baseline test they peak earlier and at a lower level, since less relative speed has to be dissipated. The accelerations with used belt start to exceed the baseline test acceleration only after about 43 ms, shortly after the crotch belt forces also start to be exerted. These accelerations too peak earlier but at a higher level, because the relative speed is almost fully developed after 43 ms and is more quickly dissipated by the greater stiffness. Thus stiffer belt material cannot produce improvement in the load values because it becomes effective at too late a stage. The earlier increase in acceleration with double belt is attributable solely to the greater pre-tensioning. This is therefore the dominant effect and merits further investigation. There are two reasons why this is an attractive line of investigation: firstly the belt pre-tension forces in the tests are below 100 N and even at this low level, alterations result in significant load changes. Secondly, it can be assumed that forces such as occur in braking are tolerable. Assuming that the shoulder belts restrain approximately half the torso and the head in decelerations of 4 G, this gives a force of 1 kN, or 500 N per belt, for a weight of 25 kg. This is almost one order of magnitude higher than the pre-tension in our tests. Thus the tolerability of the forces would not be a limiting factor.

In order to obtain an idea of the order of magnitude of the improvement attained by increasing the pre-tensioning of the belt, a simple calculation will be carried out in this chapter to approximately determine the relationship between belt tensioning, timeframe and resultant peak accelerations. Assuming that the belt is tensioned to make it 1 cm shorter than in a previous test, the dummy will travel approximately 1 cm less far and at any point in time its acceleration will match the acceleration it attained in previous tests 1 cm later. If for the sake of simplicity it is assumed that the acceleration pattern over

Fig. 4: Chest acceleration in baseline test and also with used belts and double shoulder belt




Fig. 5: Pelvis acceleration in baseline test and also with used belts and double shoulder belt
time also remains the same and the only difference is the time lag, this time lag can be calculated using the mean relative speed. Taking the mean relative speed as half the maximum speed, which in our tests was just over 10 m/s, this gives 5 m/s or 0.5 cm/ms. 1 cm will therefore be equivalent to a shortening of 2 ms. If there is a time lag of 2 ms between the accelerations and it is assumed that at the time of maximum relative speed half the peak acceleration of around 60 G is attained, the area between the curves integrates to a relative speed difference of 300 m/s2 * 2ms or 0.6 m/s. This is about 6 % of the relative speed. If there is 6 % less relative speed to be dissipated within a comparable space of time, the result is a reduction in accelerations of 6 %, that is to say in the case of a 60 G acceleration around 3.6 G. Comparing the double belt test and the baseline test, the time lag is about 2 ms and the difference in acceleration 1.9 G at the chest and 3.3 G at the pelvis. Thus the calculation supplies the right order of magnitude.

During the test period it was necessary at short notice to test a modified camera mounting. Since the usefulness of the videos in this test was uncertain, the test performed was one where failure would not have resulted in the loss of valuable or scarce material. Test 98CF17 was a baseline test with belt slack. The belts were set so that it was possible to pass a hand easily between the shoulder belt and the dummy. Surprisingly, chest and pelvis acceleration increased only slightly and HIC, forward movement of the head, rebound speed and neck load actually decreased. After 20 ms, the deceleration of pelvis and chest is slightly reduced (figs. 6 and 7), which results in a higher relative speed being developed and thus in a higher peak acceleration. The higher relative speed leads to greater pelvic movement, which is also visible in the video. Via the crotch belts the tension on the shoulder belts is increased and the chest, and thus also the head, travel a shorter distance. The interaction between head and steering wheel is modified, reducing the HIC and neck load. As John Melvin shows in SAE 942482, this need not be the case. These results raise the question whether the belt-tensioning effect of the pelvis when belt slack is present can be induced without increasing the pelvis and chest load. This does not appear to be possible since the belt-tensioning effect presupposes corresponding pelvic loads and thus decelerations.


5 Airbag tests, frontal impact

Following the positive results of the HANS tests at the end of the last report period, which were further improved in tests in December 97 – more on this later - the fundamental question as to the actual need for airbag development was raised, independently of FIA and DB. This necessitated departing from the original development plan for the airbag and first of all demonstrating its inherent protective potential, comparing it with HANS and then making an objectively founded decision about further development of the airbag.

In tests 98CF01 – 98CF04, the size of the vents in Knoll-design airbags was varied. To keep the study as simple as possible it was decided, as in the first tests, not to extract the airbag from the cockpit rim. The airbags were set up cross-wise in front of the driver

Fig 6: Chest acceleration in baseline test and with belt slack

clyde

Part 2

Fig. 7: Pelvis acceleration in baseline test and with belt slack
before the test. They had a volume of 20 l and were inflated by a Temic SHI18 gas generator 18 ms after t0. The results are shown in Table 2.

Comparison with the results of the first airbag tests, e.g. 97CF03, reveals that reducing the vent size results in a dramatic improvement in neck loads and forward movement of the head, with generally a slight improvement in head values. Chest deflection too is somewhat reduced. Thus it is demonstrated that with the airbag excellent neck values can be achieved and a large distance preserved between the head and the steering wheel, thus providing a safety margin for more serious accidents – without a trade-off in terms of higher head loads. Thus the Knoll airbag is a protective system which involves no compromises in this area.

Chest and pelvis deceleration are relatively scattered in these tests and show no corresponding dependence on the change in vent size. The fact that the airbags are hand-sewn undoubtedly plays a part in this, although the behavior is not untypical of standard-produced airbags either. The overall tendency is towards neutrality, or a worsening of chest values and an improvement in pelvis values.



Test Limit 98CF13 97CF03 98CF01 98CF02 98CF04 98CF03
Vent Baseline 52 mm 2*25 mm 25 mm 14 mm 0 mm
2124 mm2 982 mm2 491 mm2 154 mm2 0 mm2
HIC36 1000 1400 510 565 472 447 496
Head acceln. 80 G 116 G 64 68 G 63 G 61 G 60 G
Head movement 430 mm 367 mm 319 mm >250 mm <199 mm >170 mm
Rebound speed 25.0 km/h 21.6 km/h 19.0 km/h ? 21.0 km/h ?
Chest acceln. 60 G 63 G 52 62 G 71 G 62 G 66 G
Pelvis acceln. 60 G 71 G 58 64 G 68 G 60 G 65 G
Neck shear 3.1 kN 2.64 kN 1.42 kN 0.70 kN 0.41 kN 0.23 kN 0.31 kN
-3.1 kN -0.21 kN -0.11 kN -0.18 kN -0.26 kN -0.98 kN -0.12 kN
Neck tension 3.1 kN 4.29 kN 2.93 kN 2.23 kN 1.88 kN 1.66 kN 1.69 kN
-4.0 kN -1.32 kN -0.65 kN -1.01 kN -0.15 kN -2.04 kN -1.06 kN
Res. neck load 3.3 kN 4.73 kN 2.94 kN 2.16 kN 1.76 kN 1.84 kN 1.57 kN
Chest deflection 51 mm 39 mm 30 mm 30 mm 30 mm 26 mm 30 mm

Note: In tests 98CF02 – 98CF04 a 50 mm-thick foam plastic plate was mounted on the steering wheel to simulate the effect of a larger airbag. In 97CF03 the old crash pulse was used. For the neck force components, both positive and negative values are taken into account. The peak values therefore had to be used instead of the 3 ms values. The resultant neck load is the 3 ms value.

Table 2: Results of airbag tests to optimize vent size
Although in test 98CF04 best-value head loads and near-to-best chest, pelvis and neck loads were achieved, the unsystematic behavior described above meant that more detailed evaluation of the tests was necessary in order to safeguard against chance factors, to understand the causal mechanisms and in order to be able to further optimize the airbag.

In the first airbag test in this series, test 98CF01 (2x25 mm vents), it was found that the space between the steering wheel and head was not completely filled. In the following tests this space was reduced by fitting a 50 mm-thick foam plastic plate to the steering wheel and thus simulating larger airbags.

With standard airbags the correct positioning of the airbag is ensured by the fact that when fully inflated it extends as far as the forward-displaced occupant. When the 50 mm foam plastic plate is fitted, the chin part of the helmet has at this stage already been thrust about 30 mm into the airbag. Since the airbag is attached on 2 sides, correct positioning is ensured by the tensile forces which develop during inflation. The airbag must not be much larger, however, otherwise these forces are no longer sufficient or result in neck stresses.

Test 98CF01 behaves inherently somewhat differently from the other three tests because of the lack of the foam plastic plate. There is a slight increase in the supporting effect at the chest even while the airbag is still deploying, which is maintained until shortly before the peak value is attained (fig. 8). Due to this early supporting effect the peak chest acceleration is slightly reduced. The pelvis values are improved since the airbag exerts a downward pressure on the torso. This reduces the z-component of the chest acceleration. The pelvis is prevented from riding up the seat wedge and the deceleration of the pelvis increases at an earlier stage than in the baseline test. Less relative speed is developed.

An alternative explanation to be considered for the pelvis values is that due to the restraining of the torso, less tension is exerted on the crotch straps via the shoulder belts and the pelvis can thus travel further. This would mean however that the pelvis accelerations in the initial phase would have to be below the baseline values, which conflicts with the observations. This explanation is thus incorrect.

Before analyzing the remaining three tests, two observations must be noted. First of all, the video view from above shows that in tests 98CF02 and 98CF03 heavy smoke emission in the area of the air intake opening is already occurring during the inflation phase, thus the airbags are less fully inflated and also have an effectively larger vent. There is a discrepancy between this observation and the fact that in test 98CF03 a smaller forward movement of the head was noted than in test 98CF04. This contradiction can be explained by the second observation, namely that whereas in tests 98CF02 and 98CF03 the helmet turns to the right as seen from above, in test 98CF04 it turns to the left. The camera used for video analysis was positioned on the right. Due to the turning of the head, the measuring mark is moved backwards in the first two tests and the forward movement of the head is underestimated. The reverse is the case in test 98CF04. The signs ">" and "<" have therefore been added in Table 2. The smallest head movement is indeed that in test 98CF04. If tests 98CF03 and 98CF04 in Table 2

Fig. 8: Chest acceleration using airbags with different vent sizes






Fig. 9: Pelvis acceleration using airbags with different vent sizes
are swapped round, the three tests show a corresponding trend in the chest and pelvis values: as the supporting effect of the airbag, already acting on the chest during inflation, increases so the chest values improve, although they remain above those in the baseline test because at the moment of peak acceleration the airbag exerts an additional supporting effect on the chest due to the large forward displacement of the occupant. The pelvis values improve in line with the chest values due to the downward pressure on the torso already mentioned and the associated improvement in the interaction between pelvis and seat wedge.

As far as the chest and pelvis values are concerned, the results of test 98CF01 come very close to those of test 98CF04, because little gas is lost through the seams. The interaction between the pelvis and the seat wedge is less effective and the peak acceleration is somewhat higher because there is no foam plastic plate on the steering wheel and the airbag is thus effectively smaller and exerts less downward pressure on the torso.

The airbag used in test 98CF04 comes very close to the optimum. In a definitive design the aim would be to have a slightly larger vent and better seams: since the neck shear is already negative, the restraining effect in 98CF04 is too great, so that the head does not keep pace with the torso. Furthermore, in the rebound phase the stiff airbag thrusts the head against the head restraint in such a way as to produce a high level of neck compression. This compression occurs during this phase in all the baseline and airbag tests, but in test 98CF04 it is very high and even higher than the neck tension in the forward movement phase. A further advantage of a larger vent would be to damp the oscillation which is superimposed on the chest acceleration and would result in a significantly higher value if 3 ms values were not used.

Fig. 10 shows the resultant head acceleration in test 98CF04. During inflation, the head acceleration already exceeds the value for the baseline test, reaching 30 G after 30 ms when the airbag is fully deployed. At this point in time only the chin area of the helmet, that is to say only a small area, is supported. After 40 ms the pressure in the airbag starts to fall and head acceleration falls to 10 G. Then, after 50 ms, the entire helmet plows into the airbag and the acceleration increases due to the large supported helmet area to 50 – 60 G. After 125 ms, we see the effect of the helmet striking the head restraint. With the airbag this takes place 35 ms earlier and, at 50 G, substantially less forcefully than in the baseline test.

A further characteristic of the Knoll airbag is that the supporting effect takes place chiefly at the chin. Because the airbag has to be supported by the steering wheel, its center axis is level with the steering wheel boss. The driver has to be able to see over the steering wheel, which means that his chin is inevitably on a level with the airbag axis. The effect of this support being applied to the chin is shown in Fig. 11: the chin area of the helmet is supported so powerfully that the helmet lifts up by about 3 cm at the back of the head. On the basis of a head mass of 5 kg and an acceleration in the first phase of 30 G the force on the chin can be estimated at 1.5 kN. In SAE 680785, 1.56 – 1.78 kN is cited as the range in which fractures of the lower jaw can occur. The

Fig. 10: Head acceleration in airbag test 98CF04





Fig. 11: Maximum forward displacement in airbag test 98CF04
tests were carried out with a flat tool 2.9 cm in diameter, padded with a 0.5 cm-thick metal net. Since the chin area of the helmet has approximately 1 cm of foam plastic padding, the chin bone would probably withstand slightly more than the stated values, but the estimated value is so close to the cited range that the risk of fractures to the jaw cannot be ruled out.

In test 98CF15 an airbag with 14 mm vent was triggered only after 30 ms, i.e. 12 ms later than in the previous tests. The aim was to establish how critical the timing of airbag triggering is. The head values are somewhat higher (Table 3) than in comparison test 98CF04, because the head is already in the airbag during inflation and has developed more relative speed and forward displacement. However, the values are still dramatically better than in the baseline test. There is no acceleration peak during inflation. Thus the airbag is non-aggressive, which is attributed to the fact that it deploys in transverse direction to the direction of travel. The chest values are somewhat higher than in the comparative test and the baseline test. The airbag supports the chest briefly during the inflation phase but towards the end of the inflation phase moves upwards somewhat and the chest acceleration stagnates after 36 ms. At the end of the inflation phase, the chest acceleration returns to the value in the baseline test. Since like in test 98CF03 the airbag loses gas the chest acceleration increases, as in that test, above the baseline test value, with a practically identical peak value, due to the supporting effect


Test Limit 98CF13 98CF04 98CF15 98CF18
Vent Basis 14 mm 14 mm 14 mm
154 mm2 154 mm2 154 mm2
HIC36 1000 1400 447 685 721
Head acceln. 80 G 116 G 61 G 76 G 77 G
Head movement 430 mm 199 mm 190 mm 354 mm
Rebound speed 25.0 km/h 21.0 km/h 21.5 km/h 22.0 km/h
Chest deflection 60 G 63 G 62 G 67 G 60 G
Pelvis acceln. 60 G 71 G 60 G 67 G 63 G
Neck shear 3.1 kN 2.64 kN 0.23 kN 0.56 kN 0.64 kN
-3.1 kN -0.21 kN -0.98 kN -0.48 kN -0.17 kN
Neck tension 3.1 kN 4.29 kN 1.66 kN 2.52 kN 2.05 kN
-4.0 kN -1.32 kN -2.04 kN -2.46 kN -1.10 kN
Res. neck load 3.3 kN 4.73 kN 1.84 kN 2.41 kN 2.03 kN
Chest deflection 51 mm 39 mm 26 mm 31 mm 31 mm

Note: For the neck force components, both positive and negative values are taken into account. The peak values therefore had to be used instead of the 3 ms values. The resultant neck load is the 3 ms value.
98CF15: Airbag was triggered at t = 30 ms
98CF18: Overall function of Knoll airbag, no foam plastic plate on steering wheel, tensioners triggered at 12 ms, airbag at 18 ms

Table 3: Results of further airbag tests


of the airbag at the end of the forward displacement. This also applies to the chest deflection and the pelvis values. Neck shear remains low and positive because the supporting effect on the head occurs only when the torso has already moved sufficiently far forward. The neck tension during forward displacement and the neck compression forces during rebound also rise, but remain below the limits.

In test 98CF18 the overall functioning of the Knoll airbag was demonstrated. Fig. 12 shows three phases in its deployment. So as to ensure correct positioning of the airbag, no foam plastic plate on the steering wheel was used. As a result the airbag can only generate head acceleration at a later stage and the values are as in the previously described test. The forward movement of the head attains very high values which are even somewhat higher than in test 98CF01 (without foam plastic plate). This is because due to the extraction the airbag is positioned relatively high and due to the late supporting of the head a downward movement has already started which the airbag is less effective at restraining. The head only just misses the steering wheel. The absence of foam plastic plate has less effect as far as the chest is concerned since the chest is closer to the airbag and the load on it is exerted more in a downward direction. Thus chest and pelvis accelerations remain virtually unchanged as against comparison test 98CF04. Chest deflection is somewhat higher because the head receives less support from the airbag and more of the force required to decelerate the head has to be transmitted via the chest to the shoulder belts. The neck tension and compression are somewhat higher than in the comparative test due to the reduced supportive effect and the shear force remains positive.


6 HANS tests, frontal impact

After tests 97CF13 and 97CF14 had demonstrated that the HANS concept can yield excellent head and neck values with only small forward movement of the head, a component made of composite material, HANS 3.0, was built by Jim Downing on the basis of HANS 2.5. Without any optimization work, the mass was thereby reduced from 2.5 kg to less than 700 g. Next, test 97CF20 was carried out to show the equivalent performance of the metal and the composite component. The results are shown in Table 4. The only differences are that the chest acceleration and the chest deflection are slightly reduced with the new HANS. The reason for this is the slightly reduced area of the belt-bearing surfaces behind the shoulders with the new HANS. This means that the HANS rotates more before tensioning the shoulder belts and the head and chest are restrained somewhat less effectively, while the forward movement of head and chest is increased and the HIC is reduced. The greater forward movement reduces the chest acceleration and the chest deflection. The chest acceleration slightly exceeds the value achieved in the baseline test because of the late tensioning of the shoulder belts by HANS.

clyde

Part 3

Fig. 12: Three phases of Knoll airbag inflation: basic condition, extraction from cockpit rim, inflation
Test Limit 97CF14 97CF20 97CF24 97CF21 97CF22 97CF23
HANS version 2.5 3.0 3.1 3.1 3.1 3.1
Neck 0 mm 0 mm 0 mm +40 mm +70 mm +70 mm
Shoulder belt single single single single single y-belt
HIC36 1000 613 572 532 417 500 485
Head acceln. 80 G 73 G 75 G 74 G 58 G 64 G 64 G
Head movement 250 mm 268 mm 242 mm 263 mm 267 mm 257 mm
Rebound speed 18.5 km/h 19.1 km/h 19.5 km/h 19.4 km/h 19.0 km/h 17.5 km/h
Chest acceln. 60 G 64 G 58 G 59 G 57 G 58 G 60 G
Pelvis acceln. 60 G 60 G 60 G 63 G 59 G 60 G 61 G
Neck shear 3.1 kN 0.63 kN 0.55 kN 0.46 kN 0.37 kN 0.40 kN 0.57 kN
-3.1 kN -0.14 kN -0.12 kN -0.20 kN -0.21 kN -0.27 kN -0.20 kN
Neck tension 3.1 kN 1.40 kN 1.79 kN 1.15 kN 1.15 kN 1.56 kN 1.56 kN
-3.1 kN -1.35 kN -1.25kN -0.51 kN -0.91 kN -1.22 kN -1.46 kN
Res. neck load 3.3 kN 1.33 kN 1.65 kN 1.12 kN 0.95 kN 1.21 kN 1.45 kN
Chest deflection 51 mm 20 mm 15 mm 15 mm 14 mm 16 mm 12 mm

Note: For the neck force components, both positive and negative values are taken into account. The peak values therefore had to be used instead of the 3 ms values. The resultant neck load is the 3 ms value.

Table 4: Sensitivity of HANS to variation in neck length and shoulder belts


Next the question arose as to whether there are circumstances in which HANS may fail or may deliver unsatisfactory results. Given the experience as regards HANS’s sensitivity to modification of the tether attachment points it was clear that the height difference between head and shoulder could be a critical parameter. The head determines the height of the helmet and the shoulder determines the height of the HANS. This height difference determines the angle at which the tethers pull the HANS into the shoulder belts and thus determines the time at which HANS becomes effective. On the debit side, increasing this angle increases the pressure on the neck. It therefore had to be examined whether a viable compromise exists.

Bob Hubbard supplied mean values and standard deviations for shoulder and head height for seated persons of average height. Since both standard deviations were similar but the two dimensions themselves show no correlation, the standard deviation for the difference is the same, assuming the dimensions have a normal distribution. For a person of average height, the head/shoulder height difference is 285 mm. For 5 % of people, the difference is less than 255 mm. For 95 % of people it is less than 315 mm. In order to establish that these figures are more or less accurate, passing colleagues whose neck was subjectively assessed to be extremely short or long were measured at the workshop in Sindelfingen. None of them had measurements outside the stated range. It was then ascertained how this measurement varies when taking up a reclined seated position as in a Formula 1 vehicle. It was found that the difference was reduced by about 10 mm. This indicated that the target dimensions should be 245 mm, 275 mm and 305 mm. The dummy measurement was 235 mm, since the foam plastic wedges needed for the HANS tests made the shoulders higher. In order to have one test version with unmodified dummy neck, the dimension 245 mm was replaced by 235 mm. The longer necks were rendered by inserting 40 mm or 70 mm spacing cylinders between the neck bracket and the neck. Since with the 70 mm neck extension the tethers would have been very steeply angled, the attachment points of the tethers to HANS 3.1 were moved upwards by 40 mm, thus creating the same conditions for mean neck length as with the HANS 2.5 prototype.

It can be seen from Table 4 that the variation in neck length has virtually no effect on chest and pelvis acceleration, chest deflection and neck shear. The neck compression occurs in the forward movement phase and increases as the length of the neck increases because the helmet tethers are more steeply angled. The neck tension occurs in the rebound phase and is caused by two factors. In test 97CF20 the HANS is pulled very close under the lower rim of the helmet due to the low attachment point of the tethers on HANS 3.0. As soon as the HANS collar touches the head restraint during the rebound phase its height becomes fixed by the friction between it and the head restraint. The lower rim of the helmet then slides up onto the inner part of the HANS yoke, which is wedge-shaped towards the rear. It is pulled upwards and the neck tension steadily increases. In tests 97CF21 and 97CF24 this effect is less marked because in test 97CF21 the 40 mm longer neck increases the distance between the HANS and the rim of the helmet. In test 97CF24 the neck tension is reduced because of the higher attachment point of the tethers on HANS 3.1, which reduces the extent to which the HANS is pulled upwards. In the two tests 97CF22 and 97CF23 with 70 mm neck extension the second effect occurs: during rebound the helmet strikes the upper edge of the HANS 3.1 collar. Since this point is situated somewhat below the center of gravity of the head, the lever effects produce a brief peak in neck tension. At 1.1 kN, the 3 ms value of this peak is only slightly above the 0.94 kN of test 97CF21 and is on a level with the 1.12 kN of test 97CF24. Nonetheless this peak needs to be prevented by means of appropriate design.

Since the tether length was kept constant, the distance between HANS and helmet is reduced when the tethers are more steeply angled. Fig. 13 shows that the greater the length of the neck, the earlier the first small head acceleration peak occurs. This peak is caused by the fact that HANS initially remains stationary due to friction between HANS and the belts and only begins to move when the tethers are tensioned. The later the peak the higher its value, due to the higher relative speed of the head.

The subsequent pattern is influenced by three factors: first of all, significant forces can be developed in the tethers only if the shoulder belts are also tensioned. Then downward forces can be developed at the belt-bearing surfaces behind the shoulders, resulting in a balancing moment to the forces in the tethers. Secondly, the distance between the rear edge of the belt-bearing surfaces behind the shoulders and the anchor points of the shoulder belts when the above-mentioned forces occur also plays a role. The further forward this edge has moved, the further the shoulder belt has to be pulled

Fig. 13: Head acceleration with HANS 3.1 for varying neck lengths


upwards in order to develop a downward force. Thirdly, the extent to which this second effect occurs will depend on how steeply angled the tethers are.

The delay before HANS starts to move is longest with the unextended neck (97CF24). The distance between the belt-bearing surfaces behind the shoulders and the belt anchor point when HANS starts to act is smallest in this case. At the same time high shoulder belt forces have already been developed and the head deceleration is higher than in the other cases. Of the three neck lengths, the highest HIC for HANS 3.1 is with the short neck.

With the mean neck (97CF21), there is already a certain distance between the belt-bearing surfaces behind the shoulders and the belt anchor point, and thus the head deceleration is less abrupt.

HANS begins moving earliest with the long neck (97CF22). The distance between the belt-bearing surfaces and the belt anchor is larger here than in the other cases which means that the HANS has to be pulled further up before forces can be developed in it. This results in later head deceleration. The head deceleration is slightly higher than for the mean neck length, due to the relative speed that has been developed by this point. The values are smaller than those for the short neck since the elasticity of the shoulder belts and their effective length result in somewhat gentler deceleration of the head.

A surprising finding which emerged from these three tests is that short tethers are not a prerequisite for ensuring minimum load values. Thus it is not necessary to restrict the driver’s freedom of movement to a minimum. A second finding is that the distance between the rear edge of the belt-bearing surfaces and the belt anchor point does not need to be kept as small as possible. As a result, this edge can be moved forward, as is necessary for adaptation to a McLaren MP4/13, without causing a worsening of the load values.

HANS 3.0 (97CF20) should theoretically behave similarly to HANS 3.1 with mean neck length. Looking at the head acceleration (no diagram attached) however, it can be seen that the time when HANS 3.0 starts moving is very close to that for HANS 3.1 with short neck. Due to the curvature in the lower HANS collar area the tether length is effectively larger than measured and there is a long delay before HANS 3.0 starts moving. Since the tethers are more steeply angled than for HANS 3.1 with short neck, the head deceleration is even more abrupt and the HIC value is the highest for these four tests.

The fact that the drivers wear the belts running down from the shoulder was seen as a further critical point. With HANS this influences the interaction between the belt-bearing surfaces behind the shoulders and the belt. Since Professor Sid Watkins had reported that Mika Salo wears a split shoulder belt system with one belt running downwards from the shoulder and one running horizontally, it was considered what the most unfavorable case might be. This, it was decided, was the combination 70 mm neck with a double belt where one of the belts runs downwards at an angle of 45°. Test 97CF23 was carried out under these conditions. Compared with the normal belt (97CF22) there is somewhat greater neck compression because HANS is held down better. This also means HANS begins to be effective earlier and HIC and forward displacement of the head are reduced. The increased friction on the belt-bearing surfaces results in greater energy absorption and a reduction in rebound speed.

In test 98CF20 the effect of belt slack on HANS was investigated. As in test 98CF17, the belt slack was such that a hand could be passed between the shoulder belt and the dummy. In Table 5, this test can be compared with the HANS test without belt slack and with the baseline tests using the new crash pulse with and without belt slack. The only major differences are a significantly higher HIC, which is due to the later and more forceful deceleration, and an increase in positive neck shear and neck compression, both of which however are still absolutely non-critical. Both types of neck force are increased by the later, more abrupt deceleration of the head. The neck compression is caused by the downward angle of the tethers. A greater restraining force leads to greater compression. The shear is due to the greater relative head/torso movement brought about by greater resultant forces in conjunction with the elasticities in HANS and the belt system.

In test 98CF21 the amount of belt slack was set by gently pulling the belt through the adjuster until it lay loosely against the dummy. The neck length was increased by 40 mm because this is the mean neck length and because the worsening in the HIC was attributed to the less steeply angled tethers and the associated less effective interaction

Test Limit 98CF13 98CF17 98CF14 98CF20 98CF21
HANS version baseline baseline 3.1 3.1 3.1
Note belt slack belt slack reduced belt slack
Neck 0 mm 0 mm 0 mm 0 mm 40 mm
HIC36 1000 1400 1205 555 722 336
Head acceln. 80 G 116 G 117 G 66 G 82 G 57 G
Head movement 430 mm 385 mm 260 mm 267 mm -
Rebound speed 25.0 km/h 22.2 km/h 16.5 km/h 16.2 km/h -
Chest acceln. 60 G 63 G 66 G 62 G 62 G 59 G
Pelvis acceln. 60 G 71 G 74 G 69 G 72 G 67 G
Neck shear 3.1 kN 2.64 kN 2.43 kN 0.44 kN 0.78 kN 0.97 kN
-3.1 kN -0.21 kN -0.09 kN -0.19 kN -0.21 kN -0.21 kN
Neck tension 3.1 kN 4.29 kN 4.86 kN 1.59 kN 1.55 kN 0.18 kN
-3.1 kN -1.32 kN -0.94 kN -0.52 kN -0.91 kN -0.18 kN
Res. neck load 3.3 kN 4.73 kN 4.48 kN 1.35 kN 1.33 kN 0.96 kN
Chest deflection 51 mm 39 mm 37 mm 16 mm 18 mm 21 mm

Note: In test 98CF21 a Bieffe helmet was used. The tethers broke away from the slot.
For the neck force components, both positive and negative values are taken into account. The peak values therefore had to be used instead of the 3 ms values. The resultant neck load is the 3 ms value.

Table 5: Sensitivity of HANS to belt slack


between the belt-bearing surfaces behind the shoulders and the shoulder belts. The helmet model used was 'Bieffe Formula 1 – GP Compotech. K/C‘. In this test, the slots in the helmet which hold the tethers were damaged. The rear edge broke away shortly before maximum forward movement was attained and the tethers were pulled out. This resulted in the lowest HIC recorded so far, while neck shear was only slightly higher. Due to the large forward displacement of the head however, a glancing impact with the steering wheel occurred. Since no material was available at short notice, it has not yet been possible to repeat the test.

This test shows that the HANS/helmet connection is not entirely unproblematic. HANS subjects the helmets to stresses for which they were not designed. Collaboration with the helmet manufacturers must be sought. At the meeting of 15.5.1998, Ian Brown of the FIA undertook to pursue this. It should be noted however that there are helmets which are able to cope with these stresses. An Arai GP3 helmet came through 10 frontal (97CF20-24, 98CF14, 98CF16, 98CF20, 98CF25-26), 4 angled (98CF10, 98CF22-24), one side (98CF11) and one rear-end impact (98CF09) without noticeable damage.

These tests show that the protection systems airbag and HANS provide largely equivalent protection. So that the decision between the systems did not have to be decided just by the expenditure involved, further improvement possibilities were explored. In a discussion with Professor Steffan of TU Graz in which the results were submitted to him for a neutral appraisal, he suggested attaching HANS to the shoulder belt with a friction element so as to absorb energy and develop acceleration at an early stage. It has already been established in studies with belt force limiters that it is very difficult to generate a defined force using friction. In the aviation industry a type of force limiter exists which is based on the tearing of seams. Schroth Safety Products produced force limiters of this type for us at short notice designed for a force of twice 1 kN. This design was chosen in order to rule out the possible occurrence of large neck shear forces. It results in a head deceleration of around 30 G. The length of the seam, at 10 cm, was chosen so that the seam would have opened completely by the time that HANS is able to take control by itself of restraining the head. For testing of the basic principle, the force limiter was connected firmly to HANS and the shoulder belt. In practice of course the force limiter would have to be connected to HANS automatically when fastening the shoulder belts and to be released again when the belts are unfastened.

Test 98CF26 (Table 6) revealed considerably greater elasticities in the torso and in HANS than expected. After tensioning of the tethers, the rear edge of the belt-bearing surfaces behind the shoulders therefore remains stationary while the upper edge of the collar moves forward. As a result of this, 5 cm of travel is lost during which no forces can


Test Limit 98CF13 98CF04 98CF14 98CF26 98CF16
HANS version baseline airbag 3.1 3.1 airbag+3.1
Note tear strip
Neck 0 mm 0 mm 0 mm 40 mm 0 mm
HIC36 1000 1400 447 555 633 488
Head acceln. 80 G 116 G 61 G 66 G 63 G 62 G
Head movement 430 mm <199 mm 260 mm 240 mm ? 117 mm
Rebound speed 25.0 km/h 21.0 km/h 16.5 km/h - 12.8 km/h
Chest acceln. 60 G 63 G 62 G 62 G 62 G 69 G
Pelvis acceln. 60 G 71 G 60 G 69 G 67 G 71 G
Neck shear 3.1 kN 2.64 kN 0.23 kN 0.44 kN 0.28 kN 0.28 kN
-3.1 kN -0.21 kN -0.98 kN -0.19 kN -0.41 kN -0.65 kN
Neck tension 3.1 kN 4.29 kN 1.66 kN 1.59 kN 0.49 kN 1.11 kN
-3.1 kN -1.32 kN -2.04 kN -0.52 kN -0.07 kN -0.51 kN
Res. neck load 3.3 kN 4.73 kN 1.84 kN 1.35 kN 0.56 kN 0.99 kN
Chest deflection 51 mm 39 mm 26 mm 16 mm 17 mm 11 mm

Note: For the neck force components, both positive and negative values are taken into account. The peak values therefore had to be used instead of the 3 ms values. The resultant neck load is the 3 ms value.

Table 6: Comparison airbag/HANS for frontal impact
yet be transmitted to the head. The tear strips are only half opened and when the head
reaches the end of its forward movement they act in addition to the normal HANS forces. This results in smaller forward movement of the head and a larger HIC.

This line of investigation was not pursued any further because even the 50 mm of tear strip now used is not feasible in practice. A certain degree of slack is required in the tear strips in order to allow the HANS to move forward with the torso during braking. This means that only at best 30 mm of tear strip can actually be used and the benefit in terms of forward displacement of the head is even smaller. Since the benefit is so small, the greater complexity does not appear justified.

In test 98CF16 the airbag and HANS were tested in combination since the airbag can support the head in the early phase and HANS can support the head in the late phase of forward movement, thus the systems complement each other. The combination of excellent data is impressive, particularly the small forward movement of the head. Only the chest deceleration is somewhat higher, because the peak acceleration lasts somewhat longer than 3 ms and thus unlike in test 98CF04 the 3 ms value is not reduced. In the videos it can be seen that the airbag in combination with HANS produces a belt-tensioning effect: the airbag supports the chin, the helmet rotates forward, the attachment points of the tethers move upwards, the now more steeply angled tethers pull the HANS upwards and the belt is tensioned. This only occurs 10 ms (equivalent to the inflation time) later than would be the case if a belt-tensioner were used – a good indication of the latter’s potential.

clyde

Part 4

7 Angled, side and rear-end impact

The next tests explored 30° angled impacts, side impacts and rear-end impacts. The results are shown in Table 7. The first point to notice with the 30° angled impact is that the head values are worse with HANS and are actually less favorable than in the case of a frontal impact. The reason is the friction between the helmet and the raised side panels which acts additionally to the HANS forces and thus increases the head acceleration. The values in the baseline test are thus unexpectedly favorable, since the head just misses the steering wheel. As a trade-off for the good head values however high neck values are recorded, as opposed to very good neck values with HANS. Also noteworthy is the higher chest acceleration with HANS, which is familiar from the frontal impact. This occurs simultaneously with the maximum head acceleration. As a result of the restraining forces acting on the head, forces are transmitted via the HANS to the shoulders and thus the torso is decelerated more sharply, with a visible reduction in the distance it travels. The transmission of forces to the shoulders also reduces the chest deflection.

A further important aspect is illustrated in Fig. 14: due to the impact on the raised side panels in the baseline test and the associated frictional force a strong rotational impetus is imparted to the head. On the basis of the video a rotational acceleration of between 9000 and 13000 rad/s2 over a period of 4 ms was estimated. This substantially exceeds
Test Limit 98CF07 98CF10 98CF08 98CF11 98CF06 98CF09
Impact 30° 30° 90° 90° 180° 180°
HANS version baseline 3.1 baseline 3.1 baseline 3.1
HIC36 1000 585 700 952 707 755 706
Head acceln. 80 G 64 G 80 G 95 G 77 G 97 G 93 G
Chest acceln. 60 G 49 G 55 G 57 G 50 G 44 G 44 G
Pelvis acceln. 60 G 61 G 60 G 53 G 55 G 45 G 50 G
Neck shear 3.1 kN 1.81 kN 0.53 kN 0.45 kN 0.41 kN 1.00 kN 0.54 kN
-3.1 kN -0.17 kN -0.16 kN -0.06 kN -0.03 kN -0.32 kN -0.29 kN
Neck tension 3.1 kN 2.87 kN 0.66 kN 1.87 kN 1.28 kN 0.31 kN 1.10 kN
-3.1 kN 1.10 kN -0.36 kN -0.19 kN -0.17 kN -2.06 kN -1.24 kN
Res. Neck load 3.3 kN 3.35 kN 0.71 kN 1.68 kN 1.28 kN 2.04 kN 1.25 kN
Chest deflection 51 mm 32 mm 8 mm - - 5 mm 4 mm

Note: For the neck force components, both positive and negative values are taken into account. The peak values therefore had to be used instead of the 3 ms values. The resultant neck load is the 3 ms value.

Table 7: 30° angled impacts, side impacts and rear impacts

the value of 6000 rad/s2 ascertained in the TRL replication of an accident which resulted in a coma for one day and a lateral skull fracture. According to TRL fatal injuries occur between 4500 and 15000 rad/s2, and there is a general consensus that 10000 rad/s2 or more will in all probability have fatal consequences. Airbag and HANS reduce the head rotation dramatically.

In the side impact the head, neck and chest loads are significantly worse in the baseline test than in the HANS test. Since HANS was supposed to behave neutrally in this impact direction a reason for this was sought. In the video from above it can be seen that in the baseline test the dummy was sitting about 2.5 cm further away from the side with which it had contact in the impact. As a result it has a higher relative speed at the moment of contact, with the above-mentioned consequences.

In the rear-end impact the most striking observations are the increased pelvic values and improved neck values obtained with HANS. The increased pelvic values in conjunction with unchanged chest values are not due to HANS but rather to the following fact: in the baseline test the steering wheel had already been used in 5 frontal tests with HANS, whereas in the HANS test a new steering wheel was used. Apparently the steering wheel used in the baseline test had incurred damage from the shocks sustained in the frontal tests because it broke in the rear-end impact, allowing the legs and thus the pelvis to travel further.

The neck forces are reduced by the HANS collar behind the head. Whereas in the baseline test the friction between the helmet and the head restraint induces a rearward rotation of the head, this is prevented with HANS because the frictional forces are transmitted via the HANS device to the shoulders (fig. 15). The frictional forces cause




Fig. 14: Head rotation in the angled impact without (top) and with HANS (below)




Fig. 15: Head rotation in rear-end impact without (top) and with HANS (below)
compression of the neck and the head rotation imparts a flexural moment to the neck
which attempts to push the torso forward. The associated force is measured in the neck as shear force.

Since HANS produced a noticeable, though non-critical increase in head values in the angled impact and with belt slack, two questions arose: what effect is achieved with HANS in the combination angled impact plus belt slack and does the airbag yield significantly better values than HANS in the angled impact?

In test 98CF19, like in test 98CF04, a 20 l airbag with 14 mm-diameter vent was used. The results are shown in Table 8. The airbag yielded good head values, the chest and pelvic values were on a par with HANS and the good neck loads were somewhat worse than with HANS, as in the frontal impact. The good head values are achieved because the head is supported at an early stage. The fact that the chest values are higher than in the baseline test is due to the additional supporting effect of the airbag which can, however - as described in the airbag chapter - show considerable scatter.

In tests 98CF22 – 98CF24, HANS was tested in an angled impact with belt slack and various neck lengths. The amount of belt slack was not however, as in test 98CF20, chosen so as to leave enough room for a hand to pass between the belt and the dummy since this was regarded as unrealistic. Instead, as in test 98CF21, the belt was gently pulled through the adjuster until it lay loosely against the dummy. Also, in the test with unextended neck a HANS 3.0 was used, so that the tethers were more steeply angled and the HANS interacted with the shoulder belts at an earlier stage. The results are shown in Table 8.


Test Limit 98CF07 98CF10 98CF19 98CF24 98CF22 98CF23
HANS version Basis 3.1 Airbag 3.0 3.1 3.1
Neck 0 mm 0 mm 0 mm 0 mm +40 mm +70 mm
HIC36 1000 585 700 537 596 546 593
Head acceln. 80 G 64 G 80 G 66 G 70 G 65 G 64 G
Chest acceln. 60 G 49 G 55 G 54 G 54 G 62 G 54 G
Pelvis acceln. 60 G 61 G 60 G 63 G 60 G 65 G 60 G
Neck shear 3.1 kN 1.81 kN 0.53 kN 0.35 kN 0.56 kN 0.61 kN 0.57 kN
-3.1 kN -0.17 kN -0.16 kN -0.33 kN -0.16 kN -0.20 kN -0.20 kN
Neck tension 3.1 kN 2.87 kN 0.66 kN 1.52 kN 1.03 kN 0.94 kN 1.37 kN
-3.1 kN 1.10 kN -0.36 kN -1.19 kN -0.27 kN -0.22 kN -0.24 kN
Res. neck load 3.3 kN 3.35 kN 0.71 kN 1.42 kN 1.13 kN 1.02 kN 1.47 kN
Chest deflection 51 mm 32 mm 8 mm 18 mm 9 mm 13 mm 13 mm

Note: For the neck force components, both positive and negative values are taken into account. The peak values therefore had to be used instead of the 3 ms values. The resultant neck load is the 3 ms value.

Table 8: 30° angled impact
The head values are now significantly better than in the first HANS angled impact and are similar to those in the baseline and airbag tests. This is not however due to the changed geometry of the tethers but to the reduced amount of belt slack. As a result, HANS provides its restraining effect only after the friction on the side panels has occurred, and the accelerations are thus smaller, although at the price of a considerable increase in forward movement of the head. This effect cannot be relied on however, since there will always be an accident severity at which the two forces occur simultaneously, resulting in a certain worsening. It should be noted that the HIC values are still 30 % below the limits of the FMVSS208 standard. With the airbag too the above effect is unavoidable, but due to the earlier restraining effect it only occurs at a higher level of accident severity than in the case of HANS, which could result in the limits being exceeded. With HANS on the other hand, there would be a risk in such cases of a head impact with hard components. Tests are the only way of providing precise information about this but given the likelihood of destruction it would be advisable to wait to carry out the testing with the existing body until all other work has been completed.

It is interesting to note, as far as neck forces are concerned, that there is no longer any compression and instead there are actually small tension forces. Here again, belt slack is responsible. The forces between shoulder belt and torso, which are the pre-condition for HANS to become effective, are developed only at a later stage, because in an angled impact the pelvis is less effective in taking up the belt slack and tensioning the shoulder belts. The video from above shows that the pelvis does not travel as far in the x-direction: there is 13 % less speed in the x-direction to be dissipated in a 30° angled impact, while the friction between the pelvis and the cockpit wall produces an additional deceleration force. The greater forward movement of the head produces centrifugal forces which are reduced with the onset of the HANS forces. This is least effective in the case of the 70 mm neck since here the distance between the rear edge of the belt-bearing surfaces and the belt anchor points is largest, so that the elasticity in the belt webbing more than compensates for the effect of the steeply angled tethers.

The last point to note is the increase in chest and pelvic load in test 98CF22. The reason is the even greater amount of belt slack in this test. This can easily occur due to the slip-stick effect since the belt webbing was pulled only gently through the adjuster. Due to the larger amount of belt slack the pelvis can travel further, while the x-component is delayed and peaks at a higher level. The z-component is higher because the pelvis is accelerated further upwards by the seat wedge. This higher z-component also feeds through to the chest.

The x-component of chest acceleration is initially somewhat below the values of the other two tests and shows a substantial increase after about 35 ms. The peak value remains unchanged however, because by this time the pelvis has already passed its peak acceleration and its rearward movement provides the chest with more belt webbing and thus with more travel. The y-component is initially substantially lower than in the comparison tests because the shoulder belts first have to be tensioned by the pelvis before a y-component can be developed. A greater relative speed is developed and the impact against the cockpit wall is correspondingly harder. The fact that peak acceleration is reached simultaneously in all three components produces a high resultant chest acceleration.


8 Airbag/HANS comparison

In this chapter the aim is to collate the supporting arguments for a decision on how to pursue development of the airbag and HANS. Such a decision can only be absolutely objective if both systems have been developed and optimized and all conceivable cases have been tested. Rather than expending this amount of effort the wish was to develop, with the available resources, as effective a protection system as possible as quickly as possible and to be able to conclude the project if possible before the next serious accident. This means that absolutely objective conclusions are not possible; however, some characteristics have been reliably established and on the basis of the collated experience the uncertainty of the results can be limited.

As far as the load values obtained from the tests are concerned, what is of interest is not how a protection system performs in a favorable case but how it performs under unfavorable conditions. The selecting of tests from which the most unfavorable values are ascertained is to some extent a subjective process. Therefore one table was produced with optimistic values and one with conservative values. The reasons for the choice of tests are set out below.

For the optimistic table, Table 9, the baseline tests 97CF15 (old crash pulse), 97CF18 (old crash pulse, y-belt), 98CF12 (new crash pulse, double belt), 98CF13 (new crash pulse) and 98CF07 (angled impact) were used. The conservative table, Table 10, also includes test 98CF17 (new crash pulse, belt slack). The tests prior to 97CF15 are not taken into account because they took place with a different dummy. 98CF00 is not taken into account because it took place with used belt material.

With respect to the airbag, the tests with 14 mm vent are used in Table 9 because this is the most favorable airbag design. Since in the frontal test this design resulted in negative neck shear, the restraining effect on the head was too great. Since the airbag becomes disproportionately more stiff as accident severity increases, this effect becomes more pronounced. The solution would be to increase the size of the vent. Table 10 therefore includes test 98CF02 with 25 mm vent. With the foam plastic plate on the steering wheel, the positioning of the airbag in front of the head is unreliable. Therefore Table 10 must take into account test 98CF18, in which the overall functioning of the airbag was demonstrated. Test 98CF15, in which triggering of the airbag took place 12 ms too late, is not included because this is a problem with the electronic triggering system and the 12 ms is a random value. The performance of the airbag could be worsened ad infinitum by further delayed triggering.

Table 9 uses the HANS 3.0 and 3.1 values for neck length 0 and the HANS 3.1 values for neck length 40 mm and 70 mm. Test 97CF23 with 70 mm neck and y-belt is also used. Since in the angled impact with belt slack there was a time lag between the head impact on the raised side panels and the onset of the HANS restraining effect, a conservative evaluation must contain the test without belt slack (98CF10). Also, test 98CF20 was taken into account, in which HANS was tested in a frontal impact with a large amount of belt slack.

Various points must first be noted before comparing the tables. Both tables are more conservative with regard to HANS than with regard to the airbag. 3 – 4 times as many tests were carried out with HANS as with the airbag. The results obtained when varying the size of the airbag vent indicate that a number of reproduction tests ought to be carried out, which would result in a worsening of some values. Also, no airbag test has yet been carried out with belt slack. This would in all probability show a need for a larger vent size.

Looking at the two tables it can first of all be seen that both systems dramatically reduce head and neck values. Whereas the head load recorded in the baseline test was on a par with the TRL replication of an accident which resulted in a coma for one day and a lateral skull fracture, both systems produce values well below the limits, even under unfavorable conditions. In the comparison between airbag and HANS the optimistic values give the airbag an advantage in terms of forward movement of the head, while HANS produces significantly smaller chest deformation. All other values are either small in relation to the limits or else are no more than 10 % apart. In the conservative evaluation, the advantage of the airbag in terms of forward movement of the head disappears; also, HANS has a 20 % advantage in terms of resultant neck load. Both aspects are due to the inclusion of test 98CF18, in which the overall functioning of the airbag was demonstrated. Thus the effects of the extracting the airbag on the overall dynamics should not be underestimated.

On the basis of the overall data compiled, no system can be said to have an overall advantage. The use of HANS and airbag in combination gives extremely good values for forward displacement of the head while other values are similar, but this is due to an indirect belt-tensioning effect exerted by the airbag in interaction with HANS. This is an effect which should be produced by means of a belt-tensioner.

Also, as was pointed out in the airbag chapter, there is a significant risk of chin injuries with the airbag because this is the main point at which forces are transmitted to the head, and the chin is the part of the head with the lowest loading tolerance. With HANS there is no such problem, since the point of force transmission to the head is defined by a line extending from the tether attachment points through the center of gravity of the head. By attaching the tethers at a low point on the helmet, it is possible to direct the forces into the forehead, where loading tolerance is highest.

A further risk with the Formula 1 airbag is that head and torso are restrained independently of each other. Since the restraining effect of the airbag increases disproportionately with increasing accident severity, there is a risk that in serious accidents the head may be restrained too severely relative to the torso, causing neck
Limit Baseline Airbag HANS
Number of tests 5 2 9
HIC36 1000 1681 537 596
Head acceln. 80 G 120 G 66 G 75 G
Head movement 430 mm <199 mm 280 mm
Rebound speed 27.8 km/h 21.0 km/h 19.5 km/h
Chest acceln. 60 G 63 G 62 G 62 G
Pelvis acceln. 60 G 71 G 63 G 69 G
Neck shear 3.1 kN 2.64 kN 0.35 kN 0.61 kN
-3.1 kN -0.29 kN -0.98 kN -0.27 kN
Neck tension 3.1 kN 4.29 kN 1.66 kN 1.79 kN
-4.0 kN -1.54 kN -2.04 kN -1.46 kN
Res. neck load 3.3 kN 4.73 kN 1.84 kN 1.65 kN
Chest deflection 51 mm 39 mm 26 mm 16 mm

Note: For the neck force components, both positive and negative values are taken into account. The peak values therefore had to be used instead of the 3 ms values. The resultant neck load is the 3 ms value.

Table 9: Most unfavorable values - optimistic






Limit Baseline Airbag HANS
Number of tests 6 4 11
HIC36 1000 1681 721 722
Head acceln. 80 G 120 G 77 G 82 G
Head movement 430 mm 354 mm 280 mm
Rebound speed 27.8 km/h 22.0 km/h 19.5 km/h
Chest acceln. 60 G 66 G 71 G 62 G
Pelvis acceln. 60 G 74 G 68 G 72 G
Neck shear 3.1 kN 2.64 kN 0.64 kN 0.78 kN
-3.1 kN -0.29 kN -0.98 kN -0.27 kN
Neck tension 3.1 kN 4.86 kN 2.05 kN 1.79 kN
-4.0 kN -1.54 kN 2.04 kN -1.46 kN
Res. neck load 3.3 kN 4.73 kN 2.03 kN 1.65 kN
Chest deflection 51 mm 39 mm 31 mm 18 mm

Note: For the neck force components, both positive and negative values are taken into account. The peak values therefore had to be used instead of the 3 ms values. The resultant neck load is the 3 ms value.

Table 10: Most unfavorable values - conservative

injuries. HANS is a restraint system which provides restraint for the head relative to the torso and thus it does not encounter this problem.

Another problem with the airbag relates to triggering. Expert opinion is that since the deceleration in a nosecone test is between 10 and 20 G, whereas driving over a curb produces 60 G and engine vibrations have an amplitude of 70 – 85 G, it is not possible to use a conventional sensor system. In the last status report a crush sensor system was suggested. This would have to be laminated into the nosecone and side impact structure. A vehicle would have to be built which would not be usable in competitions and would be tested separately from normal testing and racing activities. This would require a test team which would first have to be assembled and financed. A cost estimate based conservatively on the distance a race driver covers during a season in both racing and testing indicates that the costs for sensor testing alone would exceed the current overall project costs by a factor of between 10 and 30.

Thus an airbag has no biomechanical advantages over HANS and entails unacceptable costs. Therefore no case can be made for continuing the development of an airbag.

downshift

My head hurts. :p

ZoomZoomH

iTrader: (1)

anybody got cliff notes on that paper? :p

or should all drivers start wearing HANS instead :p

PUR NRG

Too much blah, blah, blah?

This report deals with F1 safety and biometrics. I wonder if it has real application for autocross, where the impact speeds are significantly lower and seat height/angle relative to the steering wheel are different.

And here I was getting hyped about justification for a new Arai open face helmet...
________
DC MEDICAL MARIJUANA DISPENSARIES

clyde

I think the critical part that is causing concern is in Chapter 5:
There are a few things that I think are important to note which, in my layperson's opinion, cast doubt on how applicable the conlusions are to us. The study was on the F1 car environment which is pretty standard from car to car and is signficantly different from the varied environments of our production and production based cars that we use in Solo II and the other series mentioned in the memo. The type, nature and severity of the impacts that are reasonably forseeable in our events are also quite different than those in F1 racing.

The study doesn't suggest how much less risk (or even if it's noticeably reduced at all) there might be with an open face helmet under their test conditions than a full face helmet. For all we know, we might be better off with no helmet at all in some cars with certain airbag configurations...or, when the entire spectrum of forseeable crash situations are examined full face could be the best bet.

The only thing that this study tells us, IMO, is that we would probably be best off if a study was done to accurately assess the risks that we face doing what we do.

Just my 2 or 3 cents.

Banannie

http://www.sccaforums.com/cgi-bin/ul...i/topic/7/898?

Lots of discussion on SCCA Forums, including feedback from Doug Gill (SCCA)

Clearly he says it's not mandated.

Annie (who will continue to run a full face helmet because it will do double duty in autocross and road racing, and I'm not 100% convinced of the applicability of this F1 study to the autocross environment)