Crash Helmets and Impact Test Limits
I inquired Snell about some more of the issues of duration and HIC values recently, here's the reply. Ed Becker's been incredibly gracious with his thoughts to an anonymous questioner, and answered direct critical questions within a couple days of writing. This info may provide a little more clarity to some typical ideas or questions of riders concerned with helmet effectiveness and the various performance standards, and give a little background or better understanding of the idea various issues of current debate surrounding the use of various criteria within the available performance marks.
"Usually, I don't deal directly with issues of injury. Instead, I work with injury criteria as devised by researchers and written into helmet test standards. The test head forms we use rarely complain or show any symptoms so the most we can do is compare test measurements to our injury criteria and come up with pass/fail evaluations.
At this point, there are at least four sorts of criteria used to evaluate impact test results:
Peak G - whether the peak acceleration exceeds some value (Snell and everyone else sets peak G criteria)
Duration - whether the total time the acceleration exceeds some G value exceeds some time limit(DOT and Japan)
Severity Index (SI) - based on the time history of the acceleration (NOCSAE- sports helmets)
HIC - also based on the time history of the acceleration - an "improvement" on SI. (ECE 22-05 and FIA 8860)
During the 1960's, researchers at Wayne State University published the "Wayne State Curve" which purported to show that. Injury is sensitive both to acceleration and to the time of exposure. They had taken a pneumatic hammer to cadaver skulls and, by controlling both hammer speed and stroke, showed that the skulls would fracture at higher G's at shorter strokes and at lower G's for longer. An engineer at GM, Mr. Charles Gadd, cobbled up the Severity Index: acceleration in G's raised to the 5/2 power and integrated over time. GM was looking for a reasonable way to interpret car crash tests in which instrumented dummies were pounding into vehicle interiors. The frequently did not know impact velocities or anything but the acceleration versus time recordings their equipment recorded. Gadd gave them a way to compare their data to the Wayne State curve and evaluate the effectiveness of vehicle interior padding. HIC was proposed later to correct some philosophical problems with the SI. One was that the Wayne State accelerations readings were averages and another was that standing in a room at 1 G acceleration for 1000 seconds would expose an individual to an SI of 1000 putting him at risk of skull fracture. HIC is described in Federal Motor Vehicle Safety Standard 208 which deals with vehicle interiors. The FMVSS HIC limit is 1000.
In the 60's the ANSI helmet committee wanted to include some features of SI and the Wayne State Curve in their motorcycle helmet standard but hardly anyone had the gear or the computing power to digitize accelerometer signals or do the arithmetic on a regular basis. Instead, they came up with time duration. Dr. Snively, the man behind the Snell standards, argued that the fact that the impact velocities were measured really meant that SI and time duration were largely superfluous but time duration went into the standards anyway. But the manufacturers were careful to set duration requirements that wouldn't bite. It wasn't till they revised the standard in 1970 eliminating the old swing-away test and requiring guided fall equipment that duration became a concern. Duration, it turns out is significantly longer in guided fall than in a swing-away test that exposes the helmet to the same impact energy. The manufacturers did not realize this until later and by the time they published a correction, DOT had already made their mistake law.
Interestingly, when the DOT standard was circulated for comment, the government also proposed that the test criteria would eventually be replaced by HIC. They figured this could be done in a year and a half. This never came to pass. Manufacturers and helmet experts were able to shoot this down even if they never were able to get DOT to accept ANSI's fix for the time
In helmet testing, peak G can be violated for two different reasons, one is that the liner is too dense and the acceleration eases up over the criterion before settling back down at the end of the impact. The other, though, is very different, the liner is too thin and gets collapsed completely before the pulse is over. As long as the liner still has some crushable thickness left, everything is fine but at the end of that thickness, the interior of the helmet is the impact surface and it is hard up against our head form, the measured acceleration shoots up and, an instant later falls back down to zero. Since our titanium head form and drop assembly and the helmet shell and even the steel anvil all have some miniscule shock attenuation capability, the sharp acceleration spike does not leap straight to infinity. If the helmet almost managed the impact, the peak might not go all the way off our charts. But the problem will not be fixed by lowering the liner density, instead, the manufacturer might have to increase the density, make the liner thicker or both.
Generally, flat impacts test the liner density. Flat impacts work a broader area of the liner than hemi impacts so, for the same impact velocity, G's will be higher and but there won't be as much crush as for a hemi impact. Hemi's work a small area of the liner, the helmet shell deforms around the hemi leaving a smaller foot print on the liner. With less liner working, the G's are lower but, for the same reason, the crush depth is greater. If the helmet meets flat requirements, the liner will never be too stiff for the hemi and if the helmet meets hemi requirements, the liner will always be thick enough for the flat.
Duration requirements are almost never a problem for hemi impacts. Hemi failures are sharp narrow acceleration peaks that never last more than the slimmest fraction of a millisecond. You only see duration violations for flat impacts. HIC and SI are very similar. If a helmet is going to have trouble with any of these, it's most likely to be in flat impact. Hemi results can put HIC and SI over the criteria but only well after the same trace has already broken through any reaonable peak G limit.
The traces I presented at the HIC meeting were intended to show the non-compatibilities of Snell and ECE 22-05 testing. Much of the differences there are due to mass differences in the head form specs. The medium size Snell helmets are tested on heavier head forms than ECE calls out so ECE impact results are likely to be higher, since they call out a lower peak G criterion, in addition to a HIC 2400 criterion, our test simulations of ECE requirements came pretty close to the line. Since this head form weight disparity is much greater for extra small helmet sizes, there's hardly any way to build small size helmets that will meet both Snell and ECE. On the other hand, ECE does not use the hemi anvil and I suspect that much of the energy in their kerb stone impacts is lost to rotation and misalignment effects. Guided fall allows excellent alignment between head form c.g. and the anvil center but the European free drop systems make this sort of alignment virtually impossible. As a result ECE helmets generally do poorly in Snell type hemispherical impacts.
The ECE HIC criterion of 2400 was one of the issues prompting that HIC conference. But FIA 8860 allows HICs up to 3600. The original HIC limit was set to 1000. No one is quite sure why 2400 or 3600 should be okay for helmets, but I suspect that they reverse engineered these limits from what current helmets can do.
However, there is one aspect of helmet performance that duration, HIC and SI might get at and that peak G does not: rebound. Helmets bounce so the total velocity change is not just from impact velocity down to zero but, instead all the way past zero to whatever negative veloicty is implied by the bounce. A perfectly resilient helmet would expose its wearer to twice the impact durations and, presumably twice the HIC and twice the SI as a perfectly dead helmet with no bounce at all. My own impression is that there's not that much variance in bounce across helmets although some think that plastic shells might be a little more bouncy than others. If I'm right
and bounce is pretty constant for current helmet technology then bounce may not be an issue for standards. Right now, it appears that duration, HIC and SI all work to impose some lower peak g for flat impacts than the existing peak g criteria. We could accomplish the same purpose just by setting two peak G criteria and save a lot of heartache. But the actual value of the G limit does not affect hemi pass/fail results all that much, hemi failures are usually well above 300, 400, 500 and what have you. So the question of G level is really a matter of what G level to set for flat impact. The same level will work just as well for hemi as any other impact level. We figure that for current head gear technology 300 G is a conservative upper limit for flat impact and we've found it works just fine to eliminate hemi failures as well.
I've taken a shotgun approach to this response, I'm not really sure what your questions really are. I'd be happy to attempt a few more exchanges to see if we can get closer to an understanding.
Last edited by license2ill; 12-02-2007 at 05:20 AM.