I emailed Snell a while back about some of the issues that Rick Marlin of Suomy brought-up about the performance of Snell vs BSI helmets as well as the random vs batch testing methods and any recent research that may pertain to the issues. I specifically wanted to know about whether or not research existed to point to a specific standard's testing criteria, and understand more about the various points of criticism. Overall, I was very impressed by Becker's efforts and articulation of the issues. However, I am bothered by the relative infancy of the research and data available for the basis of current efforts and future advancements. It is a shame that more research or funding is not being done on the subject. Here's our correspondence:
Is there any current or ongoing research being conducted into the causes/effects of helmeted head/brain injuries or deaths specifically caused by head trauma?
Also, does the Snell foundation seek to fund research or ask the helmet industry to support head protection research through the funding or as part of the Snell certification requirements?
I am not aware of any ongoing research into crash helmets or crash related head injury. I'm sure that there are some projects concerning helmet effectiveness and so on but nothing on the order say of the study of bicycle helmet effectiveness the Foundation sponsored in the 90's. Too bad. At this time, we're not in a position to sponsor anything that ambitious for motorcycle helmets, the only organizations that might be are national governments and, for whatever reasons, they're attentions are elsewhere.
These days, most of the helmet investigations seem to involve sports helmets, football headgear and so on. Currently, they're worried about multiple impacts and, for that reason, brain damage accumulated from a series of incidents which, taken individually, might be undetectable but manifest over time as cognitive disabilities and so on. I do not believe that these should be crash helmet concerns. Riders do not expose themselves or others to the succession of head impacts that are seen in contact sports. Instead, the concern is the possibility of a very few but very severe head impacts over a lifetime of riding. That is, no real concern for accumulation but we want to prevent death or profound injury due to a single incident.
Before I would look to institute any big changes, I'd like to examine carefully just how well. We haven't got the money and, it appears, neither does anyone else but I'd like to see a big epidemiological study where investigators would catch and follow up on every motorcycle crash in a given area over a given period of time. I'd like to see detailed accounts of injuries and outcomes, crash scenarios and exhaustive analyses of the headgear, if any, worn in the crash. If the numbers were big enough, we might well get a sense of what we might do to improve our methods. I'm pretty sure the results would confirm a lot of what we're doing already but who knows? At the very worst, I'd know a lot more than I do now and so would everyone else. How bad could that be?
I've read your mention of the 300G "survivability" threshold, but with questions of lower force transmission numbers, I'm wondering about not only survivability, but the threshold for traumatic brain injuries, including simple concussions, serious visual damage and other lingering effects of brain injuries sustained in a motorcycling crash while wearing a M2000 Snell-certified helmet.
It appears that a helmet protects against skull fracture by decreasing the impact force, but also protects against the internal forces within the skull by reducing the acceleration or the rate of deceleration. Are these two concepts completely interwoven or one and the same, or can they operate independently?
The discussion of the 300 g test criteria brings up the woeful state of ignorance about human tolerance. When Snively began looking into it, the concerns were death and profound, long-term disability. The experts of the time were using a 5000 lb transmitted force limit. This was converted
and moderated to 400 g's and, over time, was lowered to 300 g's.
The Foundation's directors will remain alert to new findings about mild traumatic brain injury, but they are not likely to trade away any protection from death and profound injury without some persuasive evidence that motorcyclists are going to realize some real benefits because of it.
Helmets seem to reduce the incidence of skull fracture by distributing the impact forces more evenly over the skull. In a sense, the helmet shell adds directly to the skull's bending resistance. Unfortunately the shock still gets transferred to the brain. In the old days, the coroners would note the skull fracture and were satisfied they knew the cause of death. When helmeted crashes started to come in, many people were astounded that there could be serious injury and even death due to head impacts even though there was no fracture. The speculation was that the shock sent pressure waves through the brain and that the strains due to these pressure waves were producing damage. These were thought to be most serious on the side of the brain directly opposite the impact site. Others thought that rotational effects might be even more serious causing tearing of blood vessels at the sides of the brain or diffuse injuries throughout sections of brain tissue.
For all these, the hazard may increase with the magnitude of the acceleration and with the duration of the exposure. However, there is little agreement how much acceleration or how long the duration.
For flat impact, the current liner material seems to control peak acceleration, that is, when the helmet smacks into an impact surface, the shell stops moving immediately but the head inside the helmet remains in motion crushing the liner as it moves, the liner resists this crushing with a relatively gentle braking force that is limited by the liner density.
If helmets are to work at all, they've got to transfer forces to the head. The whole business is to help the head make the transition from the initial impact velocity to zero. The helmet has got to apply some well-controlled and relatively gentle braking forces to the head to accomplish this transition. It may seem paradoxical but the higher these forces are, the more severe the impacts the helmet can manage. The real limit on helmets seems to be wall thickness. As this thickess increases, the weight of the helmet goes up rapidly, the wind resistance at speed goes up rapidly and user acceptance drops into the toilet. But for a given wall thickness, the higher the braking forces, the greater the impact management.
If the manufacturer has selected the right density and if the helmet liner is thick enough, the braking force will slow the head to a relatively gentle stop before the liner is squeezed to its minimum thickness. In our tests we measure the braking force by measuring the shock acceleration delivered to the headform. Force equals mass times acceleration. The liner material puts out the force, the mass of the headform determines the resulting acceleration. I believe that most liner materials generate pretty much constant levels of force throughout their useful crush range. Although, at the onset, the forces may be lower until the head has crushed its way into a good close fit with the liner. At the other end of the range when the liner starts to use up the last of its crush, the forces start to go way up until the liner is essentially as unyielding as cement. It has been crushed as far as it can go. If the designer has done his job, we won't get this amount of crush when we test, otherwise, the shock acceleration will spike upward, right off the chart.
Since the liner is trapped between the head and the helmet shell, it also exerts forces on the helmet shell. In response, the shell flattens against the anvil and, depending on its own stiffness properties, will also moderate the braking forces applied to the head. The shell properties are much more important in determining the response to the hemi anvil but the shell properties are important even for flat impact. In general though, helmets seem to apply braking force levels depending mostly on the liner density and, so long as the impact energy is not too severe. The peak forces and therefore peak accelerations will be in the same general neighborhood.
Instead of higher peak decelerations, greater impact severity produces longer deceleration pulses. If you slam on the brakes in a speeding car, the deceleration won't vary with the original speed but the braking period will. You'll brake twice as long from 80 mph as you will from 40. Same goes for helmets. I expect helmet shell properties may affect this somewhat though. For some helmets, peak acceleration measures may vary with impact energy. However, I'm not at all certain whether they'll be higher for some helmets and lower for others or what have you. But I'm convinced that the variance is not a concern and that a helmet that performs well at Snell impact levels will be effective at lower impact energy levels.
I've recently read some material that suggests the importance of foam density in combination with a thinner shell for a wider-range of protection. The authors cite penetration test protocols, rather than impact energy levels as cause for thicker shell materials. Is there data to support these observations or any data that supports the need for specific penetration protocols?
Also the authors recommend that flat-surface testing be given priority over less common objects, like the hemispherical surfaces. Is there any practical data available to suggest changes to design for providing better performance in one area or another?
As for the argument that we may be biased towards load concentrating impact surfaces, the fact is that a rider does not get to choose where he'll crash or on what he might strike his head. Even if he hits flat surfaces 80% of the time, he'll be right to be concerned about the other 20%. There is no necessary trade off between the two kinds of protection so why not go for both. At this time, flat impact requirements seem to set the liner density and hemi impacts seem to set the liner thickness. If we dropped either requirement, helmets would change: no flats and the densities would shoot up enabling thinner liners; no hemis and the liners would get much thinner and shells might disappear altogether unless the penetration test started to become a problem. Although I see complaints about the penetration test in the magazines, right now it's not a problem for Snell helmets. The flat and hemi requirements pretty much dictate that a helmet will stand up to penetration. Right now, the best argument to maintain the flat and hemi type testing in our standards is that the helmets that do well in that sort of testing also seem to do well in real crashes.
Our guiding philosophy for selecting energy levels is mostly a matter of the biggest hit for which we can hope to equip a motorcyclist. It might seem reaonable to work toward the biggest hit a motorcycle may have to withstand but this level is so far beyond any helmet that might be devised that it's hardly woth thinking about. When you pack a lunch you can think in terms of all you might need but for a helmet, we recommend all you can carry. The reality is that there are reasonably foreseeable crashes that might call for even more. This advice is so important, it's printed on Snell certification labels.
Some of our critics raise a big hit/little hit concern. They suggest that our testing is apt to miss something important because we're not impacting at low speeds. A helmet that cushions at high speeds could conceivably
act like a brick in lesser impacts. I'm certain that there are materials and systems that could b have like this but there are none being used for helmets. My best experience indicates that, for flat impacts, the ones that produce the highest peak accelerations, the lower the impact velocity,
the lower the reading. But in the unlikely event that some one does come up with a design where the big hit/little hit phenomenon is a concern, the directors are ready to reject the technology and the helmet as necessary.
With regard to the necessity of the two-drop protocol, or any other protocols that Snell uses, is there any ongoing theoretical study of these various protocols and specifications of the current M2000 or other standard in use by the Snell foundation, either by the foundation?
Two hits or one hit is more a matter of history and convenience. When the Foundation started, test methods demanded a great deal of ceiling height, even for relatively mild impacts. Hence, the first Snell standard called out two drop protocols. Over the next ten years, Snively, our founder, adopted much better methods but Snell certified manufacturers were also building much better helmets. Snively still needed two drops. Thirty plus years later, we've gotten a lot of experience with two drop testing and, there are lots of Snell type test rigs all over the world for manufacturers to do their development work. A while ago, we cut a hole in our ceiling and extended two of our test rigs up to drop heights approaching six and a half meters. These extended rigs may allow impacts on the order of 25 mph instead of the 17 plus mph first impacts we use now. A 25 mph impact is likely beyond the capacity of any motorcycle helmet I'm likely to see in my lifetime so we may finally have the gear to consider a one impact test protocol. But nobody here wants to rush. As nearly as we can tell from crash statistics, helmets certified by current two drop methods are saving lives. Before we endanger that and force Snell manufacturers the world over to swap out their test gear, I want some definite assurance that one impact methods are going to tell us something we need to know that the two impact method does not.
The two impact protocol will be with us for the 2005 standards and maybe longer. Its virtue is that after almost fifty years of testing helmets to two impact protocols, the injury statistics prove pretty conclusively that the helmets actually protect. The theoretical speculations we do here are concerned with how well a helmet certified to our standards might do when tested to some other requirement. The concern is that something in another standard might cause a Snell certified helmet to be rejected in spite of what we feel are superior protective capabilities. It is conceivable that a different test protocol might allow us to test helmets more efficiently or even identify capable helmets that some quirk in our current procedures has effectively barred. However, I can assure you that the Foundation's directors would scrutinize any radically new procedure very carefully. Unforeseen consequences being what they are, no one wants to yield any of the advances the current methods have won without some overwhelming demonstration that newer methods are better.
With regard to batch testing: I understand you are comfortable with the current random testing approach, but do you feel that a batch-test method would be superior for gaining credibility and/or ensuring the quality control and overall safety of a consumer? Here's a quote that appears to be out taken out of context in a press release written by Rick Marlin of SuomyUSA. According to Marlin,
"...this sampling can be as low as four thousandth of one percent (.004%) of annual helmets sold for the higher volume manufacturers. In one example, only 16 helmets were tested of an approximate 400,000 helmets sold of one particular manufacturer in 2001 (Source . Ed Becker, Executive Director Snell Memorial Foundation: http://www.smf.org/)."
Are these numbers accurate, and do you feel the actual performed percentages are within an acceptable range to provide the kind of assurance we deserve?
Also, could a BSI rejected batch of helmets end-up on the shelf as Snell approved, if the Snell certification was also met by that particular model?
As for standards enforcement, we're doing more testing than any other standards body I know of. I'm not sure where Rick Marlin got his numbers. He says from me and, if he says so I guess it must be so, but you and Rick should know that if I ever said any such thing, I was wrong. Current policy has us testing at about 1 sample per every 2000 produced. We don't do batch testing because it's too easy for an unscrupulous outfit to cheat. Not that we expect our certified manufacturers to cheat but why enable it. Even if we trust them, they surely can't afford to trust each other and, once one of them suspects others of cheating he'll find it a lot more difficult to maintain his own integrity. Therefore, so everyone can rest easier, we run our enforcement program on the same helmets that people are actually buying and wearing, we get our samples from dealers and distributors.
When BSI was mandatory in England, they could afford to run batch type enforcement on foreign helmets. It was a matter of trapping imports at the docks and holding them until batch tests were complete. We don't have that kind of government clout so, as a practical matter, we do in market enforcement. A short-sighted manufacturer might try to dump a load of substandard product but it's a sure thing we'll catch him and, for that very reason, there are damn few shortsighted manufacturers. If anyone ever forgets the long view, his bankers, bond holders and employees can be trusted to remind him.
There is still the possibility of mistakes, I have not seen any serious ones though. The penalties for serious RST failures can be catastrophic. Our solution is to pull labels from helmets in distribution and may go as high as issuing unilateral public warnings. Even if only part of a model line is affected, every product the manufacturer makes will be affected. He might never recover. If we were to prevent mistakes as well as dishonesty, a whole new order of enforcement would be necessary. For big government projects, they set up independent third party inspectors and testers right at the work site. Engineers and techs haunt the workspaces, watching and testing everything and signing off at each stage. It's expensive but, for major construction it may be worth it. However, it might be overkill on a high order for crash helmets. Until the public is ready for three-housand dollar hats, that level of security is probably prohibitive as well. What seems to be true is that, with our current standards and current programs, Snell certified helmets are serving motorcyclists very well.
I hope this is useful.
Thanks for your interest.
Ed Becker
Riders Gear Forum sponsored by Sportbike Track Gear
Interesting article on the Snell Foundation (a visit to the facility)
Linear Mode
