Let's think about this. The real enemy is rapid acceleration. This is what needs to be slowly dissipated in order to reduce injury. If the bar is some distance away (and that distance is not known -at least by me) you may incur more injury by hitting it than not. If the cage could be built as a safety system with body restraint incorporated then it would be a viable solution but to just put bars in close proximity without knowing the effects of a non-lethal biff is just crazy.

Surely the future will have body retention and H&N restraint with a purpose built cage to contain all these systems. It is the natural progression but it is not there yet.

 

Woah... hitting roll cage tubing with body parts increases safety? New one to me....

Seriously, I think there is a flaw in this 2" vs 8" debate.... when you are in a high speed impact, the accelleration comes from the car stopping suddenly, and your body still going at speed .... I would guess the impact is the same whether you are 1 inch away or ten feet away... you will go at roughly the same speed until you hit something hard and stop.

Re-look fairly graphic descripion of impact by Redline.... your belts strech 10 inches in the first 30 milliseconds...or something.... I just can;t believe driving your head (even with helmet) into a metal bar with that kind of force can ever be a good thing..... its like..."Good thing that tree was there or the idiot would have gone off the road completly..."

Somebody has to show me some pretty compelling data for me to believe that.....

As far as sub belts.... I agree.... the term "submarine" is probably a poor one to use.... I see the sub belt job as assuring that the lap belt stays against your pelvic bone, and doesn't ride up into your abdomin.... if it is riding at any angle shallower than straight down seems to me it will be less effective.... an no.... I don't want to stop by using the sub belt against my *********... not a good idea IMHO

 

Remember, ain't no expert here, no test data, no sled info, just a-guessin' BUT I think if you are talking about a helmetted head ... since your body is strapped in to a seat and since at worst, a Nylon belt stretches 12% on impact (not empiracle data, just what I have heard) that hitting from 2" would be a harder impact than hitting at 8" since the latter had 6" more travel-time or more importantly deceleration time from the harnesses that are trying to restrain you.

John, we might be thinking of different examples in the same case and just not seeing in the same light. I want the room primarily but ironically, the example of the ice racer and what I was talking about on the Miata aren't too different - just that the gussets are on the top plane not the side (for more open space I'd imagine).

 

Currently the top of my helmet hits the front-rear horizontal roof bar on my cage if I move it a few inches sideways..... I have SFI padding

Part of me says.... bad news... I don;t want my head hitting the roll cage in any manner or form..

Part of me says... good news...instant head containment device... better than basal skull fracture.

Part of me says... bad news.... now I have torque applied to the top of my helmet levering my neck... won't have to worry about basal skyll fracture.. I;ll get a severe compression injury instead...

I just ordered a Sparco Circuit... given all these options I;d prefer if my head hit a padded energy absorbing surface....

 

John, point is that if you do not have time accelerate then impact will not be as great.

Think of this way.

Assume for the sake of argument and unbelted occupant.

Car stops due to impact. The person accelerates forward with respect the reference frame of the car. Since the person accelerates with respect to the interior the impact energy is less with sorter impact. Or if you will a shorter "fall" since the body (person in this case) has less speed. Longer distances mean more duration of acceleration and this higher speed at impact.


Now that I have established that an accelerating object will hit with more energy than at a farther distance we need to establish the the object is in fact accelerating. So to do this we need to establish a reference frame. For now lets use the seat of the as the reference. At time just before impact the we will assume a constant velocity of the car and there or reference frame is at zero velocity and so is our unbuckled person. We also know that just before the person impact the steering wheel they have a velocity. Newton's law tell us that to go from rest (in the car reference frame) to a velocity the body must under go acceleration. So therefore the person is in fact accelerating. You simply can't go from zero speed to XX speed without accelerating.

Now this all in the reference frame of the car. Lets look at it from reference frame of the ground.

Car approaches a brick wall some fixed constant speed. In this case there is no acceleration anywhere. At some instant the car impacts the wall undergoes a very rapid deceleration. This is clear. No what happens to the unbuckled occupant. Well again Newton's law tell us that bodies in motion tend to stay in motion so the person continues on at the original fixed speed. unfortunately the car is not longer at that speed so deceleration of the car causes an apparent acceleration of the person with respect to the reference frame of the car. So what this means is that car could appear to be moving toward the person. This is where it gets complicated. As the impact of the person to inside of the car occurs the person decelerates at rate that will eventually equal that of the car. Once this happens they continue to decelerate at the rate of the car.

So to understand the actual impact figure it like this.

Lets say it takes 5 seconds to decelerate from 150 mph to 0. Clearly not impact rates, but useful for our discussion.

So we can calculate the deceleration rate. We know that just before impact the person and car are both a 150 MPH. We know that after the impact both person and car are at 0 MPH. So how did they get there?
At time zero impact occurred. The car began to decelerate to zero. The unbelted occupant continues on at 150 mph until there is some force to restrain him. Since we all know that it takes some fixed time to cover 6" at 150 mph and a longer time to cover 12" at 150 lets make some guesses. What happens if the person is 1" from the steering wheel. Well it takes you lets guess 1 second to hit the wheel (not actual, but you can see were I am going). Then you still need to decel to 0, but now have 4 seconds to do it in. Not good. Lets say the person is 12" from the steering wheel. It would clearly take longer to hit the wheel than if you were 1" away so lets call that 3 seconds. So now in the 5 seconds it takes for the car to decel 3 seconds of that time is speed with there person still going 150 mph. That leaves only 2 seconds to decel this making the impact much more severe than if the person was closer.

Of course all of this is for a unbelted occupant. When you get belts involved the equations get much more complex. This is where belt stretch is good. If you have good belts you may in fact still be decelerating after the car has stoped because the energy is being taking up by the stretch in the belts. Thus the 5 second decel may in fact be more like 6 to 7 seconds for belted occupant. When you mix belted occupant hitting objects it gets even more complicated. If the occupant hits near the end of the belt stretch then they may have already undergone much deceleration and this have light impact. If they it happens right away they may not have had time to accel or had time for the heads to accelerate to a high speed yet. If you are in the middle you may be hosed. Enough distance to accelerate vs the insides of the car, but not enough for the belts to be slowing you down yet. Bang bad news.

So the simplest thing is to prevent contact with anything. Makes it much safer, but if contact is going to occur it best done either very early before the body can accelerate or much later when it speed is being reduced.

There the simple answer.. In theory...

And yes i am am a trained mechanical engineer from MIT. Sorry everyone in while it catches up with me.

PS... I am certain you can find holes in this logic. I have not been thinking about for long enough to root all of them out. Also engineer's can't spel! or use the gramer so gooder!

 

M758,

There's some good stuff in there, but I think you clouded it with the "acceleration" parts.

It may seem like a minor point (and at one stage you actually make it quite well) but in an impact, there really isn't any acceleration taking place... only relative rates of deceleration. For the purposes of this discussion, I think it's much easier to think in the single term of deceleration...

The real problem in an impact is that deceleration occurs at different and excessively high rates for the various bits of the car and the occupant.

 

Acceleration and deceleration are really the same. Difference it the sign of the vector. Newton's laws do not mention deceleration as they really the negative of acceleration. That said using both terms as I did may have made things more confusing.

 

Stop this! This is crazy talk!!! Too much like work, I'm not paying attention any more unless I'm paid to!!!

j/k!

 

What counts is the relative speed of the cage and the head, and that can be roughly calculated on the basis of crush distance. One of the sources cited in an earlier thread on the need for sub belts in a formula car dealt with this. The car decelerates as the front of the car is crushed, and this crush distance might be two feet maximum. The car thus decelerates from x mph to zero in two feet, and assuming linear deceleration, one can calculate the rate of deceleration expressed in g's. Say the decelerations is 45g's. You then have the cage and the seat belt mounts both decelarating at 45 g's, while the force on the driver from the belts is 45 x the driver's weight, since f=ma. Belt stretch will reduce this force somewhat, and if we know the coefficient of elasticity of the belt, we can calculate that reduction. The speed of impact of the driver's head with the cage can be calculated by determining the time it takes the head to travel from its resting positon to the cage at 45 x 32 feet per second per second relative acceleration (less the effect of belt elasticity), calculating its speed at that point, calculating the speed of the car at that point in time, and subtracting the difference. And after the head hits the bar, it stops, while the torso keeps moving. It ain't simple. I can't wait to get on the track so I'll stop thinking about this stuff.

 

Taint no signtist neether, just a thinkin...

Here's the scoop. Kinetic energy is the real demon here. Anything unrestrained will continue to build it over time. Kinetic energy equals one half the mass times velocity squared (KE=0.5MV^2). THAT is the scarey deal. Velocity SQUARED.

When the car experiences negative acceleration (the correct technical term I believe), the body will continue to move, and effectively accelerate relative to the car and whatever there is inside it to hit. This is not instantaneous acceleration, but ramping up over time until it is impeded by something. The graph would look something like this ( ^ )(upside-down V). The body accelerates relative to the car until it hits something, and then decelerates back to zero eventually.

The real problem is the length of time that the body is allowed to accelerate. Anything that decreases this time could "theoretically" be a good thing. Since kinetic energy is a function of the SQUARE of velocity, if you cut the time short, you reduce the acceleration, and therefore the kinetic energy. If you cut the velocity by 2, the "square velocity reduction" (??) is 4. If you cut it by 3 then the reduction is 9. Etc., etc.

It changes FAST... for the better. Therefore, it may be "possible" that to hit something early would indeed be far safer than to hit it later. Interesting in theory.

 

The mit engineer said:" So the simplest thing is to prevent contact with anything. Makes it much safer, but if contact is going to occur it best done either very early before the body can accelerate or much later when it speed is being reduced."


O.K., If we buy that then the question is with your typical 5 hole seat if you have a 20g crash by someone T-boning you and a 20g crash into a cement wall head on how much does the average body move when 5 point belted? This is a critical issue because if your head is going to move with seatbelt stretch on average 12 inches no rollbar is that far from your head unless the transmission tunnel is between your legs. Therefore, you might as well just pad your bar next to your head and be done with it and keep it close. If a head yeild only 2" and on the average car you can get the bar 4" away then go for no contact and let your helmet restraint device do its job and let the roll bar do its job. So the million dollar question is how far do I move in a crash? Anyone have any data on that?

 

Well, FatBilly;

It depends on things like belt length and driver mass, etc. HANS pretty much says that Stock Car drivers WILL hit the steering wheel. Period! Not too hard to imagine given their seating position. Here's the scary one. In sled testing on a formula cockpit, they say the occupant would have a moderate impact with the wheel. Their current model reduces head extension by 7.5" over a non head restrained occupant, and 3.3" over their earlier prototype. What the heck would be the travel if un-head-restrained and no wheel to hit?!?! I think the answer would be frightening.

Look at the test photos and video clip on the Test Results section of their website. This should give you some idea of why we are seing containment seats becoming popular!

http://www.hansdevice.com/testResults.html

 

The point is you need a seat with complete torso retention (like the Sparco Touring, which provides shoulder retention) and a H&N restraint (like HANS). If a driver were to design a car with the cage as an integral piece of containment, then the body part which will be contained should be in contact with the bar while driving, thereby reducing the acceleration of the body part to that of the bar at impact (which is the safest scenario).

I don't want to hit the bar at all. Let's use the left sided bar that goes by your head as an example (since this is usually the closest bar). If you were to hit this bar your head will most likely be pushed to the right while your torso continues to the right. This can't be good. Another bar is the one directly above your head in front. If you were to hit this bar, your head would be pushed back while your torso would continue forward, again not a good thing. The seat is the best place to be restrained during impact as it should accelerate at the same rate as the car. Now maybe the seat should be incorporated into the cage so it would be stable at impact, but this brings up other safety issues.

 

Ideally you want the body to decelerate at slow a rate as possible all in the same plane. Simply put you want the entire body moving as one. In the case of rear impacts this much easier to achieve since the seat does a good job of holding the body in.

Here of course where seat flex is good. Some seat flex will allow the body extra time and distance to slow down this imparting less G forces. This is good. An infinelty rigid seat will decelerate exactly at the same rate as the car.

Now in side and front impacts the driver is not containted as well as in rear impacts. In most racings seat the hips are well controlled. Legs and feet seem quite tolerant of some realtive motions and often times the foot space in the car is such that they can't move too far.

Problems result in the upper body. Holding the ribs is easy to do, but the ribs themselves are not very strong to support that loading. Much better at the shoulder level. Most all road race seats have some shoulder retention features. Right now there are really two times. Traditional wings and the newer "full containment"

The real complication comes from partial impacts. What I mean is some rear component and some side component. In those cases I think the traditional wings are quite good. In partial front, partial side the full containment has and edge. In full front it provides no added benefit since it is only the harness that holds you in.

So we are truly approaching the point of the newer safety features really being useful in only certain types of impacts.

Yes safety is a system.


When it comes to theory however it is important to understand the basics, but at the end of the day testing is still the best method to determine what is really happening. For year car companies have conducted crash testing. They still do today in spite of alot of analysis they do. In fact a good portion of their analysis is based on past test data. There so man factors involved and when you start talking about deformations and human injury levels it is very hard to do it all by theory. Testing is very important.

 

Hold on... (this from a guy who flunked 9th grade algebra...understand). In an accident the body does not "accellerate" it continues at its currrent speed until it hits something and then deaccellerates at the instant your body touches that item....one micron before that contact you experience 0 Gs.

It doesn't matter if you went 10 feet, 10 inches, or were physically touching the object... as long as nothing else restrains you the deaccelleration will be the same... and the injury..as long as the position of the impact is the same, .the fact that your head is touching the bar does not change the amount of energy you absorb... just that you absorb it a few milliseconds sooner.