Well guys I posted a while back about trying some different configurations of roll cages. I was specifically interested in what to do for the side/door bars. Because this was a street car, I really hate to put a "X" bar in, because of the difficulty's of getting in and out of it all the time. I kept debating between a single bar, and a "bolt in" X bar (same idea as a swing out bar, but in an X shape)

Finally, since I had it at my disposal, I decided to FEA the different types of configurations.

For those that don't know FEA stands for finite element analysis. It's basically a computer simulation of a crash.

What I did was a simplified model of a side impact with different bar configurations. I didn't get too crazy because these take 10 hours each with 5 computers working on it at the same time, just to run these simple models.

In addition to the single bar, and bolt in "X" bar, I decided to try a "T" style bar. The extra bar runs from the single bar, back down to the rocker bar.

I was amazed the difference that this "T" bar made!!!

The graph shows energy in the "Y" direction (up and down) and deflection in the "X" direction (left to right).

You can see the t and x bars are way better then just a single bar. The point where the curve flattens out is actually where the failure point is. They continue up slightly, just because it's still connected at the other end. But disregard anything above the point where the line gets flat. . .


The crash block was done as a constant energy. No weight, no inertia, no speed. Just a constant push, never slowed or stopped by the bars. That way you can see exactly when each one breaks, and how much each deflects.

So for example, if you take a crash of say magnitude 1.5 (in the up and down, energy is unit less in this case) the single bar (yellow) has about 70mm of deflection. If you stay on that 1.5 horizontal energy line, and move to the left. You can see the t bar only has about 40mm deflection for the same energy (crash).

And of course as you go up in energy (crash severity), the t-bar does much better.

The "x" bar in this case has a little bit more ultimate strength then the "t" bar, but not by much. Of course if you focused the crash down towards your knees a little, the x bar would excel vs. the t bar. But in all honesty, I'm more worried about being hit in the hip/pelvis area then I am the legs. You're a lot more likely to be seriously injured in this crash location.

The extra bar to make it a "t" doesn't interfere with getting in and out of the car. The weight increase is minimal, but as you can see as far as strength goes, it's a significant improvement.

Anyway. . pretty neat stuff. . .thought I'd share it with you guys.

single bar pic

"X" bar pic

"T" bar pic

Just to add too, the impact was done in the same place on all 3 bars. I hit it basically right at the hip point. The T bar intersects the single bar at this point.

If you guys cant see these pics, let me know and I'll post them in a different format.

You should just be able to click on the attachment, click "Open" if it asks.

thanks for the info!

they opened fine for me!


Dave

Great info. I might add that T bar to mine. never thought about it. and they open just fine for me.

Scotty very nice work. What material and wall thickness? Can you redo them with .085 CM and .120 MS? This is data I would like to see from SFI since they're bent on removing MS chassis from faster classes. I'd like to see data on when Koretsky removed all the bars on the passenger side of Bruce Allens pro stocker. I know the impact was a very rare circumstance but how easily they were removed was scary.

It was done with .134 mild steel 1 5/8" bars.

I might be able to re-run Moly later, but we're slammed now with new development stuff so it's gonna be a while. I just happened to catch our FEA guys/computers when they were open. . .

bump for the night crew

What caused the discontinuity (jog) in the X-bar curve? Would adding another tube at the T running back down to the main hoop to lower rocker tube improve the strength with additional triangulation? Certainly it would add extra weight, but I would be curious how much strength it would add?

Nice work. I'm suprised the T-bar was close to the X-brace, but the impact is in the middle of the tube in the X and at the intersection of the T.

I was just thinking that the other day the dart has a cage with a single side bar and want to make it a swing bar but want to put some thing more around the drivers butt might go one more step from the top of the tee up to the roof bar aka funny car style

I think the "jog" in the x bar curve was from where the tube "kinked" from the edge of the crash block.

My guess is adding the second bar to the "t" probably wouldn't do much for the crash in this direction. I think the "t" bar as added takes care of the pull tension in that direction. However, from a frontal impact, I think that yes, adding that bar would help. This certainly wasn't a perfect model, the crash location really favors the "t" bar, but like I said before, it's probably the most likely area that you would be seriously hurt in a side impact.

I noticed in your drawing that the single and X bar arrangements had a horizontal bar at the bottom that was gey in color. This bar was not present in the T shape arrangement. What significance does this hold?

Damon

OK, you have data for only 1/2 the reason behind having the side bars. The other is to stiffen the chassis. The X is much better as it ties the chassis in 2 different directions from the main hoop to the a-pillar bar (firewall bar).

I'm pretty sure I already know, but what happens if you move the point of impact down the tube on the "T" bar?

Now picture what's on the other side of that T-bar after it rips apart at the HAZ and bends inward from the impact.