tinman944

Geo,
sure it is a valid test.
now think of it this way. you have a door opening of 40"
you have a front A pillar bar and a B pillar bar. take two 40" bars and weld them together in the X .
if you where to fish mouth them so the fit tightbut did not weld them in and aplied force to them they would push through.
Now take two bars that would start out as 43" lond and put a bend in each till they fit the same opening. fish mouth them ans try and push them trough the oipening. there is no way it will ever go through. it is like putting a 41" box into a 40"hole.

bruinbro

I'll do the FEA tonight. What are some of the parameters (tube dimensions, lengths, impact force, impact angle, etc)? I'll assume rigid fixed end points.

BTW, tinnman944, the point you are making assumes that the A and B pillar bars dont move. That is incorrect as even a small force (hand force) would be able to spread the pillars apart assuming there was a vertical rotation capability (hinge joint) where the tubes form an X. This thinking scales with larger impact forces assuming a welded joint (no hinge) at the X.

Bro

Cory M

Bruinbro-

I was thinking of running on FEA model on this too because it is so controversial around here but I don't really have the time. Hopefully you have some good FEA software like Ansys and not just Grape or Cosmos.

You should run the model with 1.5" diameter, .120 wall thickness tube or 1.75"x.095 mild steel tube as these are the most popular sizes in our classes. Not sure on length - just measure your door. It should just be a relative comparison so none of the force or size values are critical as long as you use the same values for all of your analyses and just vary the geometry we'll have a good apples to apples comparison and be able to say that Model A is 10% stronger than Model B, etc. I'd like to see a comparison between a flat X, a bent X, Nascar bars, and a bent X with a sillbar and gusset.

Geo

FEA would be fantastic. Just model something besides a perpendicular hit. I will concede that a truly centered perpendicular hit would find the bent X to be stronger. However, hit to one side at an angle (45* will do) and I'd bet those tubes would be more likely to collapse because you could start collapsing both sides of the (bent) X before seriously engaging the main hoop and a-pillar tubes. At least that is my hypothesis.

complexx

I wouldn't bother wasting your time with FEA. Its extremely difficult to apply a load to nascar bar structure. Even if you try, the results won't be realistic enough to use as a means of argument. The reason why I say this is because I asked my buddy to do it for me and he pretty much told me its way to complicated and difficult.


I see the benefits and problems with both systems. I would think that the X brace would promote maintaining the cages shape in a collision and the nascar basket is really designed to deform and absorb the impact. I do feel that by bending the X brace outwards, you are losing some rigidity of an in plain X brace for the ability to absorb the impact, but it won't absorb the impact like a nascar basket would. Also, you gain the rigidity factor that the basket does not have (basically the characteristic that the cage will deform less and maintain its shape around you).

bruinbro

I guess I don't see why it would be difficult to load up a NASCAR bar structure. You might have to do it somewhat parametrically if you want absolute values, but event that is just time consuming, not difficult. As far as looking at what the differences in performance in various door bracing/protection cage designs it should be very straightforward.

Bro

bruinbro

I think I know the geometry for the flat X and the pyramid X, but can you describe the dimensions for the NASCAR bar and the X + sill bar + gusset?

BTW, I use SDRC(now EDS) I-deas FEM and Model Solution.

Thanks,
Bro

bruinbro

Does the X terminate at the main hoop or the door frame in the rear?

Bro

bruinbro

I think this correct. The vertical tubes will be restrained for the analysis at the tube ends.

Bro

RedlineMan

Awesome!

The X type is accomplishing many different things simultaneously. It is offering not only side impact protection, but front as well. It is keeping the firewall from moving back, which is a legit concern in some cars. Not only that, but improving handling through increased torsional stiffness as well. A flat X is better for front, giving up a little side, and making ingress/egress a little tough. The bent Pyramid variety is accomplishing MANY tasks too a high degree, adding a bit easier ingress/egress to these without compromising any of the others tremedously. It is a bit of a compromise, but a good one. It is also much easier to render, to a much lighter degree, than a truly effective NASCAR setup.

RedlineMan

Exactly!

I'm not sure you know you nailed it, but that is what makes them effective. I've tried to make this case before regarding head tubes, but I don't think people buy it.

A force applied to any tube welded in between two others will immediately transfer load, but it is how that is key. Say you have a straight header tube welded between two side hoops. If a downward force is applied to a straight tube, it will transfer load to the side hoops, but its first reaction will be to COLLAPSE as it pulls in. The only thing resisting is whatever structure stiffness there is in the roof and tubes in that direction. Since the head tube is really the only thing resisiting this implosion, your losing out. This collapse is counter to our goals.

A head tube with some arch in it will have the immediate reaction of pushing out on the side tubes. Its first reaction is to RESIST COLLAPSE. This resisitance is provided by the outward force trying to TEAR the roof skin and header beam, and it is very hard to tear something, relative to bending it. This resistance may be just enough to counter the relatively momentary force of a roll over. With any structure, multiple loads will be a crap shoot, and mostly down to level of over-building.

That's how I see it. Now... I'm wishing I put a little arch in MY door bars.

Cory M

There are about a million different ways to do the Nascar bars, so I would just use something generic. How about two horizontal bars with 45deg bends at the ends (enough to offset the bars about 6 inches away from the driver) and then put in three vertical bars, one at each of the bends and one in the center.

The pyramid X (calling it a pyramid sure makes it seem stronger than calling it a bent X) with the sill bar should be the same pyramid bar you analyzed before but this time add a straight horizontal tube across the bottom that connects at the ends of the X. The "gusset" is just a straight vertical tube that attaches the vertex of the X to the sillbar.

The aft ends of the X terminate at the main hoop, the forward ends terminate at the side hoops. You should probably just fix all of the ends in the X,Y,Z directions to make sure it won't move around, that is good enough for modeling purposes. Remember that it's just a relative analysis, so as long as the same constraints and boundary conditions are applied to all of the models we will be able to compare them head to head.

complexx

The design I am intending to use does not have the sill bar gusset. It has REAL gussets where the two tubes meet to form the X. If you guys are going to try to model it, model it with 2 gussets on opposite sides, and then model it with 4 gussets (one in each of the corners the X brace makes).

Keep in mind, these are GUSSETS. Not little dinky gussets. I'm in the office right now so I can't send the picture of the gt3 cup car I would like to upload. But the gussets on it are sizely. If I had to guess, I'd say that the two sides of the gusset that are equal length are probably 5-6 inches long each.

Geo

John, there is another side to the things you mention.

1) The arched tube: A nice flat impact will indeed tend to spread apart the tubes it's connected to. However, a nice sharp impact in the middle (God forbid Armco in the middle of the top front cross tube) will tend to bend inward before even pushing or pulling the side tubes to any extent. It will bend inward relatively easily since the sides are not coming into play. With a straight tube there the sides instantly come into play.

Something to think about. You've got me thinking a bit. I think everything depends upon the type of hit any tube takes. In some tests one design will win, in others another will win.

2) I'm doubting the pyramid X will have near the torsional rigidity of the flat plane X. In fact, I'd expect a 3 horizontal bar NASCAR style set-up would be torsionally stiffer, but I'm just a monkey scratching my head here. But model this with any material you like and I'll bet the pyramid X will move around much more easily.

Cory M

complexx-

Here is a picture of the gussets you're refering to (on a 997 cup). They should make the structure quite a bit stronger, unfortunately they are a lot harder to model than just tubes (I assume Bruinbro is using beam elements for the tubes). So if I were doing the analysis I would leave them out of the first iteration. The cases that I outlined before will give us a some good insight into how the flat X, pyramid, and Nascar compare.