Jeff Lamb

The thread titled “steel vs chro-moly cage??” resulted in a good discussion about material choice, however, it also included many references to the structural design of the cage really being the most important factor . . . and how many “pro built” cages are of a poor structural design (even though the fabrication and welding qualities might be good). This leads me to wonder if anyone on this list has designed an optimal roll cage using finite element analysis (FEA) or some other suitable method?

When various threads are started to show pictures of roll cage progress, there are usually some comments about fabrication quality, however, there almost always are significant structural design flaws that are pointed out. For the sake of everyone’s safety, can someone on this list point us to a properly designed roll cage using FEA that will work for say . . . 90% of the different types of Porsche sedans being converted into race cars? The inclusion of drawings or pictures to illustrate the proper design would be most helpful. This thread then could be referenced by anyone BEFORE they have their cage built OR if someone wants to have some or all of their existing roll cage modified.

When I refer to “structural” design characteristics above, I am referring to both the safety and the chassis stiffening characteristics of the roll cage design. A good starting point for some guidelines on the main aspects of a properly designed roll cage are the current SCCA and the NASA rules. They do a good job of covering the design of the main hoop (with diagonal and horizontal cross bars), rear bars and front section. However, there is not much detail around proper design of the door bars.

SCCA Rules here: http://cms.scca.com/documents/Club%2...CR/2009GCR.pdf See Section 9.4 beginning on page 88.

NASA Rules here: http://www.nasaproracing.com/rules/ccr.pdf See Section 15.6 beginning on page 59.

PCA Rules here: http://www.pca.org/clubrace/docs/08%20Rule%20Book.pdf See Appendix A.


So . . . maybe we should group the various roll cage designs into three different categories:

1. Bare minimum design to meet SCCA, NASA and PCA requirements. I have attached a photo of a very basic cage design that should meet bare minimum requirements. Please note that I couldn't find very many pictures of bare cages and this one even has an extra bend on each A pillar tube that shouldn't be there. Plus, note that this basic cage uses the "halo" design for the bar in the roof area and the door bar design using two tubes probably is best when the two tubes are placed in the shape of an "X" instead of the configuration shown in the picture I attached.

2. Cage in #1 improved with optimized door bar design. The outcome here will likely influence where you locate the main hoop in your chassis. This debate will involve whether or not the tubes should protrude into the hollowed out area of your doors? Or, whether the bars should stay in the same plane as the down bars of the main hoop and forward hoop? If you decide to have the bars protrude into the hollowed out doors, what design is best?

3. Cage in #2 now with additional tubes added to further improve chassis stiffness and/or safety. The focus here would be to point out the tubes to add in their order of priority. In other words, if I was going to add just one more tube to the basic cage, which one would be the most important to add? Or, if I was going to add two more tubes, which two are the most important to add? . . . and so forth. This is where things get interesting when you start adding tubes going into the front and/or rear of the chassis. Or, when considering the addition of a “Petty bar”, etc.

If the above has already been done in a prior thread (or threads), feel free to reference them here.

Thanks,
Jeff

DWalker

Im no artist, so sorry no cool CAD pics for you, however I will try and help-
thats a good "basic" design, but I would make a few changes, even on a "basic" cage.

1st- I would X the door bars and sheet metal gusset them with a brace to the sill, even on a basic cage. I have no idea why people seem to think the biggest danger is rollover when some of the worst injuries I have seen in motorsports happen when a competitors car ends up in your lap.
2nd- There needs to be gussets at least at the downtube to Halo joints.

Next step up-
1st- There needs to be a tube from the corner of the downtube and Halo at the A-pillar to the bottom of that down tube to triangulate that space and re-inforce the opening.

2nd-I would put a cross tube for the Halo from behind the drivers head to the pass front A pillar corner.

3rd- i would use a tube to connect the rear down tubes to the main hoop level with the door bar.

All of these "not required" tubes would be a slightly smaller diameter- 1" if the required tubing is 1.5", 1.25 if 1.750, etc. to save weight while adding strength

This EVO shows some of the things mentioned above, the only thing I would add if I did it over again is A-pillar re-inforcement and the use of smaller tubing for the rear X-bar.

Jeff Lamb

DWalker, that looks like some nice work . . . and really challenging because you were required to keep most of the interior in place. Some questions:

1. If you could gut the doors, would you have done anything differently with the door tubes?

2. I noticed that in the main hoop, your horizontal cross tube is one piece, whereas, the diagonal tube is two pieces. Is there a reason you chose to make the horizontal tube one piece as opposed to the diagonal?

3. In the rear seat area, there appears to be carbon fiber panels. What are those for? Are they actually structural pieces? Or, are they just "eye candy"?

Jeff

DWalker

1> Things I would have done differently in regards to the door bars would be to have tied them into the sills and if the doors had been gutted to have moved them a bit further out to have a better angle on the sill tie-in.

2>There was no specific reason the horizontals/diagonal were done that way, since they are in the same plane with no bends I dont think it matters much. I know there are some who will dis-agree.

3> The carbon panels were used for a couple of reasons. The fuel tank is directly under the seatpan, and a panel of aluminum, steel, or carbon was required between the tank and cockpit, the battery is behind the bulkhead, and one again we had the choice of materials. In this case it was easy beause the rally guys in Europe already had these panels done and ready to go.


One of the things thats not imediately apparent is the OEM crumplezones front and rear are intact. This means that the in the event of a front or rear impact the chassis will do its intended job of protecting the occupants. While the chassis might at that point be a throw-away, at least the driver will not bear the brunt of the force, while saving a lot of weight others might spend adding "reinforcement" bars. When this cage was constructed I was able to find a lot of imanufacturers information regarding chassis rigidity, which was 70% improved over the previous Evolution chassis. I thinka lot of fabricators jsut start bending tube and throwing the welder around without actually studying the chassis, its purpose, and what is already there.

RedlineMan

Hey;

The pictures above hint at a lot of things, but do not show much detail. One thing that does seem clear to me is that the rear stays and door bar extension tubes terminate on what appears to be the rear wheelhousing. I've never seen a wheelhousing yet that was a structural member of the FRAME (Even if it houses a shock mount), and so I do not understand using it for a major structural node. This virtually guarantees that the main hoop and door bars will rely on a non structural area for support. Since these are two major areas of crash concern, this puzzles me. Maybe I'm seeing it wrong, or don't know my EVO from my... elbow, but...

The little tubes that go from the main hoop up to the rear door bar extension tubes seem to be rather redundant. I assume they are meant to brace the rear door bar extensions in tension, but I'm not sure what is gained from this, unless it is some attempt to make up for the door bar tubes attaching to the flimsy wheelhousing?

IcemanG17

I would say the picture you posted is a bare minimum 6 point cage....most racing organizations require a 6 point, but prefer 8+ point..... Another fact is to use the proper gauge-size tube that matches your cars weight with driver and fuel....

Door bars vary from club to club.....I went a bit overkill on mine, with triple nascar style on the drivers side and double on the passenger side..of course the doors were totally cut out to allow this....but it does make it a bit easier to get in-out of the car too....heres a pic

DWalker

Thats a good point, and ordinarily I am right there with you. In this case there really isnt a readily available unibody "frame" or "tube" to get to. Our choices were the sheet metal "floor" or the double-walled wheelwell housing. Even double walled, a piece of 6x6x3/8" plate was formed to the wheelwell housing to spread the load rather than rely on a 4x4x1/8" plate. Also, the downtube was triangulated back to the main hoop and its lower mount, so that in a rollover the load would be spread back into to the hoop and its mounts which would allow the downbars to perform thier job of not allowing the main hoop to collapse.

The main hoops BTW are also mounted to the chassis using the same 3/8" plate as well as gussets tying the hoop into the sill.


Actually those little bars were put in to act as a sort of gusset to brace that bar, and thier somewhat less than optimum placement was dictated by the shape of the door card.

magnetic1

Foot protection is also a good thing to include. Like this:

fatbillybob

Iceman,

There is not overkill here at all. In fact you have underkill because your lower NASCAR bar is above the sill plate and not supported by a sill tube or sill plate. The bars pictured are a lot weaker than you think. If someone can do an FEA and post it here you will see how your design is not as strong as a simple "X".

fatbillybob

IMO the place to start with a cage is with your seat. The seat is everything. It is the first and most critical part of your safety system. Other hard barriers can sometimes prevent your seat from being placed in ideal positions at times. IMO the seat should be purchased first. Then you should be placed in the seat in the car at your ideal driving position. Then the main hoop relative to seat position then you design the rest of the cage. Ideal seat position is determined by your relative desire to be where you want to the wheel and pedals and to a lesser degree where possible tubes will be relative to your body parts. But before a mainhoop is placed consideration needs to be made for mainhop anchorage and how you will access the top of the cage to weld it by methods such as drilling a hole in the floor to drop the cage down through or taking the glass and roof off to access for welding. Weld method TIG VS. MIG mostly speaks to access rather than quality. MIG can get to tighter places. MIG and TIG done properly are equal welding processes. When seats are anchored to chassis floor extra consideration must be made for tube anchor points to the chassis. Anchorage in 2 planes is always best and lots of surface area is better but not lighter and sometimes an improperly placed plate anchor can transfer stress to a new area in the chassis and decrease a unibody's rigidity and strength. When seats are designed to be anchored to the cage then you can be a bit more compromising on cage anchor points because the goal of cage mounted seats is to provide a safety capsule that is the cage and the rest of the car can go away upon impact. When seats are anchored to cages the driver moves with the seat which moves with the cage in an impact. When seats are mounted to chassis floors sometimes the chassis will deform and the seat/driver will hit the cage in an impact. A counter intuitive tube placement issue is that a bar placed closer to you will hurt you less in an impact than a bar further away that you can still hit. So bars you can hit keep them close. Bars you won't hit the farther away the better. This is why the structural cage enhancement of a petty bar actually can hurt you more than help you because you will probably hit the petty bar in a crash. Lossening your belts so your shoulders can move 10" from your seat back will give you a rough idea of what you can hit in a crash because your belts and you yield this much in a crash. You will be surprise at all the thing you can hit. Make sure simple helmet hooks are far away because these items can pierce your helmet like a spear.

Greg Smith

I disagree, a simple X has 1 tube section in the center with a HAZ on top and a HAZ on bottom. Then this happens:

Professor Helmüt Tester

Amen. Saw a similar failure last fall, in a car that got punched hard in the passenger door. The weld didn't fail...the full-length tube itself failed. Drover spent a few days in the hospital from the hit. The simple X focuses all energy in a small area. I want more steel around me than the X bars offer, thank you very much.

DWalker

I agree-

While the picture of the failed X-bar looks bad, its obvious it wasnt done properly- no sheet metal to spread the load, not tied into the sill, and it looks like it failed at the weld which is interesting and raises questions in and of itself. Are there other pictures of the car involved? Any details of the incident?

RedlineMan

Yah...

That pic doesn't prove anything to me. It suggests that maybe that those are inferior materials, poorly designed, improperly handled in the case of CR, or that it was a big hit. If you can eliminate any of those thoughts, then I want to see it compared to the same hit with a usual " marginal" NASCAR setup like Iceman's. You either have to crash two cars, or do FEA.

I only have Intuitive Analysis to go on, but it's proven pretty accurate.

Larry Herman

Where would you tie an X bar into the sill?