LocostUSA.com

I'm kind of unclear on this as well. (and I have a structural steel background). To me the tensile strength of the tubes is a more determinant factor. Ie in bending, one side is in tension, the other in compression. in regards to space frames (even tho we can't make perfect space frames) the individual members need to be in compression/tension.

It doesn't seem to me, that increase individual member torsional strength will have as much impact on overall space frame torsional strength as members with increased tensile strength.

I would definitely like to understand this better since I seem to be missing something

The downside to thinner wall is that the tube is ALOT more sensitive to damage. And a damaged tube (kink / dimple / what have you) will fail so much easier than an undamaged tube (it will fail at the damage). Just something to be concerned about for vehicles that actually live out in the real world, and not in the computer.

JustDreamin

Hi Mach68,

Could you be confusing strength with stiffness here?
Stiffness is much more important than strength in a frame.

Cheers - Gavin.

Young's modulas for all steels common to us are similar as to be considered the same, only the yields change.

So if you have lengths of 1010 through to 1020 and stand in the middle, they will deflect the same amount just the 1010 will eventually bend (as in stay bent/kinked) before the 1020 - but the levels of failure/force we are talking about would see your body ripped apart so again, no need to think about it, just grab what looks like reasonable quality at your local steel supplier and just be careful you don't get hot rolled re-cycled steel and that's only because it's a pain in the ass to weld with..

The differences though for our level of building isn't worth a minutes thought and chrome moly is a waste of time and money.

Material size choice is the most important, I dummy'ed up a couple of DeDion tubes recently, one from 2 lengths of 1.65"x.080" box and one from 3"x.080" tube and I can stand on the double 1.65" and jump up and down in the middle and get flex (actually feels soft to stand on) whereas the 3" tube doesn't budge at all, not a bit.

Great info and all in one place. TNX

JD (Oh, no! Not another JD!?! They are multiplying! ) thanks for the extremely valuable post. It might save someone (like me) some very real grief. Anything thinner than .065 wall is not going on the bottom of my car.

Mach68, I'm not completely clear about the relationship between torsion and tensile numbers myself. I do know that the tensile number is a material number and the torsion constant is a material plus geometry number. My instinctive understanding is that when the chassis is in torsion, the longitudinal members are seeing torsional loads. I did a quick comparison of tensile stress and torsional moments on a simple box frame in LISA (FEA in the hands of the "ignorant or the unwary", a dangerous situation) and it seems to bear this out. The longitudinals connecting the fixed (rear) frame and the deformed (front) frame must twist. But i'd really like to hear from someone with a better understanding of the fundamentals. I do not understand the numbers (-92.51 to -95.19) and color assignments in the Torsion Moment image. -a subject for another post, I suppose.





-Pete

We do have a thread where we do FEA on the base Locost frame and also on the frames that both Andrew and I have or are designing. I used the program "Grape", but I think Lisa is a good choice too. The models for Grape are available in that thread, and perhaps Lisa models too, so you can work with the real thing and that might be a little better.

You need to look at all these things. You hope that in a space frame your loads will largely be compression or tension. It is not practical to use small diameter solid rods though. They would not resist bending from casual use, like leaning on a tube for instance. They would not resist buckling in compression. Lastly as the frame does deform, which it must to respond to stress, things like diameter start to become a factor.

Having said the above it turned out on my frame the highest stress was torsion in the main roll bar hoop, above the braces. I should check if that is still true (or you could download my model and give it a try - it's fun! ). At some point you have to click thru all the options on the FEA program and look at all the possible loads. On Grape the numbers are given in PSI for American units. If you look at our models it should make sense, in my model when you put a 1000 lb. load into the rear coilover mount, you can see perhaps 10,000 psi in some of the front engine bay tubes. It's only a few mouse clicks to swap different size tubings around the frame.

Here are some things to keep in mind. When you scale a beam ( round, square or I-beam etc. ) stiffness goes by the fourth power. So if you double a rods diameter and try to twist or bend it, it will take 2 x 2 x 2 x 2 as much force ( 16 times ). This is why floor joists are always mounted standing on their narrow edge. If you simply double the height of a beam or i-beam it is 8 times stiffer, because it is not twice as wide. If you have any lumber lying around a quick jump on a 2x4 and a 2x8 will show a very big difference. You could even measure it for fun.

This liberal arts grad wonders if the impact resistance is the same for the thinner tubes as the thicker ones. In my mind, it would be possible to have a very stiff frame with thin-walled stuff that then folds like origami paper when hit in the "wrong" direction. At that point, I guess the thinner stuff would be easier for the firefighters to deal with as they cut me out (okay, that is dark humor)?

However, I'm thinking that these tables allow me to somehow get away from the theory that I need 2"x2" .25 thick tubes for everything, and that's very valuable to me. Thank you for the posts!

Pe7e does that LISA run Linux?
In your examples, I totally agree, the members are in torsion because of node geometry change, but if you add bracing as in a space frame you zero the load force vectors at the nodes. See how those numbers change then.

Mach, LISA and Grape are both windows only. I see your point about the force vectors at the nodes. And yes, the numbers do change, but I still can't figure out what they (the torsion moment numbers) mean. I think if I pursue this, I'll do it on the FEA thread that Marcus mentioned. (Thanks, Marcus.)


-Pete

thats a bit of a new term for me, but i think it is simply the moment force around the node (and perpendicular to the connecting member) Also, i think, because of the nature of the frame, it is going to be couples of moments since the forces will be similar on either side of the node.

i think FEA is the reason i always seem to resort back to drawing out force vector diagrams with a pencil and paper