Elliptical tube chassis, question

[tomislavp4]

Hi guys, I´ve read about few cars having eliptical tube chassis, mostly cars from the 50´s. Now I know it´s more difficult to work with elliptical tubes compared to oval but does anyone have any experience in working with them? How harder is it?

I would guess that bending them is quite hard and then off course finding a suplier

Thanks

[marcush.]

first of all ,
I hope you have very good reasons to ask for oval tubing,of course it will look cool but then you have a lot of joining detail design to tackle...but then again
bikeframes even use confied tubes with variable gauge thickness,so I guess theres really no end if you have the recources available to tackle a job that sophisticated.
At the end the qusetion is ,will the car be faster,stiffer more reliable,or just
that much more expensive?
I guess theres god reason for contemporary chassis builders to use what they use.
Of course all shapes are available readily ,maybe only in 25crMo4 if you go for
a very weird profile,but oval tube is available in the more desirable 1.7734.4
spec wich does not require heattreatment before/after welding but still you won´t believe how hard this stuff is to work on,so you need very precise drawings and tooling to get the desired pieces to fit your chassis.It will definitevely not do
having a cuttersaw and a Migwelder and a flat plate to do such a job....

[tomislavp4]

Thanks for the replay marcush!

I think that with a good CAD program it will be quite easy to design the whole thing, if you´re good in it. So the problem is to replicate the design in the real world... Let´s say you hire a experienced guy to do it for you, how much more expensive would the elliptical chassis be over the oval one? 5-10 percent or more? What do you think?

[Ciro Pabón]

Well, I don't know anything worth of mention about welding for a chassis, sorry for not being able to contribute. I make holes when welding, so...

Anyway, I know a bit about structures. I think your assumption about stiffness is incorrect.

Yes, after a quick check, an elliptical section has around one third of the stiffness in compression of a round section, for very similar cross section areas and, hence, weights.

The failure, in the exercise I'm looking at right now, at Bresler & Lin & Scalzi's structures text, occurs by buckling. I'm talking about columns and the such, where loads are more or less symmetrical.

If you're using a member that works in pure bending, you could go with the elliptical section, aligning the long axis of the ellipse with the loads, altough if that's the case it beats me (please, understand I'm civil) why you wouldn't use an I-beam (or a box, for torsion). But if it works in compression, in principle, forget it: you're chosing a worse section.

Trying to think as a mechanical guy, I'd bet that elliptical tubes could be used if you're thinking of buckling them in a crash: even in my simple book that's mentioned (for joints). Anyway, I also think you should be aware of the stiffness you're giving away in compression, and incorporate it into your design.

Don't do anything based on my limited knowledge of structures, please. I'm used to security factors most guys here would laugh about.

[tomislavp4]

Thank you for the arguments Ciro I guess I´ll stick to ovals, I was just curious about the elliptical ones.

[Ciro Pabón]

Well, I'm also curious here. What's the difference between oval and elliptical?

[Belatti]

I really really doubt of you this time Ciro, but anyway:

An ellipse has this mathematical property: is constructed by the path traced by a point whose distance from a fixed point (focus) maintains a constant ratio less than one with its distance from a straight line not passing through the focus, called the directrix.



An ellipse can be obtained as the intersection of a cone with a plane.
An ellipse has an equation that can be represented in the Cartesian plane:
Ax2 + Bxy + Cy2 + Dx + Ey + F = 0

In the oval, the figure is constructed from two pairs of arcs, with two different radius. Any point of an oval belongs to an arc with a constant radius (shorter or longer), whereas in ellipse the radius is continously changing (see property above)



I really dont remember the name of that figure that is constructed by two straight lines and two arcs.

[ReubenG]

I've had the benefit (pain) of building three space frame chassis (or is the plural chasses ?) from circular cross section tubing. Because we did not have access to CNC profiling for the tube ends (or a mill with easily adjustable shaft angles), we did all the end profiling by hand (files for the new people, angle grinders for the more experienced / brave / stupid). Needless to say there were many discussions as to whether the welder could fill a gap left by the filer, or whether the filer should get another length of tube and start from scratch..

If you are using some kind of FEA for your chassis design, then I assume you will be optimising the orientation of the major / minor axes of the ellipse. The issue then becomes whether your manufacturing technique can align theses axes to the tolerance that the design / analysis demands. If you are profiling by hand, it's enough of a mission to get the tube axis at the right angle, with the nodes where you want them, without the welder filling in big gaps, for plain old circular tubes.

Ciro talked about elliptical cross sections being less stiff in compression and more prone to buckling. An elliptical section is only really beneficial if you can align the major axis with the plane of the bending moment. If you have a well triangulated space frame, then the tensile / compressive stress should be more significant than the bending stress. So elliptical tubes don't appear to have a significant benefit over circular tubes - but as one of my colleagues says, "You can't make that decision till you've done the numbers!".

An aside - most chassis failures I experienced were all at welds (perhaps this has something to do with the afore mentioned filling of gaps between tubes by welding ) with a single case of buckling. So if you are going to design tube members close to the limit, then designing the joins properly becomes even more important.

[tomislavp4]

Well if you want to make a chassis without body pannels, something like the Atom, they would be better from an aerodynamic point of view, because of their streamlined profile. And I was also under the impression that they are stiffer than ovals if the loads are parallel with the longer side of the ellipse, untill you said they are not. If they were you could design a stiffer suspension with carefull planning. In the given example they would be better from an estetical point of view too....

That´s why I was curious

[Belatti]

I cant seem to understand the differences regarding difficulty for welding/construction between oval and elliptical, so I guess tomislavp4 confused circular with oval

The problem here is how do you select the beam section to design a chasis?

I´ll go further: why do you need to build a chasis with only one type of beam? Maybe there are cross sections that suits better some positions than others...

After all, it all comes down to geometry, mechanical properties and desired failure modes (Ciro mentioned buckling, thats for example the failure mode of rectangular sections to bending impacts, or almost any other slim compression case )

If you have not got FEA, I would check stiffness in compression first, because its not a very "intuitive" property. To "see" torsion instead is a bit easier (remember membrane theory). I will risk and say simple traction is more or less the same for all sections and depends just on material, but we all know you never gonna have simple traction

[marcush.]

the cost side was answered already at least in part:
with some experience and good basic equipment ,I´d say around 15k € would give you a top notch outfit to build a space frame from square or round tubing.
And yes ,a 1,5mm gage tubing does need a perfect zero weld fit to avoid distortion from excessive heat when bridging the gaps .

If you´d use sophisticated oval or elliptical or teardrop tubing ,first you get the nightmare of getting your calculations regarding stiffness in all directions correct,of course you are lost without 3D-CAD design and someone doing the machining on the the parts with the needed precision.So most likely you will need
100% outsourcing on getting every single bit of your tube frame ,this will easily get very costly ,,my guess is around 100 machined parts ,most of them individual
or mirrored parts at best so you might be in for a lot of CNC hours just because of the many hours needed to programme and fit the bits into the machine.
I´m sure this is a lot more than 10% additional cost,it is at least 3-4 times as expensive and if you don´t happen to have a CNC Machine shop definitevely out of scope to do as a hobby project.
So ,I´d say it is definitetively only an option if you really need the elliptical tubes for a special reason ,but I doubt it is justifiable but of course it would look the part..
Luckily with modern CATIA V5,ProE etc you will always be able to put these dreams to paper ,err desktop.

[Ciro Pabón]

Thanks, Belatti. Of course I know what an ellipse is (c'mon! ). I had no idea that ovals were so strictly defined. I've handled ovoids (for road pipes) but never ovals, nor did I know that there were steel sections in oval shape.

[donskar]

From the Jurassic Park Dept of racing memories . . .

I recall reading (many times) that 1950s Ferrari race cars and 1960s road cars used oval tubes for their main members because they gave greater torsional rigidity (I suppose that might make sense). They were not in widespread use, supposedly because the greater difficulty in joining them (that does not make much sense to me).

[Belatti]

Thanks, Belatti. Of course I know what an ellipse is (c'mon! ). I had no idea that ovals were so strictly defined. I've handled ovoids (for road pipes) but never ovals, nor did I know that there were steel sections in oval shape.

A little more visual info



A construction of an oval (using 4 arcs)
A comparison between an oval and an ellipse (both having same size long and short axes)

[Miguel]

Could anybody here explain in a few words and without equations why is an oval tube stiffer than an elliptic tube? Or why would there be such a high difference between an ellipse and a circle? I fear that the answer will be "it just is", but I don't lose anything by asking the engineers.

[bad humor sense on]
After all, and for me, an ellipse and a circle are basically the same. So with a change of coordinates I can transform one into another, and can use symmetry arguments like a politician in order to support my point of view.
[bad humor sense off]
I would have thought that the oval stuff would have some tension in the "tangent points" between the circles, since the second derivative of the curve is probably discontinuous there. Please note I'm well aware all this could be just trash talk by a bored guy.

BTW: Thanks for the pictures, Belatti.