Does anyone know if there exists a standard for testing the chassis torsional rigidity by FEA? Ideally I was wanting the boundary conditions and loads for a simulation?

Cornell University wrote a good paper on this subject. Search for a SAE technical paper on chassis design at sae.org.

Mark
Cal Poly Pomona

Tortional test is set up with the chassis mounted rigid at the rear, no suspension interference, front tie rail is pivoted at the centre,no suspension interference, tortional twist is calculated from the centre of the front tie rail at a measure of 4000nm dialed for four cylinder enines and 6000nm dialed for six or V8s. the twist has to sustain that weight with a maximum of two mm deflection.

Stephen,
Where is that procedure from?

Deejay,
The most accurate way is to restrain and load the frame by the hubs, and include suspension links and "dummy shocks". We've had very good success correlating these FEA results to physical testing.

If you try to clamp the frame at the suspension points, and apply a load at the pickups, without a-arms and pushrods installed, the loading will be different.

Since the search feature on this forum is intermittent at best, I'll post another picture of our torsion test rig:

Denny, this procedure is from the Australian design rules Tech dept. it states that no tortional test is to be taken from any suspension component.So the test would have to be initiated from the outer extremeties of the chassis corners.

Australian design rules for what series? How many 6-cylinder or 8-cylinder FSAE cars have you seen? I've only seen one 8-cylinder...

It just sounds to me like the procedure you listed is for some production-car based series, and the 2mm deflection limit at 4000 N*m (3000 ft-lb) is arbitrary, maybe a "safety" rule.

Can anybody else confirm that this is in the FSAE-A rules?

We do a torsional test every year to validate our Stressed engine setup and i've never seen this procedure before, in the FSAE-A rules or anywhere else.

quote:

Tortional test is set up with the chassis mounted rigid at the rear, no suspension interference, front tie rail is pivoted at the centre,no suspension interference, tortional twist is calculated from the centre of the front tie rail at a measure of 4000nm dialed for four cylinder enines and 6000nm dialed for six or V8s. the twist has to sustain that weight with a maximum of two mm deflection

that measurement technique is certainly not in the FSAE-A rules, I think it is actually a rule for individually constructed road cars, for roadworthy testing. CAMS (australian motorsport governing body) have similar rules for home-built race cars in some series. It sounds like a simplfied method of testing so that anyone can do it without having to construct proper test rigs. Not really useful if you want proper stiffness results though.

"Australian design rules"?

That would be "Australian Design Rules", aka ADR's, aka the bureaucratic bull$#!t that petty-minded, puerile, pencil-heads write to spoil the fun of hard working, honest citizens trying to, ahem, improve their road cars...

Z

Ahem Z,

ADR's are performance criteria, and don't specify design in most cases.

Sure, they are difficult to read initially. but it seems to me that once you're in practice, they're not that hard to summarise.

Sounds like your state gov's are the likely cause of your frustration.

Steven J,

Could you please reference this document?
Is it an administrator's circular?

Personally, It seems illogical to do a torsion test anywhere other than from the hub mounts, (using dumb shocks), and the degrees of freedom of your FEA should be the same as your physical test.

I've never seen the mention of torsion stiffness values in any Australian Design Rule, if you know where it might be, please point it out to me.

Some state regs for ICV's do call out a torsion test for state registration requirements.

Interestingly, The Queensland state gov spec clearly defines hub mount torsional tests.

I've said this before, and I'll say it again. The compliance in the rocker, rocker mounts, shock, and shock mounts will rob you of torsional rigidity more than you might think. (As the loads are being fed through the shocks)

Regards

Frank

Repeating what seems to be a general consensus, I'd say it's nice to have a number for the stiffness of your chassis (so you can compare future and past chassis, if tested in exactly the same manner) but, at the end of the day, the car sees 'non-sprung' torsion everywhere between the contact patches. The most useful comparator has to come from measuring at the uprights.

Frank,

Yeah, I guess I forgot to put these in etc...

(Still trying to figure that stuff out...)

Its not so much the ADR's, its the, err (think of polite word...) queueing in the Rego offices...

Z

As regards the pictured torsion test rig:

You need to add a deflection measurement bar at the rear to deleat the inherant slop in the suspension. Just subtract it's measured deflection from that at the front to get the true twist from end to end.

Where are you attaching the front deflection bar?

The front pivot should be raised to the bottom of the chassis. Putting it that far away will cause the frame to try to move in an arc laterally, and will upset your deflection reading, especially if the rear is constrained.

Elaborate rigs like this are not really necessary to get good and accurate results. All that is needed are dummy shock struts, 2 bars attached to the chassis at the axle c/l's for deflection measurement, a pivot under the frame at the load applied end, and wheels without tires mounted ( the test can be done with tires, but will require longer travel indicaters). The torque load can be applied to just one hub - both ain't necessary.

Richard,
Not pictured are the 4-foot long 1" square tubes we attach laterally to the chassis to serve as deflection "moment arms". We attach one at each end of the car, near the fixture bars shown, and a few in between. That way, we can measure fixture twist vs. frame twist, and get "through the hubs" stiffness and chassis stiffness from the same test run.

The fixture would be better if the pivot were higher, but that's what I came up with in an afternoon last year a couple days before we did the test. It's particularly ugly in the rear, where I had to use scrap hanging around the shop to get enough clearance between the table and the chassis.

Also, we record vertical displacements on the left and right sides of each measurement tube, and average them. Plus, we measure them as we load the beam, and as we remove load from the beam. That way, we can look at the hysteresis / play, and factor that out.