chassis physical torsion test

[mg0199]

Just researching options for doing physical torsion testing for the chassis and basically from what I've read so far come up with basically making plates to lock the two rear hubs to the ground and the front left wheel while resting the chassis at the centre of the front axle on a pivot and loading weights onto the front right hub followed by taking readings from the dial gauge on the front bar and subtracting the vaule on the rear bar dial gauge from that then calculating from there.

The question I have is I have seen some do it with just locking the rear two wheels and front just rest on the pivot and load one hub or lock 3 hubs which way to do some of you guys do it?

[mg0199]

Just researching options for doing physical torsion testing for the chassis and basically from what I've read so far come up with basically making plates to lock the two rear hubs to the ground and the front left wheel while resting the chassis at the centre of the front axle on a pivot and loading weights onto the front right hub followed by taking readings from the dial gauge on the front bar and subtracting the vaule on the rear bar dial gauge from that then calculating from there.

The question I have is I have seen some do it with just locking the rear two wheels and front just rest on the pivot and load one hub or lock 3 hubs which way to do some of you guys do it?

[mg0199]

Also is there any other method of carrying out a physical torsion test on a bare chassis with no suspension arms or anything else on it?

[sbrenaman]

I would recommend doing as much as possible to not over-constrain the wheel hubs. In our fixture, only one corner is fully restrained, but not quite "ridgid." The Neil Roberts fixture in his Think Fast book provided some inspiration, and is what I feel to be down the path of doing it correctly.

Yes to your first question, people have done that. But in our FEA simulations, we showed that doing such a thing can artificially over-stiffen the chassis and give you results that are incorrectly high. Some of the FSAE fixtures we saw artificially increased hub-to-hub stiffness by up to 40%.

As to doing a bare chassis test, it would be better just to do a test with the whole suspension installed. Not worth your time to do bare chassis IMO.

FEA simulation - Make a torsion beam in FEA with a known torsional stiffness. Then mount that in your torsional test fixture. Then load exactly as you would in real life and you should be able to see how much extra stiffness your design adds.

[Canuck Racing]

Well, are you trying to find the torsional stiffness of your chassis or your suspension components?

Think about it.

[RobbyObby]

Well, start from the beginning...
Why is it the torsional case that you want to test? Why not longitudinal bending? Or the axial stiffness? Once you understand this, you can then decide how to load the car for the test.

As Canuck mentioned, are you driving around the track on just the bare frame? Where do the loads from the tire go? HINT: It's through the suspension! However, is it important to know the stiffness component of just the suspension? That's for you to find out.

And lastly, why do want to test the torsional stiffness of your chassis? Because other teams do it? Without actually understanding how it affects the overall vehicle characteristics, testing for an arbitrary number is useless.

Hopefully that gets you thinking along the right path. There are a few good SAE papers out there as well that should come in handy. Look up SAE papers 2002-01-3300 and 2000-01-3554. Those should help you out.