It wouldn't matter where you put the chassis. The chassis could sit 100 foot up in the air and LFS could still get it's data from it. So long as points of information (In a real car the bushs that hold the suspension to the chassis) are connected to the chassis and the suspension in one way or another and that that information passed between them is the correct information then the chassis could be racing around on the moon for all LFS cares.

And the backbone method could and would communicate chassis flex in a perfectly adequate way. But, and you'll like this but, if it can be determined that LFS's engine can cope with a more complex chassis structure then so be bit. But advance beyond the backbone and the complexity will rise exponentially. From a system that has 6 moments of rotation to one that will be having to cope with over a hundred and will you be able to taste the difference? I wouldn't have thought the difference would be readily apparent to anything but the most stringent testing.

But anyway, if Lord Scawen was implementing chassis flex anytime soon then I'm sure that the starting point would be a simple model, i.e. the backbone and he too would then work upwards from that if he felt that you needed too.

But as the clever people here are tyring to determine can LFS handle the frequency that chassis flex needs to be sampled at.

Untill thats determined then alot of this talk is moot.

Quote from Hyperactive :

The evil is the details.
(...) And how are you planning to get the suspension joint deformation data from that model? The model has attachments to the tires, not suspension joints...?

Eh ya, didn't we mention this earlier in this thread ?! Obviously it will
be connected to the suspension points, we just simplified it.
That's a, ehem, detail.

Quote from Fonnybone :

Eh ya, didn't we mention this earlier in this thread ?! Obviously it will
be connected to the suspension points, we just simplified it.
That's a, ehem, detail.

Ok then

Just a bump - any progress with your singularity model?

I had a bit of personal life in the last days, so few progress. Though I did create a model and it does everything I demanded in the first post. I obviously didn't have any time to balance the stiffness-values in the different rods to a realistic measure. But I spent some thoughts on dynamic oscillation:
A real chassis has segments with different stiffness against deformation and different oscillation frequencies (because the nose of a car is shaped differently than the middle section, read up on the torsion pendulum for more information). Also the two ends of the chassis are interlinked to the flat ground, thus with each other.
I didn't have time for numerical analysis of the problem, but I think that due to the interlinked ends of the multi-part torsion-pendulum the chassis will stop it's oscillation very quickly.
Notice that the ends of the torsion pendulum are extremely under-dampened, because the chassis oscillates at frequencies like 20-50 Hz while the dampers are around the critical dampening of a 2-8 Hz oscillation. But the interlink of the two ends should cause the different parts of the multi-part torsion-pendulum to oppose each other's movements quickly and thus the chassis should come to rest soon. If that is the case the oscillation is neglectable.
But again, this is just my understanding of the matter from looking at and twisting a lot of my pens ( ) and thus guesswork. I need to get my hands on a full featured simulation software to check that.

Vain