E90 M3 vs. E92M3

[alpineweissM3]

Originally Posted by lucid

Thanks for the data. What axis are these measurments obtain around exactly? Also, do you know what exactly is behind the seats in the sedan?

I see Sedan Clan pulled the info for you...he's quick!

Anyway, I have an '01 325i that I am fanatical about. That being said, I do know a thing or two about that chassis. My understanding is that there is a pretty thick bulkhead that spans the width of the coupe / sedan w/o folding rear seats. This helps to tighten everything up just as a front shock tower brace would do.

I am shocked that the e92 is more rigid than the e90. Guess I have alot of learnin' to do over here!

[lucid]

Originally Posted by alpineweissM3

My understanding is that there is a pretty thick bulkhead that spans the width of the coupe / sedan w/o folding rear seats. This helps to tighten everything up just as a front shock tower brace would do.

If there is a thick relatively flat metal panel that is running through, that doesn't make sense to me from a structures and optimization perspective. It would be much more efficient to run thinner individual members across with the optimal cross section for the torsinal stiffness one is trying to increase. Regardless, it is clear that there is something structural running across behind the seats, and I did not expect it to make such a difference. It would still be useful to find out through which axis the data were taken though.

[Sedan_Clan]

Originally Posted by lucid

If there is a thick relatively flat metal panel that is running through, that doesn't make sense to me from a structures and optimization perspective. It would be much more efficient to run thinner individual members across with the optimal cross section for the torsinal stiffness one is trying to increase. Regardless, it is clear that there is something structural running across behind the seats, and I did not expect it to make such a difference. It would still be useful to find out through which axis the data were taken though.

I may just pull out my back seat and take a few photographs.

[swamp2]

Originally Posted by lucid

If there is a thick relatively flat metal panel that is running through, that doesn't make sense to me from a structures and optimization perspective. It would be much more efficient to run thinner individual members across with the optimal cross section for the torsinal stiffness one is trying to increase. Regardless, it is clear that there is something structural running across behind the seats, and I did not expect it to make such a difference. It would still be useful to find out through which axis the data were taken though.

Pure speculation on my behalf:

There are not too many possibilities for torsional testing of a chassis. I would bet that they either mount the chassis on the suspension mounting points with some rigid arms assembled so that they can not pivot relative to one another or simply use the shock tower mounts for the front of the car. Then opposing forces are probably used so it is a pure torque and no net force is applied. The angular measurement can then be across the front or rear shock or suspension mounting points.

You may also be able to use the 1st torsional mode shape frequency to get the stiffness. If you used a free-free mode shape the mounting issues may be totally removed from the picture.

[lucid]

Originally Posted by swamp2

I would bet that they either mount the chassis on the suspension mounting points with some rigid arms assembled so that they can not pivot relative to one another or simply use the shock tower mounts for the front of the car. Then opposing forces are probably used so it is a pure torque and no net force is applied. The angular measurement can then be across the front or rear shock or suspension mounting points.

So you are basically saying they would somehow fixate the chassis at the suspension mounts which would transfer the suspension loads. Makes sense. So at that point, the car is clamped down at 4 corners, potentially with more than 4 attachment points. I am not sure about the force application. If you were to simplify it to 4 mounting points in the corners, which points and directions are the forces being applied exactly? I can see different permutations resulting in different axis.

[Sedan_Clan]

Good discussion guys (..in the latter part of this thread). I am so curious to see how the sedan performs.

[swamp2]

Originally Posted by lucid

So you are basically saying they would somehow fixate the chassis at the suspension mounts which would transfer the suspension loads. Makes sense. So at that point, the car is clamped down at 4 corners, potentially with more than 4 attachment points. I am not sure about the force application. If you were to simplify it to 4 mounting points in the corners, which points and directions are the forces being applied exactly? I can see different permutations resulting in different axis.

More speculation...

For example you could do it this way. Have the rear suspension mounting points totally rigidly locked down. Then have the front shock towers fixed together with a U shaped rigid tool as well. The force actuators could then be one upward near one front hub location and the other downward near the other front hub. If the test fixture for the front was made very stiff but with some pivots instead you probably could insure that the two forces were very close to vertical with equivalent application locations right at the shock tower mounting points.

I'd love to see the actual test set up, again might be a 4 wheel suspension dyno or modal testing as well.

[lucid]

Originally Posted by swamp2

More speculation...

For example you could do it this way. Have the rear suspension mounting points totally rigidly locked down. Then have the front shock towers fixed together with a U shaped rigid tool as well. The force actuators could then be one upward near one front hub location and the other downward near the other front hub. If the test fixture for the front was made very stiff but with some pivots instead you probably could insure that the two forces were very close to vertical with equivalent application locations right at the shock tower mounting points.

I'd love to see the actual test set up, again might be a 4 wheel suspension dyno or modal testing as well.

How about applying equal but opposing forces on the front (FR & FL) and the rear (RR & RL)? Or some other scenario like grounding FR and RL, and applying equal but opposing forces on FL and RR.

[swamp2]

Originally Posted by lucid

How about applying equal but opposing forces on the front (FR & FL) and the rear (RR & RL)? Or some other scenario like grounding FR and RL, and applying equal but opposing forces on FL and RR.

Fun, "armchair engineering"... Well at least we are/were both engineers. Many possibilities here, but the fewer places to keep track of a force input the easier and less expensive.

[aequitas330]

I like the looks of the sedan a lil better...

[lucid]

Originally Posted by swamp2

Fun, "armchair engineering"... Well at least we are/were both engineers. Many possibilities here, but the fewer places to keep track of a force input the easier and less expensive.

Ah, that brings up to question of which test best simulates driving scenarios that affect performance the most so that torsional stiffness is measured around axis that count more than others...