The whole torque tube issue is new to me. I’ve literally just thought about it for a couple of hours over this weekend. All help would be greatly appreciated in making me understand the issues related to it.

There seems to be two separate issues here. First, the vibration of parts that rotate. Second, the vibration of parts that don't rotate.

My instinct is that eliminating dangerous resonances of rotating parts is difficult but important for reliability. Eliminating resonances of parts that don't rotate is much easier and mostly just about eliminating annoying noise.

By my current understanding, the drive shaft suppport bearings deal with the first issue, as they support the rotating shaft and effectively break it into many shorter shafts that resonate at higher frequencies than the whole unsupported shaft would. If furthermore the bearings are placed such that none of the short segments resonates at the same frequencies, then we’ve mitigated any potential resonance problems.

The cast iron damper deals with the second issue, as it only touches the non-rotating tube. It’s just a weight attached to a rubber spring. A similar device could be attached to the outside of the tube, for the same effect. A similar device could also be used in an exhaust pipe or a muffler to get rid of drone.

For the vibration dampers to impact the driveline reliability and the rotating shaft itself, there would have to be some vibrations that exist at or get transferred to the clutch housing or transmission case and that are so severe that they shake those whole heavy components. This I find unlikely, but then again I’ve only though about this over a weekend.

Now, to even more speculative views: My guess is that the main determinants of the resonant frequency (or really frequencies) of the tube are the length and diameter of the tube. My guess is that Porsche cast and built dampers that were intended to match each tube length and diameter they used. My guess is that those dampers never come out resonating (or really dampening) exactly at the specified frequency, so they get measured empirically after having been built. Again, guessing, when the car was built, the damper was matched with the tubes using the empirically measured damper-specific frequencies and model average tube frequencies, on a best efforts basis. The damper placement by my guess is relatively unimportant as long as it’s attached to the tube by the rubber spring.

If these guesses are correct, retaining the dampers is safe, provided that the rubber is new and the damper can’t move around. I don’t know if the old rubber parts can be replaced on those dampers. The potential benefit from retaining the dampers is a quieter car at some rpms. There is no meaningful reliability benefit by my guess from the dampers, but who wants to drive a car that booms or drones at a cruise rpm?

The above is all about the stock configuration. The next question is what happens with three super bearings inside the tube. Has anyone measured the sound spectrum of the torque tube at different rpms with and without the vibration damper installed, using those super bearings? Even just without the damper would be interesting, because if there’s no strong audible resonances at any relevant rpms, there’s nothing to dampen. I’ve printed some measurement articles, and since this is not a rotating part one should maybe just figure out the resonance with a sound generator and microphone?

Thinking these issues out loud at this point...

Ok, steep learning curve. The first question I needed to clarify for myself is why does the torque tube even have bearings? Talk about starting from basics. In any case, suppose that we have a steel drive shaft inside a torque tube that doesn't have any support bearings. The SAE design guideline is to compute the elementary beam theory equation resonant frequency for the shaft and then stay under 80% of the resulting frequency in exitation. There's two kinds of excitations in this case, I think, first just the pure rotation and the firing frequency. The formula for elementary beam theory resonant frequency is pretty simple, that is, frequency = (pi/8)*sqrt(E*(D^2 + d^2)/(rho*L^4)). E is Young's moduls of elasticity for the steel used, D is the outside diamtere of the shaft, d is the inside diameter of a hollow shaft (zero for solid shaft), rho is density of steel used, and L is the shaft length. Easiest with SI units. I did the computation below for an unsupported shaft using the 300M material values, not too different from other good steels:

Variable value units value in si units si units
Modulus of elasticity 205 Gpa 205000000000 Pa
Shaft OD 25 mm 0,025 m
Shaft ID 0 mm 0 m
Shaft density 7,87 g/cc 78,7 kg/(m^2)
Shaft length 37,5 cm 0,375 m
Resonant rotating frequency 213 785,46 rpm 3 563,09 Hz
Resonant firing frequency 53 446,4 rpm 890,77 Hz

I'm sure that the terrible, horrible, bad, no good, Rennlist editor will garble the table above, but that's not news, not even to me...If you care about the values you have to parse it out.

The punch line is that the unsupported shaft would only resonate at 13400 rpm under continuous drive. With firing frequency exitation, however, it would go critical and fail at 3340 rpm. I'm ballparking the shaft length to 150 cm here, which should be measured from the transmission bearing point to the clutch bearing point. So driving a lot of V8 torque thru a 928 drive shaft without torque tube bearings is a no go.

With equally spaced two bearings (50cm segment lengths), the resonant frequency increase to 30000 rpm. With equally spaced three bearings, the criticl frequency increases to over 50000 rpm. Well above the guidelines even if one bearing fails its support job completely.

So my conclusion is that those bearings inside the torque tube are very necessary and they do solve an actual critical problem. This may not be news to anybody else but me!

Gun drilling the drive shaft would further increase the resonant frequency a little bit without hurting the torsional load capability measurably, but with three bearings that's not worth the cost by these computations.

[THERE'S A UNITS CONVERSION ERROR IN THE ABOVE COMPUTATION, CORRECTED IN A LATER POST.]