ptuomov

Thank you, there was indeed a units conversion error in my calculations. I was typing that in Europe and decimal points are commas and '000 separators are periods and dog ate my homework etc.

Corrected for unsupported bearings, assuming about 150cm from support to support for the whole assembly:

Variable value units value in si units si units
Modulus of elasticity 205Gpa 2.05E+11Pa
Shaft OD 25mm 0.025m
Shaft ID 0mm 0m
Shaft density 7.87g/cc 7870kg/(m^3)
Shaft length 150cm 1.5m
Resonant rotating frequency 1,336.16rpm 22.27Hz
Resonant firing frequency 334.0rpm 5.57Hz

Resonates and breaks without the support bearings inside the torque tube.

With two bearings, dividing the length by three:

Variable value units value in si units si units
Modulus of elasticity 205Gpa 2.05E+11Pa
Shaft OD 25mm 0.025m
Shaft ID 0mm 0m
Shaft density 7.87g/cc 7870kg/(m^3)
Shaft length 50cm 0.5m
Resonant rotating frequency 12,025.43rpm 200.42Hz
Resonant firing frequency 3,006.4rpm 50.11Hz

Doesn't resonate at relevant rotating frequencies, but may resonate at firing frequency at 3000 rpm.

Variable value units value in si units si units
Modulus of elasticity 205Gpa 2.05E+11Pa
Shaft OD 25mm 0.025m
Shaft ID 0mm 0m
Shaft density 7.87g/cc 7870kg/(m^3)
Shaft length 37.5cm 0.375m
Resonant rotating frequency 21,378.55rpm 356.31Hz
Resonant firing frequency 5,344.6rpm 89.08Hz

With three bearings, doesn't resonate at relevant rotating frequencies, but may resonate at firing frequency of 5300 rpm.

Thanks for catching the error. Anybody else find any remaining errors?

The forum editor is lovely...

ptuomov

Not so, as there is units conversion error in that computation.

Constantine

Interesting stuff.

ptuomov

With the corrected computation, neither two bearing nor three bearing torque tube drive shaft resonates at any relevant rotating rpms. The three-bearing manual tube and the two-bearing automatic tube have some relevant resonances at firing frequencies. The firing frequency resonance is at lower rpms for automatic and higher rpms for manual transmissions because the unsupported section lengths are longer for automatic (two bearings instead of three bearings). This might be related to the noise-vibration-harshness concerns that Porsche may have had and that the vibration dampener inside the tube appears (in my opinion) to be mitigating.

Also note that the formulas that used (initially incorrectly) are just regular beam equations. They aren't the exact rotating shaft equations, but per SAE get you within 80%. So don't pay too much attention to the exact rpms, rather just to the ballpark of those rpms. It could be for example, that the NVH concerns being at different rpms for auto than manual transmissions is explained by the auto having two support bearings and manual three support bearings and the torque tube and shaft resonating at firing frequency:

Also worth noting is that when Porsche moved to two bearings in automatics, they increased the size of the torque tube shaft to 28mm from 25mm. That may have been motivated by the resonances, as the increased outside diameter moves the resonant rpms up (as you can see from the spreadsheet). Perhaps as a pure coincidence, if you plug in 28mm shaft OD and 50cm bearing-to-bearing distance into the spreadsheet, you'll get the 3300 rpm or 55 Hz number that Porsche bullet references for automatics.

Constantine

Still questioning why Porsche went to 28mm drive shafts and only two bearing units. However they did use 25mm drive shafts up through MY 1987 and only two bearing units. Seems a bit haphazard but after seeing other questionable changes Porsche has done to torque tubes in other Porsche models, doesn't really surprise me.

One question about the above computations is how the bearing unit design has an effect on the drive line NVH.

To add further to this whole discussion of controlling frequencies in the 928, one cannot overlook the changes done from the early drive lines to the later versions with their very specific engine and transmission mount designs. I strongly believe that owners who are not using Porsche OE parts for these two areas invite frequencies back into the 928 that Porsche designed out with the design of these parts.

Besides the other problems now surfacing with out of true drive shafts, worn suspension parts, modifications to the 928 by their owners and work being done by technicians who have no clue about the 928.

ptuomov

In my opinion, the torque tube vibration damper can't have an important direct effect on the reliability of the driveline, because it is not actually touching a rotating part in any way. It however can have an important effect on the user experience due to reducing noise, vibration, and harshness. And there can be some meaningful indirect effects: for example, if the manual shifter **** is vibrating like crazy, maybe those vibrations can actually transmit into the transmission controls?

I think that simply hitting an installed torque tube with rubber mallet and recording and analyzing the sound is going to tell you a lot about the resonances. Sounds silly (pun intended), but that's a valid test method. Doing that with different configurations would tell us a lot.

In my opinion, the motor and transmission mounts can be modified to address specific concerns, but whoever makes those modifications better have the engineering exactly right. For example, one might suspect that going to solid motor mounts in the front while maintaining flexible mounts in the rear might make the life easier for the transmission case. Without actually modeling the whole system, that would just be a guess, however. I've got stock mounts all around because I don't understand the system well enough to deviate from factory stock configuration.

ptuomov

If one wants to retain the torque tube vibration damper but is afraid of it possibly moving and hammering the bearings out of place, what's the best way to "pin it" into its installed location without accidentally turning it into a non-damper dead weight?

ptuomov

What new ones are you having made? Torque tube vibration damper rubber springs? Or whole dampers? Or something else?

Constantine

The Super Bearings were specifically re-designed many years ago to address not having to re-use the vibration dampener. There have been many successful rebuilds with no increase of NVHs and in fact the Super Bearimngs reduced NVH problems caused by secondary sources they where not designed to mitigate.

However we are also in the process of re-designing the old vibration dampeners for those who want to use them. Pinning or re-using old style vibration dampeners can lead to problems.

Also have a few vibration dampeners of Porsche's newer design used in later TTs to include the GTSs in good shape.

ptuomov

Someone rebuilding and installing a lot of these torque tubes should start systematically measuring the system natural frequencies along the lines of this article: https://www.reliableplant.com/Read/2...tion-vibration

Landseer

So much science here!

Will know in 2 weeks or so how my setup works.

Again, to anybody reading this, wondering how best to proceed, answer is super bearings from Constantine.

dr bob

I'll throw something in here just for fun. A decade or so ago, give or take, Greg Brown shared some 944 factory race program issues with torque tube life in their cars that had more solid engine and gearbox mounts. The tube itself was failing, and from memory the failure pattern and the visible fracture faces indicated a longer term fatigue failure that points to what Constantine was working towards. Apparently the factory was on to the same thing with the tube dampers. They would reduce the chance of fatigue failure in the pipe, and in the process help the bearings last a lot longer, and resonanct vibrations would not migrate the same way from bearing to bearing through the tube.

Somewhat-related: In the early 1960's Oldsmobile sold a car with aluminum engine in front, a torque tube, and the gearbox in the rear. The driveshaft in the torque tube was purposefully bent, with the center bearing not concentric with the end bearings. Killed a bunch of resonance issues in the driveshaft and the tube. Guess Who was with GM on that project. Yes, the 928 project's very own Guess Who.

ptuomov