Aluminum Diff Cover on 911 Gearbox

[jluetjen]

I'm in the process of installing an aluminum CMS differential cover on a magnesium 911 gearbox that I have. The differential spacers were originally set-up (with the mag diff cover as follows):

S1 (ring-gear side of diff) = 2.81 (1 thick and one thin spacer)
S2 = 2.72 (1 thick and 1 thin spacer)

At that point the diff preload felt fine (I never actually measured it).

So I started from there when I installed the aluminum cover. This diff cover uses the heavy duty Turbo diff-bearing. After assembling things up (note that the gears and associated shafts are not installed), the diff preload was so high that I could hardly turn it. So I took everything apart and removed the thin (0.15) spacer from the S1 stack. My logic is that if the diff assembly was centered as originally set-up, and nothing has changed in the relationship of the diff to the far side of the case -- which is a fixed reference, that it would make sense that the only thing that has changed is on the ring-gear side.

You can visualize it as follows: (The red items are the only things which changed from the original set-up)

Transaxle case>> bearing >> S2 (2.72) >> Diff << S1 (2.81) << Turbo bearing << CMS Diff Cover

By removing the 0.15 spacer out of the S1 stack, I reduced the preload to about 40 inch-pounds, or 2X the maximum. So it looks like I'm going to need to take another ~ 0.15 out of the S1 stack, which will mean getting a smaller "thick" spacer, or taking about 150 um off of the existing spacer. (If I end up doing this, I'll increment up to the value while checking the preload).

My questions are as follow:

1) The factory shop manual suggest that you need to change both S1 and S2 together. Is that still needed in this case?

My fear is that if I do this, I'll end up changing the relationship of the diff to the pinion gear, which will open up a whole new can of worms in trying to set that relationship.

2) Is it normal for the preload to change so drastically when installing an aluminum diff cover?

The 911 gear box uses an O-ring to seal the diff cover, so there doesn't appear to be any gasket under the cover that I can change in order to bring the cover back out a little bit. That would seam to be the easiest fix. I have a left-over diff cover gasket from an aluminum 901 transaxle, but that turns out to be a smaller size then the magnesium 911 transaxle.

I'd appreciate people's thoughts and inputs if they have done a similar upgrade in the past.

[sjf911]

I changed the carrier bearing in the case at the same time I did mine so I started from scratch finding the optimal combined stack thickness for preload then playing with S1/S2 combinations to find the right R/P backlash.

[jluetjen]

OK. I finally got the preload dialed in pretty close. I actually had two transaxles with this update to work with. The secret was to use the nominally 2.8 mm spacers as normally used on the S1 side on the S2 side. Then use the 0.25 mm thin spacers as needed. In my case I ended up using just the 2.8 mm spacers for each transaxle. I tried to use a 2.7 + 0.25 mm spacer in one case but the preload was still too high. So fairly straight forward -- but a lot of disassembly and reassembly of the diff bearings and side-covers to get to this point.

[Geary]

Always a good idea to measure turning torque prior to disassembly. Your used bearings might have registered considerably lower turning torque with the original cover.

You're definitely correct in not wanting to change the S2 shim.

Smart thing would be to check backlash. If it's at least .006", you're safe.

Final check is with bluing. Peace of mind with a nice pattern.

[jluetjen]

Thanks Geary. Yes, in hindsight I should have checked the value prior to disassembly.

I've been struggling to think of how to check the back-lash without a bunch of special tools. I've got dial gauges and stands. Is there a way that you could share?

As far as checking the pattern with blue die, based on input from this board (maybe yours) I did this with my current transmission when I built it a few years back and it worked out great. That is what I was going to do after getting the shafts assembled.

[Geary]

I'd forgotten you've got a new bearing in the mix .. You're probably right on with your shimming.

The VW 521/4 tool affixes within the output flange bore. So long as you realize the importance of this function, you can fabricate something from a long bolt or 10mm thread stock combined with HD washer .. something that will rotate solidly with the diff body.

Lock your pinion shaft from movement, of course ..

Measurement takes place at approximately the outer edge of the ring gear, which is about 3/4" inside the outer edge of your new side cover (I've loaned out my manuals, or I'd give you precise radius). I said .006" as probably being OK, but .007" is the backlash I actually shoot for.

Bluing check will tell all. Near full pattern on drive side, with centered pattern on coast side.

I think you'll sail through all of these checks with flying colors. You've only altered S1.