Yes, that is the correct way to install the lock plates.

http://forums.pelicanparts.com/uploa...1516481205.jpg

Yes, I reused them. Clean and degreased, green loctite, torqued to 120nm.

Thanks for the lock plate help!

Regarding the discoloration to the shafts, try cleaning them with a Scotch-Brite Ultra Fine pad, also called #000, or its equivalent. They can be found in the furniture refinishing department of your favorite hardware store. I cut them into smaller squares, which work really well to clean varnish off of shafts, bushing, floating gears, as well as that last bit of sealant from sealing surfaces. The stains in your pictures will probably disappear with a few quick passes of the pad.

I hadn't seen Paul Guard's LSD rebuild video; thank you to whoever posted it. Good info in there. I don't agree 100% with him on all points, but we all have our own ways of doing things, tips, and tricks. The one thing that stuck out was his suggestion that the internals are best measured individually rather than as an assembled pack. Wish I knew why he suggested that, but I believe measuring the assembled pack gives you a less error-prone measure of the pre-load dimension, less chance of accumulated error.

I don't install components piece by piece into the case, as Paul demonstrates. I assemble the "clutch pack" with cup springs (Belleville washers) together (with lubrication), install the thrust washer into the case with a smear of assembly lube (as Paul shows), then lower the case onto the clutch pack.

A couple things the workshop manual and video do not tell you to do but are important. First, measure the unloaded height of your cup springs (Belleville washers). There is no published spec for them, but new ones for 911-series cars like Evan's are roughly 3.5 mm, and for 924S/944 series cars about 3.6 mm. If yours measure much less than those figures, they are fatigued and should be replaced.

http://www.connact.com/~kgross/FAQ/assets/cupspring.jpg

Second is, check the thrust washers for cracks at the pressed-in tab. This was a design flaw on ZF's part: the tab creates a natural stress riser at the end of the cut. I frequently see these come out cracked and broken, and sometimes see both sides of the tab cracked through and the tab floating loose. Guard Transmission corrected this in their excellent LSD by locating the thrust washer with an "ear" that extends radially out into the case rather than a pressed-in tab bent into a small bore in the case.

http://www.connact.com/~kgross/FAQ/a...rustwasher.jpg

Good luck!

Everything I ordered (just steels and clutches) was available within a normal delivery timeframe. I checked Pelican as well, and everything in the list you posted earlier was available except the Bellville washers. Some items were in stock, others special order. Most of the parts for this build have come from our host, but in the case of the LSD parts it was such a small order I went local.

It will be interesting to see how the diff performs.

Thanks!

The shafts cleaned up fine with scotchbrite. I got the whole gear stack back together a page or two back.

I posted the link to Paul's LSD rebuild, it was a big help and I was able to measure all components and decide what to replace (with good help from Matt, Jon, Peter Z, and this board.)

I checked the thrust washers; fortunately no cracks.

Hopefully I'm good to go!

Backlash help please...
 

With the diff all back together it was time to check preload again and then backlash. As always, I'm following Peter Z's guide, so I won't detail that here.

From the factory the speedo side shims were 2.8+.25 =3.05mm. The ring gear side was 3.2mm. Preload was a bit low when I tested prior to the diff rebuild, so this time I changed the ring gear side out for a 3.3mm shim and kept the speedo side as it was. Preload measured between 36 and 40 lb-in, low to middle for the new FAG carrier bearings I've installed, but well within the 30-56 lb-in spec.

So on to backlash: I've got a dial micrometer that reads in 0.01mm. I've got to get it set up a bit better, but if the measurements I've taken tonight are correct my backlash is way out at about 0.85mm.

I reused the original pinion depth shim (0.30mm.) The face of the final drive case was machined flat by the shop that installed the Wevo insert, possible material removed should be less then 0.1mm according to the shop. So we have an unknown there.

So besides verifying that I'm taking good measurements, it seems some shims will need to be changed around. I've got extra carrier shims at 2.8, 2.9. 3.0, 3.1, 3.2, and 0.25 not including the installed ones I detailed above. I also have extra pinion depth shims at 0.10 and 0.15mm, in addition to the .30 that is currently installed. I have no way to measure pinion depth, only preload and backlash.

All tips greatly appreciated as I'd rather not go blindly changing shims 10 times.

Rebuilding the lsd shouldn’t change any settings. If it was right before don’t change it.

Much more has been done then the LSD though. New carrier bearings, and case machining in particular. So if backlash is way out of spec, that can't be something to ignore? I have no idea what the values were before teardown.

If the machining for the Wevo insert did not change the installed depth of the pinion shaft bearing, then you should not have to adjust the pinion depth shims. I have never worked with a case that's had this done to it and cannot say. The 0.1 mm machined sounds like a small number, but given that Porsche indicates a 0.05 mm tolerance in shimming, it is significant with respect to correct depth. You may want to consider finding someone local who has the tool and can measure depth for you, if indeed the machining changed bearing position.

Because you changed the differential carrier bearings, it is not surprising that that has changed the shims needed for the diff carrier bearings. My experience has been that they always change, at least that is, following the factory instructions you'll find different shims needed to reestablish preload and backlash (free play). Whether those changes are necessary, we could argue about. Sometimes they change a lot, though, and I can't say for sure it's because of wear, bearing tolerance, etc.

When you say "Preload was a bit low when I tested prior to the diff rebuild", was that with the old differential carrier bearings? If so, not surprising: break-in of the bearings lowers turning torque a whole bunch, and the factory manual will tell you the ranges.

By the way, you can use the 996 series shims to set the diff carrier bearings. They are the same dimensions as the 915 shims and available in many more sizes. Last time I checked, many of the 915 shims were NLA from Porsche.

Good luck!

Evan, are you certain that your pinion and differential are both locked effectively, and that your measurement is not being affected by additional movements?
A backlash measurement of 0.85mm is enormous.
Maybe you should tell us, or show us, how you've prepared the transmission for these backlash measurements.

Since you didn’t measure the baseline before teardown you have no reference point. With the case work you must redo the pinion depth before you can do backlash.

Thanks for all the input everyone!

So I clearly needed to step up my game. Tonight with a fresh set of eyes and a new tool I did a much better job of locking the pinion shaft, and set up my dial indicator correctly. I also preloaded it about halfway and zeroed it properly. The good news is that my backlash reading is much lower, just out of spec at about 21-22mm. EDIT:
I meant .21-.22mm ;)

Just for fun, after this I took the diff out again and loaded the ring and pinion with white grease to see what the contact pattern looked like (I have to order some prussian blue from Amazon, it doesn't exist here.) Printing test was of course done with the cover fully torqued. Lots of photos below, and a video. Thoughts please...

A summary of where we are:

-Original pinion depth shim is installed, 0.30mm
-One piece retainer bolted down
-Pinion nut torqued to 181 lb-ft
-Main shaft locked
-Pinion shaft locked
-Side cover fully bolted down, torqued to 17 lb-ft
-Backlash just over 21mm in one position, 22mm at 90 degrees from that. EDIT:
I meant .21-.22mm here too ;)


<iframe width="560" height="315" src="https://www.youtube.com/embed/hIbFV6kAbNE" frameborder="0" allow="autoplay; encrypted-media" allowfullscreen></iframe>


http://forums.pelicanparts.com/uploa...1516741443.JPG
http://forums.pelicanparts.com/uploa...1516741443.JPG
http://forums.pelicanparts.com/uploa...1516741443.JPG
http://forums.pelicanparts.com/uploa...1516741443.JPG
http://forums.pelicanparts.com/uploa...1516745950.JPG
http://forums.pelicanparts.com/uploa...1516741443.JPG
http://forums.pelicanparts.com/uploa...1516741443.JPG
http://forums.pelicanparts.com/uploa...1516741443.JPG

You meant to say that the backlash reading is 0.21 - 0.22mm ;-)

Evan, did you lock the axle flange to the differential with an appropriate disc, or are you relying only on the differential clutches?
If you're not locking the flange to the diff housing, you might still be getting some additional movements.

Hi Evan, nice job on the custom tools! Want to sell me one of the pinion shaft lock tools? I've been looking for the factory tool for years, long since unavailable. Your tool looks better than what I have cobbled together for my own use!

Regarding using lithium grease to read the pattern on the ring or pinion teeth, I don't think you will have much luck with that. A somewhat better approach is to use a compound designed for this, for example Permatex Prussian Blue. It will at least give you a clear pattern.

But my experience has been that even with the prussian blue, you won't get useful information. The problem is that the factory (Getrag or VW Kassel) tested the pinions under load, looking for quietest running (this is my understanding), and under load the pattern moves. In the field, we cannot apply drive and load to the transmission so that the pinion-to-ring pressure center moves to actual operating conditions. So, the factory does this for us, tells us the correction needed for a particular finished r&p set, and we humbly duplicate that correction using our VW385 tool set.

I could be completely wrong... would enjoy hearing Matt, Jon, and others' opinions. Cheers,

Evan, regarding the contact pattern that you've made with grease, I really can't see anything clearly enough to offer an opinion.
I agree with Kevin that it's more precise to duplicate the factory settings. Porsche never discussed or recommended the use of contact patterns.

Here's the factory tool for locking a 915 pinion. I'm not keen on pushing into the rear of the pinion shaft, but it's probably OK.

http://forums.pelicanparts.com/uploa...1516770200.jpg

http://forums.pelicanparts.com/uploa...1516771655.jpg

Oops! Yeah, forgot that ever-important decimal point! Noted above.

Back on page 5 of this thread there was a bunch of head scratching and discussion around this. I tried to confirm that the washer "tool' wasn't necessary with the LSD, as the flange is tight to the carrier anyway and spins the entire diff. And in any event, the washer is exactly the same dimension as the output flange-- the washer would just be duplicating this so I can't see that it would have any effect. Matt also mentioned that I would need to overcome the LSD preload in order to have movement. My LSD breakaway is 47 lb-ft, so it seems unlikely. Feel free to shed more light on this...


Thanks-- took me 10min to make out of a couple pieces of stainless L-channel! Easy.



Prussian blue-- can't get it here. If its not of any use I'll just skip it. I'm trying to see if someone has a VW385 tool I can use, or I'll have to send off the box somewhere expensive.


Thanks Jon, I'm familiar with the factory tool and debated building one like it, but Peter Z clamps the 5-R slider in this manner in his tutorial, so I took it as OK. The pinion shaft doesn't budge with it in place. Here is Pete's:

http://forums.pelicanparts.com/uploa...1516785583.jpg


So now the question is, what next? Swap some shims? Measuring pinion depth probably...

Also, does my technique for measuring backlash look OK? How much force should I exert on the flange, rocking it back and forth?



evan

915 Rebuild
 

You can make a reproduction P258 tool to set your pinion depth by modifying a VW289d. The VW 289d is used to set the pinion depth on VW transmissions and is not hard to find on EBay or other VW sites. The P258 appears to have been made from the same casting as the VW 289d; it is just machined differently.
To modify the 289d, you need to cut down the spigots to 50mm and a width of about 120mm (same as the 915 differential carrier). Then modify the setting block to have two steps: 63.20 and 66.30mm. For the 915 transmission, you need the 66.30mm step but you might as well make a step so it can be used to set up mag case 901 and 914 transmissions.
Attached are some pictures of an original P258, a VW289d and a modified 289d. Note that the P258 was originally made to work on the aluminum case 901 transmissions that used differential bearings with an ID of about 45.xxx (inch size bearings) and a differential width of about 109mm. When they introduced the mag case transmission, they supplied bushings so the P258 could still be used.
http://forums.pelicanparts.com/uploa...1516788147.jpg
http://forums.pelicanparts.com/uploa...1516788257.jpg
http://forums.pelicanparts.com/uploa...1516788277.jpg
http://forums.pelicanparts.com/uploa...1516788327.jpg
http://forums.pelicanparts.com/uploa...1516788350.jpg

Evan, the washer tool will LOCK the flange to the differential housing. Zero movement between the two.
I don't have to scratch my head about this.

If you're certain that there's zero movement now, not even 0.02mm or 0.05mm of movement, then proceed as you're doing.
But then why are you still getting high readings? Everything you've done to this point should have decreased the backlash. You added 0.10mm in shim to the bearing on the ring gear side.
How did your backlash measurement suddenly go from 0.85mm to 0.22mm? Because you eliminated some of the extra movements in your set-up.

There's a significant difference between measuring the bearing preload and measuring the ring gear backlash.
Some initial movement inside the diff, with the flange rocking or with the clutch splines or tabs, will not affect the bearing drag reading. It will affect the backlash reading.

But if you're certain that you now have no extra movements in your measurements, then proceed as you're doing.

Great information! Thank you!


Understood. The head scratching is all mine BTW. I had my local machine shop waterjet me a washer tool in the correct size today so I can do this right. I'll have it tomorrow and will put it in place, then test again. I hope you haven't experienced any of my responses as disrespectful. I'll be the first to admit I'm a complete rookie. Without help from experienced guys like you, I'd really be out to sea.

Does it look at though my overall method for testing backlash is OK? I have no idea how much force to exert on the flange when testing. If I really hang on it, I can get higher readings, which is probably a good indication that there is some movement somewhere.


evan

I'm back! Been sick with Bakersfield's winter head cold; bad this year! Evan, once again I'm sorry for any confusion I might have caused with the Carrera LSD spec sheets. My manuals have been reorganized (I still don't have an instruction sheet in my SC book!) , and I've said it before - having the part in hand sure beats trying to help from xx,xxx miles away! It looks like you're making good progress, just keep that washer in place to lock the flange to the diff! Carry on!