BMW 323is LSD 2-disc To 3-Disc Upgrade

I have a '98 323is that I bought a few years ago and converted to race duty. I'm currently running in NASA's ST6 class even though the car is technically illegal due to displacement. I made an agreement with the regional racers to let me run with a restrictor (55mm). I have no engine mods other than a full exhaust with only a resonator. The ECU is was modified to get rid of security requirements, increase the rev limit to 7200 and to ignore the absence of wheel speed sensors (ABS removed).

It came with an open diff which was a huge liability. So I bought a used LSD off eBay with the 3:15 ratio. I installed it over the winter and did one half of a track day in Feb. and then ran last weekend at VIR. Either the LSD was on its last legs or I burned it up because I was still dealing with the one-tire-fire in every hard corner. 

So, I ordered a set of clutch discs for it. The stock LSD comes with 2 friction discs and two dog-ear plates. My understanding is that these provide 25% lock-up. According to the internet, one can replace the 4mm spacer in the LSD with an extra friction disc and dog-ear plate. The discs I ordered are from Racing Diffs. With the 3 discs, it's supposed to increase to 40% lock-up.

This will be my first time tearing into an LSD of any kind. There are plenty of how-to videos on this so I won't repeat that. Just going to document how mine comes along.

After draining the case of it's perfectly new Royal Purple diff fluid <sigh> I pulled the rear cover off and checked the magnetic pick up for schmoo. There is certainly a little bit. I'm guessing it's friction material off the clutch plates. It's very fine. Almost looks like anti-seize paste.

The pinion gear looks good to me other than the gray discoloration at leading edge of the gear teeth. Not sure what that would be or if that's normal.

The ring gear looks good too. No discoloration. Every tooth intact.

More to come

Here are the new discs. They're vented. 

The internet said these bolts strip easily. I can't really understand why they're allen head when they require more torque than the heads can handle. Seems to me a standard bolt head would work just fine. I suppose I got lucky. Only one of mine stripped out. 

I hosed the entire surface down with brake cleaner. I'll let it sit until tomorrow so it all evaporates. Then I'll weld this nut onto the bolt head. I'll need to find a way to shield the bearing and things from weld spatter. I have some sheet aluminum laying around. I should be able to use that.

The forum seems to be making its own decisions on how to orient my pictures. Thanks for that. 

Decided I couldn't wait until tomorrow...

Welded a 3/4" nut on top of the stripped allen head bolt. That made it easy to get out. Used flux core because it deals with impurities better (like coatings on nuts and bolts).

Disassembled the diff, laying out everything in the order and orientation that it came out.

I left the final spacers at the bottom of the diff in place as they're clean and there's no need to remove them since everything else goes on top of them.

Soaked the new plates in oil while I was doing the disassembly.

This is the spacer that gets replaced. It's 4mm thick and is replace by a friction plate and dog ear plate, each 2mm thick.

But, as always, things don't go exactly as planned. The friction plates are stamped or cut very precisely but then they're coated. The coating is not precise. So, out of the box, they don't mesh with the side gear splines. I emailed my supplier and he called me right back (after hours I might add). He said I'll just need to clean the teeth up with a wire brush or, worst case, will need to file them a little. So I'll need to clean off the oil and go after them in the morning.

Great work so far. Thanks for documenting. I'm definitely bookmarking this one.

This is really great info! I just bought a 3.38 LSD as an upgrade for my 95 m3, but it came from a 200k mile car, so it’s probably due for a rebuild. This will be really helpful

This makes me want to order parts. Great info and thread!

Thanks, guys!

Here are some links to some very thorough DIY articles and videos:

https://www.eeuroparts.com/blog/10767/upgrade-your-stock-bmw-e36-lsd-with-porsche-parts/

https://www.youtube.com/watch?v=pVG-RUfXo4Q

https://www.youtube.com/watch?v=gHD3Z54sV_U

Here's how I tackled the ill-fitting friction discs...

I used the old dog-ear plates and sandwiched each of the new friction discs in the vice like this. I did this to avoid contact with any more of the friction surface than absolutely necessary. Maybe overkill but better safe than sorry.

Then I used a square needle file (triangular would work as well) to file the tips on each tooth and the valleys between them. Mostly, I just wanted to take off the friction material. Each disc had one tooth that was pointed and need to be filed down to a flat profile like the rest of them. It's tedious but this process took me about a half hour to clean up all three discs. I should have shown it but I also cleaned off a side gear and had it close by so I could check the fit of each disc. The discs should slide onto it with no force.

Then I went after it with a wire brush. This is why I sandwiched the new friction plates. I could be as aggressive as I wanted without fear of knocking off too much of the friction material.

She's all back together. I'm just waiting on new bolts, which should be here next week.

I’ve officially added the 3 clutch upgrade to my parts/project lists. What started off as a diff fluid change has now morphed into a complete rear end rebuild with new bushings, reinforcement plates, and a completely new diff. Scope creep is not my friend lol

In reply to Turbine :

Awesome! Start a thread on it. I'd like to see what all is involved.

I was just looking through the Racing Diffs online eBay store and saw that they have what they call "Adjustable Pressure Plates". Is that technically the correct name for the part? Dunno. But it looks interesting. I had read about shops that change "ramp angles" and I knew it had something to do with how the lock-up occurs. But seeing the pics of the product cemented (I think) how it works.

BTW, I promise I'm not a schill for Racing Diffs. I just went with their stuff for my rebuild. I'm withholding judgment until I run a few events on the new friction plates.

Anyway, what I now understand (I hope) is that the cups (Adjustable Pressure Plates) that contain the spider gears and side gears are what provide the clamping force on the friction plates. So when you accelerate or decelerate, the cups are forced apart by the spider gears. The friction plates are located on either side of this assembly. If you look at the pic in my previous post of all the parts laid out on the table, these Pressure Plates, the spider gears and side gears are on the far left side.

I'm guessing that the ramp angle determines how fast and with how much force the Pressure Plates are forced apart. Does a lower angle take longer to lock up but provide more pressure? Does a difference in angle on one side of the notch mean a difference in lock-up under acceleration and deceleration? I think yes, but I'm not 100% sure. If we have any experts here that would like to chime in, please feel free!

Now, the kit that I'm using (3-disc) will obviously be asymmetrical, with 2 discs on one side and one on the other. So it will be interesting to see how that affects lock-up in left vs. right turns. I believe the 2 discs will be on the right side.

Got the remaining bolts in the mail today. Started all the bolts by hand. Tightened them in a star pattern as best I could. Still managed to strip the threads on one bolt. Not sure if it was started wrong. They were all surprisingly difficult to turn, even before the head contacted the surface. They felt almost like self-locking threads. The one bolt stripped when I was on the final pass. It's staying in there for now. I've got it marked. I'm going to run the next event as is then take it out again. 

I think next time I tear into this diff, I'm going to replace those allen head bolts with these:

https://www.belmetric.com/m8x125-flange-bolts-c-2_1208_1221/bfd8x16ylw-flange-bolt-109-yellow-zinc-p-7182.html

I'm sure there's a reason BMW spec'd soft bolts I'm just not buying into it. There will be a tiny weight difference.

Obviously, I'm not recommending this. I'm willing to take one for the team though.  

This diff will come back out after my race at Summit Point the weekend of 4/13. I'm not expecting that stripped bolt to last all season! I'm not sure what I'll do if the threads in the housing are stripped. That would just suck. If that's the case, I'll take it to a machine shop and get a helicoil put in there. I guess the other option is to weld the hole and redrill it. And the other other option is to get another LSD.

Reinstalled the diff last night. Swapped new front rotors and pads all around. Went for a careful drive back and forth this afternoon to bed in brakes and make sure nothing fell off. Seems good. After putting her back in the garage, I took the dogs for a walk and noticed a state trooper cruising by with lights flashing. Looking for me? I think I need a quiet muffler I can easily swap before I do one of these runs.

I wanted to update this with results from Summit Point after the rebuild. It turns out, my one-tire-fire problems had little to do with the diff. My first session out was exactly like the previous race weekend on the junkyard LSD and all the races from last season on the open diff. The inside tire was spinning in every corner.

Short story: After the second session on Saturday, I unscrewed the shock body to drop it 1" and that fixed about 90% of the wheel spin. I think I can drop it slightly more and get closer to 100%. There are other suspension adjustments that need to be made as well.

Back story: My car has been "set up" by me with no real input from any professionals. So that's problem one. I did have it corner weighted and the cross weight was, by some miracle, damn near perfect. Don't remember exactly what it was but the gist of it was that it didn't need to be changed.

After the second session on Saturday, my assessment of the problem was that the inside tire wasn't making enough contact with the ground. What should have been a big clue for me was that, anytime I jacked up the car, the suspension barely drooped. Seriously. Like maybe 1/2", if that. I had certainly noticed it but I was a little slow on the uptake as to the wider ramifications.

Okay, so on Saturday between qualifying and the race, I studied on the problem. I jacked up the rear, removed the tires and took a good look at the rear shocks. I clearly had not paid enough attention to how they work. Like clue #2 should have been that, when unbolting and removing the bottom bolt from the shock, the suspension drooped at least two or three inches. So, I set about unscrewing the shock body to see just how much I could drop it. I got it about 3" longer and still had more. I decided that was enough information so I put it back to where it was (6" of thread showing) and then dropped it 1" more.

In the afternoon race, it was performing much better. It was hooking up in all but the tightest of situations or when coming off curbs. 

Over the next couple of weekends, I plan on removing the diff again and dealing with the stripped bolt. Plus, I'm going to go through the suspension again and see if I can make some improvements in ride height and camber.

As I stated a couple posts ago, I stripped one of the bolts when I was putting the diff back together. I decided to run it anyway and it held up fine. I pulled it out over the weekend and disassembled it again. I was able to get the stripped bolt out after some fiddling. All of the bolts had thread galling. There are even pieces of thread that separated completely. I think it was the second one down that was stripped.

The holes don't look too good. I can't understand how that much crud works its way down into the threads. I wish I had paid attention to their condition before I reassembled it last time. This time, I'm going to clean them out thoroughly. I need to run out and pick up some brake cleaner. I'm also considering replacing these OEM bolts with harder hex head bolts of the same thread size. I have another set of the OEM bolts but I'm not thrilled with how soft they are. I ordered a bottoming tap in case I need to clean up the threads. And if the one hole is too far gone, I guess I'll have to take it to a machine shop to get it repaired.

Okay! I got the bolt holes all cleaned out. I then tried threading a new bolt into the holes. I couldn't get more than a tiny bit of thread to catch so I decided to bite the bullet and run a bottoming tap into each hole. There was a lot of gunk on the tap when I removed it. I cleaned it each time and recoated it with cutting oil before each hole. Then I took it outside and thoroughly flushed each hole out again with brake cleaner, using the tube to be sure it was getting all the way into the hole.

After reassembly, I opted to use these flange head bolts instead of the crappy allen head bolts spec'd by the OEM. The weight difference should be nominal. Note that the part the bolts go through (not sure what to call it) is raised up. It only seats so far down into the diff housing before it runs into the stack of internal parts.

I thread each bolt down by hand until its flange touches the surface of the part. At this point, I wanted to be sure the bolt heads won't cause any clearance problems so I set the outer axle flange cover (BMW calls it transmission cover) over the bearing. There's plenty of clearance here. The only other question will be when I put the assembly back into the case. But I think it will be fine.

The next thing I did was to number the bolts with a black sharpie. Last time I did this, I kept losing track of where I was and these things need to be tightened in a crossing pattern very evenly to make sure the part they secure is pushed evenly down into the case.

One other thing I thought folks might find valuable is how to keep the assembly from spinning as you tighten the bolts. I used a small pry bar as below. It's best to keep the handles of the tools roughly parallel so you're squeezing them together. At least, I found that to be the most stable position. You can use this technique when removing the bolts as well. I also used blue Loctite on the bolts. Tightening torque is hard to find internet and it's not listed in the Bentley manual. It's 25 ft/lbs.

Thought I'd update this post as I have still not solved my inside tire spin issue. I went to VIR a few weeks ago and had Quantum Speed Works look at the suspension. They could not find anything that should have been causing the wheel spin in the suspension itself. They felt that the diff was not locking up.

So, I decided to upgrade yet again. This time, I'm stuffing a 6-disc conversion in there. I went with RacingDiffs Stage-3 package: https://www.racingdiffs.com/collections/lsd-clutch-plate/products/bmw-188-mm-lsd-clutch-packs?variant=33470318739591

VERY IMPORTANT IF YOU ORDER: They need to know what your final drive ratio is. Email it to them after you order as they request it on the product page but there's no place to enter it when ordering.

It's a bit of a CF trying to communicate with those guys. They're responsive and they try to be helpful but their English is only marginally better than my Bosnian.

The biggest issue I have is finding good descriptions of how the parts are supposed to go into the case.

I found a TIS guide on LSD disassembly/assembly here, but it is for the stock 2-disc LSD:

https://www.newtis.info/tisv2/a/en/e36-323i-cou/repair-manuals/33-rear-axle-rear-wheel-drive-guide/33-14-locking-differential-with-bearings/BIHpnmD

I'm praying that I'm understanding the orientation of the conical spring washers (#3 Disc spring).

RacingDiffs doesn't use the same nomenclature as is found in the PDF above, so the mapping for the parts that are replaced is:

BMW Nomenclature ------RD Nomenclature
________________________________________________
1    Case Covers-----------Billet Steel Cap
2    Thrust washer---------N/A
3    Disc spring------------N/A
4    Tab washer------------Oil shim set
5    Shim------------------N/A (4mm OEM shim not used)
6    Disc spring------------Preload shim set (Belville Washer or Spring Washer)
7    Outer disc-------------Part of 6 clutch set - dog ear plates
8    Inner disc-------------Part of 6 clutch set - friction discs
9    Thrust ring------------N/A - this would be what RD calls a Ramp
10   Differential side gear--N/A
11    Differential pinions
       with differential axle--N/A
12    Expansion housing---N/A

The parts they list as "6 clutch set" are 6 dog ear plates (#7) and 6 friction discs (#8). The RD dog ear plates are coated with friction material (Molybdenum-Ceramic coated) as well as the friction discs. Their oil shims (#4) are also coated.

The one thing that RD includes that is not OEM is the Ramp Saver Shim (basically another dog ear plate but it is not coated with friction material). This is only necessary if your Thrust Rings (Ramps) are scorched and need to be surface ground. If they do need this, they recommend taking them to a machine shop for the grinding. Have them take off .5mm (I assume .25mm from each). The you will add back that space with the Ramp Saver Shim. They provide a PDF when you order that explains this and where the shim goes in the stack.

The main differences from what's in the PDF above is that you don't reuse your old #5 shim (the 4mm thick flat shim) and there are 3 Inner (#8) and 3 Outer discs (#7) on each side of the Thrust Rings. You'll alternate these so against each Thrust Ring you'll have inner, outer, inner, outer, inner, outer. Everything on each end of this stack should be the same as OEM. (NOTE that RD does not provide any instructions. I had to email them to find out what order to put the parts in)

One other thing is that you have to use their Case Cover (Billet Steel Cap) because it has a shorter shoulder than the OEM cover to make room. That means you either have to press off your old bearing and put it on the new cover or you have to order a new bearing. RD offers a kit with all the bearings and seals but it's 165 Euros. I could not find the part number for the bearing anywhere. I emailed RD and they graciously gave me the part # and I picked it up for $26 from Pelican Parts: BMW Carrier Bearing - Differential Output Shaft - 33131204568 - FAG 103133

I chose to order a new bearing because I was afraid that pressing off the old bearing could damage it. Installation should be easier. I found a YT video where the case cover was put in the freezer and the bearing was heated with a propane torch. Then the bearing slipped on easily. So I'll give that a try tomorrow.

I got the new bearing fit to the case cover. I cooled the case cover in the freezer for an hour and a half and heated the bearing to 350 degrees. I put the bearing into the oven while it was preheating and left it in for about 10 or 15 minutes after it reached 350. The parts plopped together perfectly and cooled to room temp in about 5 minutes. That's my first time trying this technique so I was pretty impressed.  

My wife was less impressed with the smell of the oil cooking off the bearing.

But now I have another problem. With all the parts stacked carefully into the case, they take up ALL the space. 

As you can see, there is no space left between the surface of the case and the top of the spring washer.

And that means the case cover doesn't even recess into the opening. Since RacingDiffs doesn't provide any instructions, I'm not sure what to make of this. I have my doubts as to whether putting in the bolts and cranking them down would compress the springs enough to allow it to seat all the way down. And even if it did, we're talking about a centimeter of preload. I think that's more than the springs can move.

Race is next weekend. Hope I can get this sorted before then. 

UPDATE: See two posts down for solution.

DELETED. SEE NEXT POST.

PROBLEM SOLVED!!!

When last we left off, the stack of parts seemed too large for the space inside my diff case. I was uncomfortable with trying to compress all 12 discs into my case. So I took two discs out and torqued it down yesterday.

This morning I took the cover off again to see if perhaps the stack had compressed after being torqued and it had. It sprung back out but I gained about 2 mm.

So I re-installed the two discs I left out yesterday and put the case cover back on. It was still not contacting the opening but it was close enough to get each of the 8 bolts started. I decided to go for it.

In order to make sure that the case cover is aligned perfectly, I stuck the axle stub through it into the LSD. Then I very incrementally tightened down the all the bolts. It compressed the internals and the cover seated just fine. 

Keeping my fingers crossed that this ends my one-tire-fire issue for good.

In reply to jimgood :

So with the new cover and bearing do you need to check gear mesh and possibly adjust the side shims (I think there are shims?) to get things centered properly? My old diff could use a rebuild but I'll probably just go with the 3 plate upgrade.

Thanks for posting your progress.

Man, checking in here. Blue NB Miata TT6 car that will probably be jumping to ST6 next year. I saw your car in paddock, was positive I had read the rules poorly, and was really going nuts that I hadn't thought to build exactly what you've got. Now it makes sense :).

What happened with the front end of your car? 

 

In reply to SpeedTheory :

Oh, the front end of my car? I managed to do that during qualifying. I was behind a 944 when the yellow flags dropped. The stack of cars in front of us was lollygagging. He made a move to go around them and I followed right on his bumper. Then he checked up and boom. Damage to his car was minimal. Mine, not so much.

In reply to jimgood :

The paddock reconstruction after that episode remains one of my fondest track memories. 

Old thread, but a very good write up!