Help! Brand new clutch installed with swap is slipping!!!

[tsbpenguin]

So I recently completed a ej22t and 5mt swap on my 93 legacy. The transmission is out of a 98 forester. So I got a brand new factory flywheel, clutch disc, pressure plate, throw out bearing, pilot bearing (for a 98 forester of course). Used master and slave cyl out of the forester the tranny is from. First drive, the pedal was pretty much all the way to the floor and not returning all the way, clutch slipped. Adjusted the fork on the master cylinder and the pedal is much better, but the clutch still slips. It disengages the clutch completely when I press the pedal, and seems to grip decently well when I am starting out in gear, but once I am in gear, it slips even under a minor load. It's almost like the clutch isn't fully engaging when I release the clutch pedal. Like the throw out bearing is maintaining minimal pressure on the pressure plate or something. Any ideas? Bad master or slave cyl? Something worse? I'm praying I don't have to pull it all apart again...

[Legacy777]

Is there any play on the clutch fork with the clutch fully disengaged and car sitting there? If not and if it feels "tight", then you may have adjust the clutch master cylinder push rod clevus pin too much.

There is a point on the master cylinder where if the push rod is not fully disengaged, fluid from the slave cylinder and lines cannot make it back into the master cylinder. So once you start building heat in the engine, the fluid expands and exerts pressure on the slave cylinder which partially engages the clutch.

If I had to take a guess, this is what I think your problem is.

[wtdash]

The '98 Foresters had a TSB/recall on the clutch MC or slave. I'd bet you need new/rebuilt parts....but I've read that bleeding these is a real PITA, too.

Re-bleed first.

You can also adjust the clutch pedal from inside the car (@ least I could on the '98 Fozzie I had). There's a hard-to-reach adjustment behind/near the GO pedal...but you probably saw that when you did the pedal swap?

GL,
Td

[tsbpenguin]

Is there any play on the clutch fork with the clutch fully disengaged and car sitting there? If not and if it feels "tight", then you may have adjust the clutch master cylinder push rod clevus pin too much.

There is a point on the master cylinder where if the push rod is not fully disengaged, fluid from the slave cylinder and lines cannot make it back into the master cylinder. So once you start building heat in the engine, the fluid expands and exerts pressure on the slave cylinder which partially engages the clutch.

If I had to take a guess, this is what I think your problem is.

There is play on the fork when the car is just sitting there. The clutch was slipping before I made the pedal adjustment, so I really don't think it's an adjustment issue. I jerry-rigged some rubber bands as a "return spring" to see if it was an issue with the TOB constantly maintaining pressure on the pressure plate, but it didn't seem to make a difference. At this point I'm thinking hydraulic issue, or defective parts. When I got back from testing the rubber band - return spring I was able to come to a complete stop with the car in gear and the car did not stall.

The '98 Foresters had a TSB/recall on the clutch MC or slave. I'd bet you need new/rebuilt parts....but I've read that bleeding these is a real PITA, too.

Re-bleed first.

You can also adjust the clutch pedal from inside the car (@ least I could on the '98 Fozzie I had). There's a hard-to-reach adjustment behind/near the GO pedal...but you probably saw that when you did the pedal swap?

GL,
Td

Thanks for the heads up on the recall, I'm starting to think it's a hydraulic issue. I have re bled probably half a dozen time. Also, yea I did see the adjustment, and adjusted it accordingly since the pedal was not fully returning at first.

[Legacy777]

Ok, well if there's play in the clutch fork, then yeah I'd investigate the TSB Todd mentioned.

[tsbpenguin]

Pulled the Master and Slave Cylinders completely off the car and started the car in 1st to see if the slipping persisted without hydraulics, and it did. So now to pull the motor back off and see wtf is up...

[Legacy777]

That sucks! But good troubleshooting to eliminate the hydraulics as a possible cause.

[tsbpenguin]

Ok this is starting to get quite frustrating. Pulled everything back apart (getting quite efficient at these motor pulls), and pulled the pressure plate and disc back off the flywheel. Everything was installed in the proper direction, and I even double checked the torque of the pressure plate bolts and they were correct. No scoring, heat marks, or excessive wear. The disc surface even still showed most of the wirting that said "flywheel side". What now? Return the parts and get a new full kit?? Or is there something I'm missing here?!?!

[tsbpenguin]

Ok so I believe I am an idiot. I have been running the car with only the front two axles because the flange from the driveshaft to the rear diff is different. I had the FWD fuse installed running in FWD mode, and did not realize that the manual cars did not come equipped with this. I am assuming it was not the clutch slipping, but the center differential slipping due to no resistance in the rear?? Could installing the new flange on the rear diff and connecting the driveshaft be my magical fix to the slipping??

[wtdash]

Yes....the center is an open diff, so you'd only be getting @ the most 50% of the 'torque' to the front wheels, IIRC...

EDIT: @ least on a manual. I believe the automatics / 4eat do have a variable split from 90/10 to 50/50 when needed.

And no idiots allowed on here, so worst case you're just human!

TD

[tsbpenguin]

Yes....the center is an open diff, so you'd only be getting @ the most 50% of the 'torque' to the front wheels, IIRC.

And no idiots allowed on here, so worst case you're just human!

TD

Bingo! Haha I'm just frustrated that I didn't realize it until I took everything back apart. Oh well, peace of mind knowing everything clutch related is legit and installed properly. Might do a couple other cleaning tasks and gasket replacements while I've got everything accessible.

[Alphius]

That makes sense. You will only get as much torque transferred to the front wheels as the viscous fluid is capable of transmitting (4kg-f/m for every 100RPM difference). Any extra torque over that would translate to a slipping center diff and be converted to heat.

Hopefully you didn't drive too much and damage the center diff by driving it in constant slippage.

[tsbpenguin]

That makes sense. You will only get as much torque transferred to the front wheels as the viscous fluid is capable of transmitting (4kg-f/m for every 100RPM difference). Any extra torque over that would translate to a slipping center diff and be converted to heat.

Hopefully you didn't drive too much and damage the center diff by driving it in constant slippage.

Drove less than 5 miles doing simple road tests under 35mph so I'm thinking it should be fine but keeping my fingers crossed.

[mike-tracy]

Similar thing happened to me, in that one front axle was completely broken and one rear axle wasn't seated in the diff, when I first got my SS running. Wasn't the most obvious thing since there were no nasty noises.

I prob put 5 miles on mine in that state, and in 18k miles since then, the tranny has been fine.

[tsbpenguin]

Similar thing happened to me, in that one front axle was completely broken and one rear axle wasn't seated in the diff, when I first got my SS running. Wasn't the most obvious thing since there were no nasty noises.

I prob put 5 miles on mine in that state, and in 18k miles since then, the tranny has been fine.

Thanks for the response, that definitely gives me hope. Everything is almost all back together but I'm still having trouble sourcing the right flange to match up to the driveshaft... Oh well I have all day tomorrow to find the right one and get it on the road.

[rallyak]

Was you car originally 5spd? If it wasn't it may have the wrong rear gear ratio. I believe 98 Forster 5spd is 4.11, I maybe wrong. Also which driveline are you using?

[tsbpenguin]

Was you car originally 5spd? If it wasn't it may have the wrong rear gear ratio. I believe 98 Forster 5spd is 4.11, I maybe wrong. Also which driveline are you using?

Car was originally 4EAT, but 4.11 rear diff was swapped in with Forester tranny. Just haven't found the correct flange for the rear diff that matches the driveshaft I have. That is my task for the day.

[mike-tracy]

You can grab the flange off the old 3.9 diff if you have it.

[tsbpenguin]

You can grab the flange off the old 3.9 diff if you have it.

I would have long ago if I had known but I bartered the diff for some other parts so I don't have it any more.

[James614]

This thread actually answers the question I used to have regarding how much torque the VLSD in the 5MT was capable of transferring (4kgf/1000rpm doesn't sound like much at all, and doesn't say much about ability to move the car). I would have thought the fluid would overheat rapidly and just leave you sitting useless spinning the wheels with 0 traction in very short order. But apparently you can actually disengage a set of wheels completely drive the car around using ONLY the torque applied through the fluid to 1 set of wheels. The slippage shows it's nowhere close to a 4EAT in full lock, but still no slouch of a 4WD setup.

[mike-tracy]

^^ The stock one is 4kg, but you can buy Subaru 12 & 20kg centers as well, although the 20kg is "motor sports only" basically. I'd love to find a 12kg center and put it in.

[Alphius]

I've heard the 12kg center and definitely the 20kg cause some binding feeling in sharp corners. The 4kg is enough for a road car definitely.

It's actually 4kgf-m/100rpm difference, which is actually quite a bit depending on degree of slippage. This is dependent on the difference in speed between the front and rear wheels at the output shaft of the transmission. This will lead to a more readily lockable center at lower speeds with minor wheelspin and more open at higher speeds for freeway driving and long sweeping corners where the front and rear wheels must turn at different speeds.

In my experience, you have to really abuse a Subaru VLSD center before you harm it. The easiest way is with sustained overheating, which really only happens with different sized tires requiring slippage 100% of the time driving the car. You won't hurt it or lose effectiveness even by hard driving in slick situations. Consider that the higher the input torque is, the more locking effect you get which reduces slippage and therefore reduces heat. A friend of mine used to do e-brake slides with his 5-speed WRX all the time, and even after 6 months of that behavior the center diff was not noticeably harmed. (Measured by torque required to turn one wheel off of the ground against the VLSD compared to a good condition 5-speed trans)

When the fluid overheats and eventually fails it has basically two failure modes:
1. solid lock which will torque bind badly in corners and eat up tires quickly, and probably eventually break CVs. This would basically be like a pickup truck locked into 4wd.
2. No locking at all. Another friend of mine had a '97 Legacy that had been run for who knows how long with different sized tires on all 4 wheels and it had this problem. It was an open center diff at this point, so it could easily spin one tire. On wet pavement, if you launched it even semi hard it would spin one front tire all day long just like a FWD car, and in the snow it often spun just one rear tire on hills. This works like the old RX center diff lock in unlocked mode, basically just another open differential between front and rear, driving all 4 unless there is wheelspin.

[jamal]

This thread actually answers the question I used to have regarding how much torque the VLSD in the 5MT was capable of transferring (4kgf/1000rpm doesn't sound like much at all, and doesn't say much about ability to move the car). I would have thought the fluid would overheat rapidly and just leave you sitting useless spinning the wheels with 0 traction in very short order. But apparently you can actually disengage a set of wheels completely drive the car around using ONLY the torque applied through the fluid to 1 set of wheels. The slippage shows it's nowhere close to a 4EAT in full lock, but still no slouch of a 4WD setup.

The diff will actually protect itself when it gets too hot through what is called "the hump effect." When the diff gets hot enough the plates physically contact each other and lock together (which will allow it to transfer like 100% of the torque and get the car moving). However, that's not a permanent thing so it will cool off, separate the plates, slip, then heat up, lock again, and repeat. Not a good situation for the diff.

Also, 4kg/100rpm is not a linear curve. So for the first couple hundred rpm speed difference, yes you get a pretty direct increase in torque. But with like 1000 or more that's not the case, but by then things are hot and it locks together anyway.

Here is the relevant section of the FSM:
http://ken-gilbert.com/wrx/STi_Manual/0 ... ential.pdf

[Alphius]

Good call on the FSM. I have seen that information before but couldn't find it again when looking. My explanation was a little simplistic and didn't take into account the "hump" or non-linearity. For practical purposes (ie. street driving) those things shouldn't make much difference but it is definitely good to have that info here. I enjoyed reading it over once more.

Regardless of this carrying on, I think tsbpenguin's center diff will most likely be perfectly fine after his little jaunt.

[James614]

That hump effect is pretty darn cool, I've never heard of such a thing before. Is that a characteristic of VLSD's in general, or specific to the MT's center differential? Cause I've never seen such a thing mentioned when reading about VLSD's anywhere else.