Subaru Outback Forums banner

Replaced clutch - clutch doesn’t work

6K views 40 replies 10 participants last post by  Numbchux 
#1 ·
I pulled the motor from my ‘02 OBW to replace the HGs and clutch. I Installed a new Sachs clutch kit (pressure plate, friction plate, pilot bearing and release bearing). had flywheel resurfaced. Clutch install want smoothly, apart from the frustrating process of getting the motor back onto the transmission. I went to it the car in gear prior to torquing down the crank pulley, and the clutch pedal went to the floor and stayed there. I didn’t touch the hydraulics or release cylinder during the clutch install, but it feels like there’s no fluid in the system (I checked the reservoir and it‘s full to the max level with decent clear fluid). I did pull the release lever and degreased the ball pivot, but then reinstalled it. It seemed to move the release bearing normally, so I didn‘t think there was any more to it than that. Even with the car in gear, I can manually turn the crank pulley, suggesting that the clutch isn’t engaging. Any ideas? Otherwise, I’m contemplating having to pull the motor again to see what’s going on ?
 
#2 ·
Sounds like a pressure plate issue. Do Sachs PPs come with that star-shaped lock that you remove on installation? Are these 'self-adjusting" PPs? On another make, I've heard of folks getting new LuK or Sachs self-adjusting kits that were actually returns and thus missing the star lock, requiring that the PP be reset either with a special expensive tool or ******* ingenuity.

Although I think the symptoms would be reversed, could you have installed the friction plate backwards?
 
#3 ·
Thanks for the reply. I don't think Sachs PPs come with any sort of locking device. This is the second Sachs clutch I have installed (The first was on a Triumph TR6) and it looks no different than that earlier one or the clutch I just replaced on the Outback. The friction plate has a clearly-marked "gearbox side" which faced me as I installed the clutch onto the back of the motor. Assuming that the hydraulics are OK (will bleed the system, but I didn't touch the hydraulics during this whole process, so no reason to believe that they're compromised, short of some coincidental and catastrophic failure, but the reservoir is full), the fact that the pedal went to the floor suggests that the release lever wasn't pressing on anything. The new release bearing was in place, and clipped on, so there's no reason why is shouldn't be hitting the PP fingers. Even if I had forgotten the release bearing, the clutch default condition should be engaged, so once I put the car in gear, that should lock up the motor, but it spins free even in gear, suggesting that the PP isn't pressing on the friction plate (it was when I assembled the clutch onto the flywheel).
The only other thing I can think of (and it's a stretch) is that all the joggling around to get the motor onto the tranny somehow disrupted the linkages between the gear lever and the tranny itself, so, although I think I'm putting it into gear, the tranny is still in neutral, which would explain my ability tho spin the motor while "in gear", but doesn't explain the odd clutch behavior. I'm mystified.
 
#4 ·
Sounds like the lever isn't seated correctly. Like maybe popped off the pivot ball and slipped down into the bellhousing enough that the slave cylinder isn't engaging it.
 
#5 ·
Worth checking - although the "dimple" on the top of the lever that engages the slave cylinder piston is lined up perfectly. That being said, the lever itself is very loose and rattles around a bit. I don't know if this was always the case before I embarked in this job. Tonight I'll get someone to press the clutch and observe how the slave cylinder piston engages with the lever
 
#6 ·
I was going to ask about the clutch fork, wondering if perhaps the opposite end isn't resting flush on its pivot. I'm no help as I'm not clued in on Subaru clutch work, but also wondered if the trans input shaft has enough foward/backward play that when you re-installed the engine, its tip might not have lined up and fit square into the pilot bearing and is somehow not engaging the friction plate splines.
 
#7 ·
So, when I was in the throes of trying to get the motor back onto the tranny (a week-long bloodbath best forgotten...!), at one point, I "dry-fitted" the motor, just to reassure myself that it did indeed fit and that the problem was in aligning the splines, the input shaft and the pilot bearing. I removed the PP and friction plate and fitted the motor with the flywheel alone. I also placed a strip of masking tape across the pilot bearing as a telltale to confirm that the input shaft went in. Motor fitting went pretty smoothly, and when I pulled the motor, the input shaft had "cookie-cuttered" a perfectly circular hole in the tape, telling me (1) that the input shift was getting into the hole, and (2) that the alignment of the shaft with the pilot bearing was critical, as the fit was extraordinarily snug.
Once I pull the motor again (sigh....?), I'll do some measuring to ensure that the input shaft splines are engaging those of the friction plate
 
#8 ·
I'm thinking this is all connected. Were you able to slide the throwout bearing onto the snout around the input shaft?

I have a theory. Your previous clutch kit maybe had a repair sleeve around that snout, and your new throwout bearing does not fit around it. This would explain the resistance putting the engine back in, as you were compressing the pressure plate as you tightened the engine back up. That would release the clutch, with no resistance on the slave cylinder.
 
  • Like
Reactions: plain OM
#9 ·
Throwout bearing installed fine and moved back&forth on the snout without issue, but I think the answer (when I finally steel myself to pulling the motor again) will lie somewhere in the lever/bearing. Except for the work involved, I'm looking forward to solving this puzzle - the more you know, etc etc
 
#10 ·
A followup: I pulled the motor, and the release bearing and lever looked fine. I cycled through the clutch, and apart from the clutch pedal bottoming out because there was no resistance against the release bearing, everything moved as it should. I regreased everything anyway
473783

However, when I compared the old and new pressure plates, I noticed that the fingers on the replacement plate were recessed by ~1 cm with respect to the pressure plate body, compared with the fingers on the new plate: Here's the old PP - the fingers are flush with the rim of the PP body:
473787


and here's the replacement PP - the fingers are ~1 cm recessed below the PP body rim:
473788

Since the total movement of the release bearing is only ~1 cm, it's possible that the bearing is at the end of its travel before it even engages the fingers of the new PP, which would explain why the clutch pedal was bottoming out when the motor was mounted. It doesn't explain why the clutch appeared to be disengaged (ie, even when the car was in gear, I could still spin the motor with a wrench in the crank pulley). As it is currently mounted on the motor, the replacement PP is gripping the friction plate. My concern is that, with the fingers as compressed as they are, is there even enough travel in them to release the friction plate, assuming even that the release bearing has enough travel to engage these recessed fingers?
I may have the wrong clutch for the Subaru, although I ordered the appropriate kit from the Rock Auto site. However, the return period is expired, so I guess this clutch is mine, regardless :(
 
#12 ·
Further update - I checked the numbers on the order and on the box, and everything matches. I also loosened the new PP, and the fingers lifted up. The difference I was seeing was the difference between a "relaxed" uninstalled PP and the installed PP "under tension" against the friction plate. So, I'm back where I started, with no choice but to reinstall the motor and see what happens
 
#13 ·
Not being familiar at all with this specific set up, I can only offer general input.
  • Did you compare all the new parts to the old parts? T/O bearing, clutch plate, etc?
  • Is it possible if the clutch plate is installed backwards, the transmission input shaft splines will not reach? Some plates are not symmetric front/back. The splines may be offset towards one side of the plate.
 
#15 ·
Thanks Folks - yes the friction plate is clearly marked “gearbox side”, which faces outward when installed on the rear of the motor. The new parts look identical to the old parts, except the friction plate is thicker obv. There’s no reason why this shouldn’t work - check everything once again, reinstall the motor & bleed the clutch
 
#16 ·
From your other post about having trouble getting the engine in:
rattled that mofo until it just “went” - quickly got some nuts onto the studs and locked them down.
Ultimately, this is your problem. When you had to force the engine and transmission together, it released the clutch. If this happens again, STOP.
 
#19 ·
From your other post about having trouble getting the engine in:


Ultimately, this is your problem. When you had to force the engine and transmission together, it released the clutch. If this happens again, STOP.
No - my description was likely misleading - there was a lot of wiggling to get the motor and the tranny lined up (getting the snout aligned with the pilot bearing, I imagine) and getting the tranny upward angle and the motor downward angle lined up, but at some point, everything lined up and the motor slid on. I didn't have to force it on, and I certainly didn't tighten any fasteners to pull it on - once the motor was sat on the bell housing, the first thing I did was screw in a couple of fasteners to keep it in place (remember the motor was dangling on a hoist - the last thing I wanted was for it to swing off the bell housing again when my back was turned
 
#18 ·
I just don't know what I'm missing. I checked everything. The replacement clutch parts look identical in shape and size to the old parts, and the part# on the box matches that of the appropriate clutch kit on the Rock Auto site. The friction plate is installed in the correct orientation. When the pressure plate is installed and torqued, it grips the pressure plate. The release bearing is installed and moves smoothly on the snout. I installed the motor again - the installation went smoothly with minimal wiggling to get it to mate snugly with the bell housing (practice makes perfect, obv). BUT - something's not right. Here are the symptoms:
1 - when I push the clutch, there is NO resistance until right at the end of travel (ie, pedal is on the floor), the slave cylinder piston is up against the release fork, which appears to be at the end of its travel, so I guess the fork/release bearing has reached the pressure plate fingers, but that's all.
2 - With the clutch "engaged" - pedal pulled back up, release fork slack (ie, release bearing isn't contacting pressure plate), car in gear, I can still spin the motor with a wrench on the crank pulley - I can see the flywheel spinning through the access panel. Either the clutch isn't actually engaged, or the car isn't actually in gear. I visually inspected the gear linkage while someone operated the shift lever - everything looks fine (ie nothing obviously dangling loose) - the shift action itself feels normal

Symptom 1 suggests that the release bearing has to travel too far to contact the pressure plate, and that the clutch pedal runs out of travel before it can actually actuate the pressure plate - ie, the clutch would be permanently engaged
Symptom 2 suggests that the clutch is actually permanently unengaged, although as I said earlier, the pressure plate clamped the friction plate fine before the motor was engaged.

None of this makes sense. All I can think of is to simply buy another clutch kit - maybe from another manufacturer
 
#21 · (Edited)
Symptom 1 suggests that the release bearing has to travel too far to contact the pressure plate, and that the clutch pedal runs out of travel before it can actually actuate the pressure plate - ie, the clutch would be permanently engaged
Symptom 2 suggests that the clutch is actually permanently unengaged, although as I said earlier, the pressure plate clamped the friction plate fine before the motor was engaged.
In regard to Symptom 1, with the clutch pedal up, and the slave cylinder pushrod returned to the rest position, can you then manually move the release lever the equivalent distance without meeting resistance from the release bearing? (Just verifying that there is a significant "gap", i.e., slack, between the bearing at rest and the clutch fingers.) I might be wrong, but I'm fairly sure the throwout bearing should be very close if not up against the fingers, so there should be very little if any slack in the release lever.

Now, for Symptom 2, are you sure that when the clutch is installed on the flywheel, the friction plate is in fact against the flywheel? In the earlier photos you noted that the fingers are recessed about 1 cm when the clutch case (with pressure plate and fingers) is installed to the flywheel. I believe the thought is the pressure plate is pressing against the friction plate which, in turn, is against the flywheel, and that's deflecting the fingers. But is that correct? After the clutch cover is installed on the flywheel, can the position of the friction plate be checked relative to the flywheel? In other words, is the friction plate indeed against the flywheel? Could it be a mm or two away, thereby causing the other end of the fingers to be depressed more, such as 1 cm? If the friction plate isn't against the flywheel, it won't be adequately "clamped" so it would act as if the clutch isn't engaged.

Edit: kudos to @seagrass for the additional ideas. I had drafted submitted this post without looking back to see that your post had already been added.
 
#20 · (Edited)
I really sympathise with you for all the trouble you are having. I replaced the clutch in my daughters 2007 Legacy yesterday and with the help of my son in law the job was done in just over 4 hours. We went on to flush the brake and clutch fluids, rotate the tyres, replace front brake caliper pin rubbers and a power steering hose and we still finished in under 6 hours.

Getting back to your problem, if the clutch release fork and bearing are not in contact with the pressure plate then there is a problem somewhere, maybe the clutch fork is bent?

In Australia there are different non OEM clutch kits available depending on whether your vehicle originally had a solid mass flywheel or a dual mass flywheel. I am aware that there are different release forks for each type so maybe this could be another cause of your problem (wrong kit chosen).

If you have not owned the vehicle since new, could a previous owner have replaced the clutch and flywheel resulting in you ordering the clutch kit that should be correct but is not due to the change by the previous owner? My daughters Legacy had had this done and luckily I had purchased a complete flywheel and clutch kit or I would not have been able to complete the job due to a mismatched flywheel and clutch kit.

Seagrass
 
#22 ·
Thanks Seagrass - I've owned the car from new. This will be its 3rd clutch, the original having been replaced at a Subaru dealership at ~90k miles (currently at ~170k). I decided to replace this one since I was doing the head gaskets anyway, and the clutch was starting to slip, but otherwise everything was working fine. The flywheel is original - I had it resurfaced for this job. A bent fork is an interesting possibility - when I reinstalled the motor the first time, there was a lot of wiggling and shoving (a bloody nightmare all-round!) - could that have damaged either the fork or the pressure plate? I wouldn't have thought so, but there's clearly something at play here that I'm just not seeing, so I'm willing to entertain any possibility. However, this doesn't explain why the clutch isn't engaging at all (motor spins in gear). I think I need to reinstall the old clutch kit (god help me!:rolleyes:) to see if it's some issue with the new clutch. If it was some other part, at this point I'd just bite the bullet and buy OEM, but Subaru no longer sell an OEM clutch kit for this car (at least not in the US), so I'm going aftermarket whether I want to or not
 
#23 ·
Thanks Plain OM - yes, when the pedal is up and the piston returned, there is significant slack in the lever - I can push it back ~1 cm until it hits abrupt resistance - presumably the release bearing hitting the PP fingers - indicative of some distance between the "resting" bearing and the PP. On the second issue - yes, the friction plate is certainly contacting both the flywheel and the PP. I can see all three parts through the top of the PP. The friction plate ceases to turn with the alignment tool long before the PP bolts have reached their final torque. When I pull the motor again, I'll use a sharpie to put index marks on the edge of the friction plate and on the flywheel to see if the two are actually rotating separately when I spin the motor.
 
#24 · (Edited)
when I pull the motor again, I'll use a sharpie to put index marks on the edge of the friction plate and on the flywheel to see if the two are actually rotating separately when I spin the motor
Good idea. Hope that will help resolve the seeming inconsistency between symptoms 1 and 2.

Another thought regarding the difference: Any chance the input shaft to the transmission was damaged when there was that initial difficulty getting the engine lined up? Can the shaft be turned manually, the same as when turning the engine at the crank bolt?

I think I need to reinstall the old clutch kit
Perhaps first the full kit, and then perhaps the old cover but with the new friction disk, comparing the position of the fingers. Similarly, could install the new cover with the old friction disk and again see where the fingers land.

As noted, a bent release lever could lead to the gap between the bearing and the fingers, but it wouldn't explain the engine being free to turn as if the clutch isn't engaged (unless something beyond the tranny input shaft has failed . . .)

EDIT/ADD

The friction plate ceases to turn with the alignment tool long before the PP bolts have reached their final torque.
Not much torque possible when just turning the tool. Different when turning the crankshaft with leverage, and the transmission input shaft is held by engaged gears (supposedly). That same inability to turn might be caused by the friction plate contacting the pilot bearing (e.g. if the bearing isn't set in fully), yet it wouldn't resist turning the crank. Just some more thoughts on the various observations and what they might mean . . .
 
#26 ·
OK, so yes the drive will be sent to the front before the rear if you have no axles in.

Excessive play in the fork seems like the clutch fork inspect for cracks, particularly around the pivot ball.

One way to test the box would be when you have the motor out put the old clutch plate on the input shaft and see it everything feels ok and you will be able to check the front output splines for rotation.

I would recommend replacing the fork now as three clutches on it’s probably soon to fail.
 
#29 · (Edited)
So, one symptom down, one to go. I installed new CV axles and put all four wheels on the ground. Symptom 2 (engine spinning while in gear) is addressed. I torqued the crank pulley to the requisite 130 ft.lb, so the clutch is clearly engaged. I really shot myself in the foot with that one ?. I removed the axles for better access to the lower bell housing nuts (could've simply jacked up the motor a couple of inches, but..duh) and never considered the consequences to how the transmission would behave. Been chasing my tail (and bringing you guys along for the ride - apologies) for a couple of weeks over this.
Symptom 1 (inability to release clutch) remains. I pulled the motor and retrieved the fork (pics below). Apart from some shiny wear at the contact points, there's no sign of trauma or cracking, and this is a stout piece of metal, so I can't see it bending without obvious signs of damage - and there are none. The problem seems to be that the resting release bearing is just too far away from the PP fingers, so by the time it contacts the fingers, the clutch pedal is pretty much at the end of its travel. Any number of things might address simply moving the resting release bearing forward a bit - a "deeper" bearing or some sort of spacer, a straighter fork, a longer slave cylinder piston - but this system was working fine before I started this, so I don't know what has changed, except maybe the friction plate, which is at least a couple of mm thicker than the plate it replaced. This would cause the fingers to deflect inward by a few mm - would this be enough to essentially pull the fingers out of the reach of the release bearing?
In any event, something to be addressed first - all the pushing of the clutch pedal against no resistance resulted in the slave cylinder piston popping out (can feel it wandering about inside the rubber dust cover), so I'll clean up the hydraulics before I go any further - change the fluid and bleed the system. See where we are from there. I might also order up a new fork - as Timberwolf says, it's been on the road a long time, and $27 is a small price to pay for peace of mind. My other car is a '72 Triumph TR6 - getting to the transmission involves pulling out half of the interior, so the general wisdom is - if you have to pull the tranny to do the clutch, replace every potential wear item preemptively, regardless of how good it looks - having to go back in to replace a failed part that would've cost you $10 while you had it apart would be heartbreaking.

474167

474168
 
#32 · (Edited)
but this system was working fine before I started this, so I don't know what has changed, except maybe the friction plate, which is at least a couple of mm thicker than the plate it replaced. This would cause the fingers to deflect inward by a few mm - would this be enough to essentially pull the fingers out of the reach of the release bearing?
I don't think so. A new friction plate would be thicker than an old, worn one and that difference would be translated to a different deflection of the fingers as the case is tightened down. However, the hydraulic system should work the same as the brakes.

With the brakes, as the pads wear, the caliper pistons move outwards to take up the loss and keep the pads very close to the rotor. The additional space behind the caliper piston is filled with brake fluid from the master cylinder reservoir. When the pads are replaced, the pistons are moved back into the caliper and the displaced fluid moves back to the reservoir.

The clutch hydraulics works the same way. When the clutch is new, the fingers will be deflected in, toward the flywheel. When the pedal is pressed the slave rod moves out against the lever, and after the first few "pumps", when the pedal is released, the fingers push back on the bearing, which pushes back on the lever, and through it on the rod. The slave piston moves back only as far as the fingers push the lever, in other words, the bearing remains touching or very close to the fingers. The slave piston remains out more, with fluid from the reservoir filling behind. As the clutch wears, the fingers settle further away from the flywheel. The lever will be pushed back somewhat more at rest, and that will push the slave piston further back, with the fluid going back to the reservoir.

The problem appears to be that the slave rod can't initially reach far enough with the new clutch to keep the bearing at the fingers. I doubt there should be a need for a longer rod, but I now wonder if perhaps the rod was out of position after the engine was first re-installed. Perhaps the need to reset the rod now might turn out to be a benefit.
 
#30 ·
Wow Spud I can almost smell the blood from here ....

I may be missing some insights that the previous posters have already provided, but If the new kit is stacked on the old FW, did you check for any variance in total, combined thickness between the old and new? Also ... Not having firsthand experience with a manual Subie, so maybe this is another question that doesn't matter, but could it be possible the pilot bearing isn't seated all the way into its recessed position in the crank?
 
#34 ·
Hi there - yes, the new parts - pilot bearing, pressure plate and release bearing - are dimensionally indistinguishable from the old equivalents. Pilot bearing is flush with the faced surface of the flywheel, as was the original.
 
#31 ·
Silly but necessary question follows,

Is there a small rod installed between the clutch slave cylinder and the clutch release fork? If this is missing it would be the cause of your “gap”.

If the rod is in place, the simplest way around your problem could be to use a longer rod to take up the slack. I am not sure this is the recommended fix BUT it should resolve the issue.

Seagrass
 
  • Like
Reactions: walker
#35 ·
Thanks Seagrass - yes, the rod’s there. That’s what was moving about under the dust cover, but it only became clear that the rod was distinct from the slave piston when I took the dust cover off. A longer rod would be pretty easy to fabricate, but I’ll hold off until I replace the fork and bleed the system before I go down that route. The whole thing worked fine before, so it’s more likely that there’s some mundane thing I have missed (see the whole “Symptom 2” shenanigans earlier ?)
 
#33 ·
That all makes sense. I’ll order a new fork - may make a difference, but won’t hurt in any case. Bleed the clutch system. another tweak to consider might be to put a washer under the pivot “ball”, which would shift the whole lever forward, shortening the instance for the slave rod to reach, and for the release bearing to reach. Even a 1mm washer might help, but I’ll wait to see what a new fork and bleed will achieve
 
This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.
Top