TSB on Transfer clutch

Mr. Pockets · #1 04-07-2004, 01:33 PM

I decided that, while I was replacing duty solenoid C on my wife's Legacy, I might as well replace the transfer clutch plates themselves. Even though the car has 150k miles on it, the clutch was working fine, but I decided it'd be better to do it while I have the thing apart.

It turns out that Subaru has changed their part numbers for transfer clutch sets, and you now have to disassemble it to find out what you need. The first driven pressure plate, just behind the snap ring, must be measured for its thickness. This procedure must be followed for all Subaru models. You'll find one of four thicknesses - 3.3mm, 3.7mm, 4.1mm and 4.5mm. Why the first is $25 and the rest are $37, I don't know.

What I'm also wondering is this. I found a 3.3mm plate in my tranny, but what if I put a 3.7mm - or even a 4.5mm - plate? Is this like a shim, accounting for different clutch hub dimensions, or would the thicker plate just displace some of the fluid in the cylinder, like a new brake pad?

Edit: Here's a diagram of the transfer clutch in the SVX. The Legacy's is pretty much identical.

http://www.subaruparts.com/diag/?mod...category=170-A

svx_commuter · #2 04-07-2004, 04:41 PM

Well I do not know but I can take a shot at it. That piston moves and applies contact pressure to the clutch pack. The thinner the whole pack is the more the piston has to move. So I think the thickness is as you said similar to a shim. The flatness of the plates is important too. I got to go cook the hot dogs.

Mr. Pockets · #3 04-07-2004, 10:38 PM

Quote:

Originally posted by svx_commuter
Well I do not know but I can take a shot at it. That piston moves and applies contact pressure to the clutch pack. The thinner the whole pack is the more the piston has to move. So I think the thickness is as you said similar to a shim. The flatness of the plates is important too. I got to go cook the hot dogs.

Right, but what would the thickness matter? The system is made to apply a certain pressure, not a certain volume.

UberRoo · #4 04-07-2004, 11:42 PM

I'll also venture a guess.

You may notice that the difference between each of these plates is four tenths of a millimeter. I suspect that the manufacturing tolerances on the clutch discs (or more likely, the friction material on them) and the offset on the clutch basket when it is welded onto the gear, cumulates to as much as 1.2 millimeters variance. I would think that this is largely trivial, but whatever free play exists in the clutch pack must be taken up by a certain volume of hydraulic fluid. The time it takes to generate this volume will lengthen the delay in AWD response...unless there is always a small amount of pressure on the AWD clutch.

Again, I would think this is all very trivial, but that's my guess.

NomadTW · #5 04-08-2004, 01:25 AM

i believe that there always is a small amount of pressure on the plate, unless the FWD fuse is in, and that is why the torque split w/o FWD fuse is 90/10 not 100/0

svx_commuter · #6 04-08-2004, 06:31 AM

The hot dogs were good.

I think as NormadTW says there is alway a small amont of the pressure on the piston. The pressure on the piston is totally independant of the position of the piston. There is a spring in there to return it. Right?
I can think of no reason to have the plates (shims) in a particular order other than they had to tell the tech's to do something to make sure the length came out right for the stack. If the plate stack is too short there will be no 10% at the RWD because the spring may hold it off the plates and if it's too long it may be no less than say 40% all the time because the piston cannot retract fully.

Hay, my 91 Legacy has 4.111 dif's and it still going strong at 215k miles.

oab_au · #7 04-08-2004, 05:36 PM

It could be so that a selection can be made, to ensure that the pack is not too thick, to cause binding.

Harvey.

Mr. Pockets · #8 04-08-2004, 06:08 PM

I think that these suggestions would point to a very flawed design. I don't think that the clutch pack has to stack up to such specific dimensions - otherwise, as it wore down, it'd be useless after the first couple thousand miles.

Again, I think that the system determines the pressure of the fluid behind the piston, not the volume. If it was the volume, we'd have the same problem that I just described. The clutch would be useless very shortly.

oab_au · #9 04-08-2004, 06:24 PM

Adjusting the thickness of the clutch pack is a normal practice in assembling a transmission. The pack must have a clearence to prevent binding, when the temp increases the thickness.

The transfer clutch has line pressure applied to it all the time. The transfer valve bleeds the pressure off, so it does not work till the transfer valve moves, to block the leak, and the clutch is applied.

Harvey.

UberRoo · #10 04-09-2004, 02:47 AM

I believe what the springs inside the clutch pack are supposed to do, is preload the clutch. There seems to be a lot of springs, much like a motorcycle clutch. It appears that they are supposed to apply a significant amount of force against the discs, thereby partially engaging the clutch. When fluid is added, the pressure supplements the springs, adding more force and friction. The amount of friction the springs generate would be the 10% of the 10/90 split. Perhaps when a different shim is used, the minimum torque split is modified.

When I disassembled my transfer clutch, there was no preload. The snap ring would have been holding back the springs, but it came out without effort. Also, the springs and transfer piston did not fall apart on their own, (which I would have expected them to do,) so it's possible that my springs were not working or that something was jammed. I didn't investigate. In hindsight, I would have looked a little closer. Perhaps my clutch discs were badly worn, though they're working just fine now.

If my second guess is correct, (that the reason for different pressure plates/shims is to adjust preload and minimum torque split,) that would mean several things.

1.) No extra volume of fluid must be generated to engage the clutch, as it would already be primed by even the lowest of backpressure from the transfer valve actuated by solenoid C.

2.) It would be quite easy to increase or decrease the clutch preload by changing shims. People who want more aggressive AWD response would use a thicker shim. People who find the binding in parking lots bothersome would use a thinner shim. (I'd probably use two shims.

)

3.) Even with a thin shim, the preload would probably be quite a bit, and I'd bet that assembly and disassembly would be challenging, particularly with regards to inserting the snap ring.