not a good day

[Mr. Pockets]

Originally posted by oab_au


Hi Nick, I would love to take the credit for pulling down the box and welding the basket back. Sadly This ability disappeared along with the right arm. That was Uberoo that did that nice piece of work. I think there was somebody else that had the same problem lately, and I think Immortal Suby broke the first one that I heard of some time ago.

Harvey.

Well, I didn't mean that you were the one who did the actual welding, just that you had it welded back together.

Still, I might look into that if the same happens to mine.

Of course, I guess I have a spare basket in my old tranny...

[oab_au]

Originally posted by UberRoo
oab_au is exactly right about what's wrong. As for why it broke, his explaination is as good as any. Personally, I think the design of the clutch basket is horrible and I blame the design solely for the failure of the part - not abuse by the driver or any other fault of the transmission or controller.

I agree, the way the basket is welded to the transfer gear is not what I would have designed. Still this is the same unit that has been used in all the 4eat transmissions, except the Euro, and has not given this type of trouble. So you just gota ask why does it fail in the SVX box.

Of course the answer is, as I have said, the transfer clutch not responding quickly enought. It is the only way this componet can suffer the server shock loading that will cause it to fracture. Why doesn't it happen in the other models that use this gearbox? Well The short answer is no other Subaru engine delivers the torque that this engine has at low rpms.

Normally the clutch is tightened, as soon as the TCU sees a difference in f/r wheel speed of more that 20%. If the clutch takes up then, with only a 20% difference in rotational speed, there can be no shock loading to the clutch. If it is delayed till the front wheels are spinning hard, when the clutch engages, it will either try to stop the front wheels from spinning, or start the rear wheels spinning. Either action will result in a server shock to the clutch, that it was not designed to take.

This Transfer clutch breakage and the front differential breaking are both due to the AWD not working as it should. The next question, why does the transfer clutch fail to operate properly? is another story I will try to post when I can write it.

Harvey.

[mranderson]

Looking forward to that story Is there any way to fix this problem is my question. And when i go to 4.44 gearing will it make it more of a possibility of breaking? Theres got to be something to do to prevent it

[UberRoo]

If any of you want your basket welded, here's what I learned doing mine:

1) The primary weakness in the basket is caused by the stamped offset at the bottom of the basket. There is a thrust bearing that needs to be centered, and there are three "ridges" that have been stamped into the bottom to align the bearing.

The first problem with this design is that to create these ridges, parts of the metal were sheared so the ridges could be pressed up. The shear lines themselves are unreasonably long, and run about 20% around the circumference of the bearing retainer.



The second problem with this design is that there are no relief holes at the ends of the shear! Any engineer would scream about this major oversight, but alas the problem exists. (You can observe these holes in circular saw blades sometimes when they cut the large vibration-dampening / self-cleaning / product-selling serrations into the disc.)

2) The resistance weld (giant spot-weld) was not complete on my basket. The part that was welded onto the gear tore off leaving metal still attached to the gear. The part that was not welded left nothing and the casting texture was evident on the gear. The only way to check an unbroken unit for this problem would be to MagnaFlux it, but because of the close proximity of the bearing alignment ridge shear marks, this may not be possible. The incomplete welds may simply be an intermittent quality control problem and may not affect everyone.

3) The metal of the basket is actually quite mild. When I lathed away the torn metal off the cast gear, I noticed that the metal the basket is made of was very soft and carved away smoothly. The gear however gave my carbide bit a good workout.

4) The basket must be offset from the gear slightly. By some careful measuring and application of the TLAR system, I concluded that a 1/16-inch spacer between the gear and the basket was about right - though it's not terribly crucial. If anybody has an unbroken basket to measure, the proper offset would probably be nice to know.

5) I built up the material on the back of the gear. I welded up the shear marks for the bearing alignment ridges on the backside of the basket. (Which should have been done from the factory.) I surface-welded another layer of metal to the back of the basket to build it up. I also built up the area on the gear where I welded the basket to it so as to have a large, deep and ultimately strong bond.

6) Getting the gear centered is quite a challenge. I carefully aligned the gear with every measuring instrument I had before I tacked the gear in place. Then I checked it again several more times to be certain it was centered. (With the clutch discs in place, that's a lot of mass to be spinning out of balance at propeller shaft speed.) I cut my weld-tacks three times and went through that process again before I was satisfied with the placement.

7) There will almost certainly be a slight warpage of both the basket and the gear. The clearances on the basket are quite large and the warpage, though measurable, is insignificant. The clearances on the gear are incredibly small, but the warpage is also incredibly small. The gear should slide off the shaft with only slight persuasion, but may not go back on so easily. I've no doubt it's well within spec, but I had to press mine back on, though with little effort.

********

If anybody would like me to do this to their basket, please let me know. I was thinking about setting up a jig so as to make perfect alignment possible and reduce warpage. I may also create a reinforcement backing plate for the basket so surface welding can be avoided and virtually eliminate warpage completely.

[Mr. Pockets]

Originally posted by oab_au
Of course the answer is, as I have said, the transfer clutch not responding quickly enought. It is the only way this componet can suffer the server shock loading that will cause it to fracture.

I would think you could easily blame this on a worn transfer clutch going unnoticed or uncared for, combined with the manufacturing problems UberRoo outlines above. Thanks for all the super-detailed info, guys. This is awesome stuff, and what this site is here for.