352 transaxle finned nut & backlash adjustment

[SKlein]

I am about to install my rebuilt rear suspension on my TCS. In prep for that, I removed the finned nuts to renew the lip seals & o-rings (8.3 turns of the driver/crownwheel side nut to remove). All seemed ok on that side.
The passenger/diff side output shaft had what seemed like excess end float that concerned me. Since all the suspension was already detached, I decided to remove the trans & detach the bell housing to investigate. I removed the finned nuts & replaced the lip seals. The bearings are as installed.

I have fiddled with the finned nut adjustment a bunch of times and things aren't right. There is excessive crown wheel/pinion backlash & a lot of output shaft end float with the diff turning freely. In my first measured try, I started with both finned nuts bottomed against the bearings finger tight (FT) & marked. I then tightened the diff-side nut 1.5 fins past FT at 35 ft-lb torque. I tightened the crown wheel-side (CWS) nut to 1.3 fins past FT at 30 ft-lb torque. At this nut setting, the force to turn the diff was 13 lb where 2-7 lb range is specified on p.24 of the shop manual. I then backed off & retightened the CWS nut to 20 ft-lb, putting the fin position about the same as at FT. At this nut setting the force to turn the diff is 6 lb. The crown wheel/pinion backlash is 0.020" where .005-.010 is required.  The end float of the output shafts is either .010 or .014" depending on whether the opposite shaft is pushed in or out. There's no spec for output shaft end float. On various other occasions, the best backlash number I got was .016". Fiddling with the finned nut adjustment only seems to affect the free play turning the diff.

It seems to me that I have to split the case & re-shim the primary shaft to reduce the pinion backlash, & change the washer in the diff to reduce the end float. And maybe press the bearings a little out so they re-seat at reassembly. But to measure the pinion depth requires a dummy mandrel & spacer & I don't know the dimensions of those to duplicate them. The car has 87K miles on it.

In other words, I'm stuck. Do I leave the current measurements alone, or do I start from scratch & rebuild maybe the whole trans? I thought I saw an outfit on the web selling Renault trans parts, but now I can't find it.
Any ideas out there besides the scrap heap or a swap to a 5-speed & replace the whole back of the car?

Thanks,
Steve

[Kendo]

Do NOT scrap. Everything else has compromises. I haven’t done a trans, but we have experts here who will tell you what needs doing. And, I will eagerly read to know for next time.
(Personally, I smoked my 365’s fifth gear from starving it for oil. Back in 1990’s, so I got a replacement. But I’m very cautious now.)

[SwiftDB4]

I would just go with what you have. Backlash is very close as is. Don't worry about output shaft end float. What you have is not excessive. I remember years ago I raced my S1 and a recommendation from the Jensen Motors manual was to grind a washer to take up all end float so the open diff acted somewhat like a limited slip. For a street Europa that is not necessary. For peace of mind check the gear mesh pattern with Prussian blue or equivalent.

[jbcollier]

Ring/pinion backlash is adjusted by the finned nuts.  Diff carrier bearing preload spec you quoted for NEW bearings.  With old bearings, the preloaded is set to zero.  True zero is hard to assess so I go with a tiny amount of pre-load so I know I'm not too loose.  You need to set your preload to spec and then shift the ring gear sideways by turning each nut equally until the freeplay is within spec.  Measure in three different places/rotations.  For the record 20 thou is way too much if the spec is in the 7 thou range!

Output shaft play is adjusted by shimming the internal diff gears.  Nothing to do with the finned nuts.

The pinion shaft is NOT the input shaft.

The pinion and input shaft endplay is controlled by a double sided, tapered roller bearings under the rear cover of the transaxle.  ANY play in the pinion shaft is unacceptable, full stop.  The input shaft is two-piece and the section poking out is on a spline and will be "loose".  That is normal.

[314159td]

But to measure the pinion depth requires a dummy mandrel & spacer & I don't know the dimensions of those to duplicate them.
...
I thought I saw an outfit on the web selling Renault trans parts, but now I can't find it.

I was actually just firing up Solidworks to make the dummy mandrel, as I'm tasked with the same job in a few days here on my 336 transaxle. I think the Renault manual has enough dimensions to make it work iirc, checking now. If it goes well I'll try to make a few, already made most of the custom tools required for disassembly. Seems like I could be a little sneaky and make it in 3 simpler parts rather than using a big piece of round. 

Renault 16 shop had a decent number of transaxle parts, and I'm sure if you ask around you can find some people with spare adjustment washers floating around. Depending on how you need to adjust, off the shelf sized shims can be added. I bought one of each of the thinnest washers, in anticipation of shimming them where possible rather than reordering just to get the right washer.

[SKlein]

Yeah, I wrote primary shaft but meant pinion shaft.

I fiddled some more: I backed off the diff-side nut & tightened the other to push the ring gear toward the pinion. I got it to where there is little to no perceptible backlash. There is also no perceptible output shaft end float, but the force to turn the diff is 10lb, should be 2-7lb. So I think it's just a matter of iterating both nuts back & forth to move the ring gear bearing to the right (sweet) spot & then the backlash, force to turn, & end float should all be ok.

I'll let you know the final result but I think all is well...

Thanks Guys,
Steve

[jbcollier]

The early transaxles require shimming, the later ones do not.  Renault got really, really good with their machining and all their pinions used the same shim.  I used to remember what size it was.  Been awhile though.

I'm in the middle of fitting a 352 ring and pinion to a 365 gear set and into a set of 395 cases: 3.56 final drive/evenly spaced steps between gears (except 5th)/easy shift design for a cable shifter.  I had the 352 pinion machined to accept the splined extension for the 5th gear.  Chap in England did it.  He asked for the measurement of the 352 pinion shim and it was exactly as he expected.  He went on to explain that all the newer 3 series transaxles used the same shim.

[SKlein]

I concentrated on the ring gear side nut while backing off the diff side nut & got the bearing in the right position to give .005" ring/pinion backlash - just right. Then adjusted the diff nut to just past finger tight. The force to turn the diff housing is about 1.5 lb. Pretty much just right. There's still .015" end float on the diff side output shaft, but I'm gonna ignore it. Time to button it up.
It took me quite awhile to understand what the manual was telling me to do.
Thanks everybody for your input.
-Steve

[314159td]

As this is the most recent conversation on the topic;

My pinion height (E) is 26.8mm for whatever reason, vs the 26.4 specified in the manual. There are no markings to indicate an offset, and 0.4mm is quite large. That's what motivated me to make the measurement tooling in the first place, and take measurements with the original setup. Then again, per the Renault manual this configuration of pinion depth and tooth count doesn't exist so there may be other weirdness with the Lotus-specific ring and pinion set.

With the mandrel built the same as the factory tool, I'm getting approximately 52.9mm as the pinion depth with both the original and new double taper roller bearing, both with the same pinion depth adjusting washer. This is of course, 0.1mm too "tight" to the ring gear. I have on hand a pinion adjusting washer that would space the pinion back about that amount to the correct 53mm.

Considering that the differential has already worn into this spacing, and appears to have a good wear pattern, I think it makes sense to just keep everything as-is? The side-to-side backlash adjustment should be able to compensate for some amount of wear, and prevent things from being too tight leading to premature wear.

Manual section for those following along at home, page 28

If anybody needs this mandrel tooling, send me a DM and I can send it to you to borrow. It's not perfect (my lathe cuts on a bit of a taper and the tailstock is slightly off) but with a bit of care I was able to get consistent measurements that make sense. That's what those sharpie lines are, making sure the inner hole is centered to the bearing bores and not offset front or back. 
 

[TurboFource]

Is yours an “exceptional case” that it mentions?

[jbcollier]

Mount the ring gear and set its backlash and preload.  Then use Prussian Blue to do a contact pattern check.  If the pattern is good, no issues.  If not, use online pattern guides to direct adjustments.

[314159td]

Is yours an “exceptional case” that it mentions?

Nope, no markings. You can see the top of the pinion in the photo of the installed mandrel, just two scribbles that also match the scribbles on the ring gear.

Technically, the pinion "thickness" should have very little to do with the pinion depth due to how the bearings work and the gear geometry, and I was vaguely concerned before taking a measurement with the original setup.

Prussian blue is an idea, I think I have some of the yellow stuff too.