Hey all,
I have been in the middle of rebuilding my 1995 RX-7 FD differential, changing out the 4.10 pinion and ring gears with a 4.30 set from a later FD model. I am reusing the old bearings and races as they seemed to be in excellent condition, and am replacing everything else.
I've gone to the moon and back to follow the manual to a T as much as possible, made sure to order correct new LSD spacer rings in new sizes after measuring everything by the book (new LSD instead of original), and I even fabricated new pinion spacers at a friend's machine shop to factory precision since you can't buy them anymore, after using the original SST to measure the correct pinion height.
When installing the pinion in the housing following the manual steps for installing the pinion and companion flange for setting your preload (new companion flange nut and washer, new crush sleeve), I tightened it down to 94.1 ft-lbs but the pinion still had a lot of play up and down, a good 1/8" or more.
I continued to increase the torque manually with a long torque wrench in 20ft-lbs increments (the companion flange was braced against the concrete floor with a long bolted on metal bar with the driveshaft bolts and holes), but even after reaching the manual max torque of 210ft-lbs, that pinion STILL had up-down play, like as if I had some ways to go still, and it wasn't like I was going to run out of nut thread on the pinion any time soon.

Is this normal because of the new crush washer? Should I just keep going to crush that washer down properly until the pinion has no play, and then start checking the preload?

My bearing races are sat properly, and I followed every step as correct as possible and I am 100% sure I didn't put anything backwards or anything...

I just wanted to know if I should keep torquing it or breaker-bar'ing it down until that crush washer, well... crushes, taking it slow so that I can set the preload correctly (10-12in-lbs aimed since the bearings are used)

I'd love to know if anyone has any thoughts on this, much appreciated!

 

When installing the pinion, after you set the companion flange on, new washer, and new nut, as you start torquing up to the min amount you should feel the crush washer collapsing. That should be the first indicator that it’s not in correctly.

The variance in the amount of torque on the pinion nut is determined by when you reach the correct preload.

Disassemble it and hold it along side of your original pinion. Also, post some pictures of what you have. I can’t really comment what the issue is without seeing what you have going on.


~ GW

 

I have attached some photos of my whole setup for the pinion gear and its many layers. On some of those you should be able to see how far down I got on the pinion nut and the thread, and how it was sat inside the carrier. I have also attached photos of the stack of parts and how I had them, following the instructions in the manual on how to stack them, and where to put them. The new crush washers and nuts were bought from Atkins Rotary and are legitimate Mazda parts, the pinion is the same, and the bearings are being reused. On one of those pictures I put a new crush washer next to the one I took out of the car originally and there's still ways to go. When I compared the one I was trying to crush, vs. a new one, I barely managed to crush it 1mm, and I was already at 210ft-lbs at that point which is the maximum the manual says to go.
By any rate, despite getting to that point there was still slop on the bearings, AND now I stripped the end threads of the pinion gear when taking the nut out so I feel EXTREMELY fucked.... I'm hoping I can salvage it somehow otherwise I'll have to get another 4.30 gear set and I am not looking forward to that. Sigh.

 

 

Kind of an update on me ******* up the threads on the pinion. I was able to recut them properly with a dye, but they sure are beaten up to hell. I'm thinking of what else I can do for this damned crush washer....
FYI, I tried getting the solid spacer kit from that guy that used to make them, he never replied to any of my emails, so I'm just out of luck there.

 

In your photos I do not see the spacer that goes on the top bearing. This spacer sets the height to the crush sleeve. Is that in there under the crush sleeve? Also did you change the spacer under the pinion bearing? The crush sleeve number for the FD is R004-27-171A. I would confirm that you got the right crush sleeve.

As to your pulled threads. That happened after you continued to torque it. Like I said in my previous post, when you crank the companion flange nut down, you can feel it taking up all the slot. If you still have slop after this, i’m talking like three or 4 foot pounds, you can tell immediately, it’s the incorrect length . Those pulled threads are not great. The side shot of that isn’t good either. It appears they’ve lost roughly 50% of the height. It may not be a total loss though. There is a chance that when you get the pinion in there correctly, the nut will be a lot further down on the thread to where it’s good. You could even go with a thinner washer if that gives you the couple more threads needed. You would have to measure all of that just to make sure. You need like 80% engagement in thread height to be safe. Of course you’ll need a new nut.

My suggestion at this point would be to make sure that the bearing closest to the companion flange actually fits all the way down on the shaft. You can do this by installing the pinion bearing on the shaft outside the housing and see if it will go all the way down. Next I would install the pinion back in the differential housing, install the top bearing and hold it down with your hand. There should be no back-and-forth play. If you have play, then there’s an issue with one of the bearings, not seating all the way down. Or the spacer closest to the teeth. If you have zero play, I would move on to installing the top spacer, new crush sleeve and then the companion flange. I would hold the assembly from inside the differential housing with one hand and push the companion flange down with my other. If I have slop in there, I know there is an issue. If I have no slop, I know I’m good to go , install the washer and new nut and start at my minimum torque. Of course, then you would have remove the companion flange, install the seal and then follow the procedure for setting the preload.


Hopefully how I described it makes sense to you. What you’re doing is checking in stages to see where the space issue is.


~ GW

 

Thanks a lot for all the info, and good ideas as well, I'll be doing just as you described to test everything. I already ordered new nuts as well.

To answer some of your questions:
Yes I have the top bearing washer in place as per manual, its the same that came out of the car originally.
I have the right crush washers, correct parts number.
I fabricated a new space washer for the bottom larger bearing based on manual specifications after doing the SST measurements, in my case its the 3.26mm one.

So my pinion stack, starting from the pinion head, is:
3.26mm spacer washer, main bearing, crush washer, top bearing, top bearing spacer washer, companion flange, washer, nut.

I will test fit the bearings without crush washer to see if they fit properly in the carrier with the pinion.

 

Yes, that’s the correct orientation of the stack. It looks like you understand how it’s supposed to be done. So we just have to figure out where the hitch is. If you had a race that was not totally seated, you would have the inverse to your issue. #36 spacer or 3.26 is middle of the road in size so I don’t see that being the issue either. The top spacer, I can not remember the measurement on it. I will have to dig thru my box of diff parts to see what size that is. Since you just reused the one that was in there I don’t see that being an issue.

If your picture of the parts is how it was installed the crush sleeve is in the correct orientation as well. Off the top of my head, in addition to what I suggested about your initial test, I would also double check that the companion flange can push all of the way down on the splines. If that was to hang up, it could also create the issue you have. I haven’t seen your specific issue personally. I’m kind of interested to see what is causing it.


~ GW

 

Yeah I am at quite a bind as well, beyond the tests you suggested I di, I'm also going to install the stack in the carrier with the original crushed sleeve, since I am re-using the bearings there is a possibility that it will be crushed to the right size already and preload properly without any slop from it or the bearings, at the very least it will be a good testing point as well. If that one is too loose and giving little to no preload at all, but no bearing slop (my guess), then at least I have a starting point, and I can use my hydraulic press to pre-crush 95% of a new crush washer, and then drive that last 5% with the nut to get to the correct preload.
Rather out of ideas otherwise, aside from different accounts in other forums that the crush sleeve sometimes require way more torque to crush, like upwards of 400ft-lbs with the nut, but then again the Mazda manual states to stop at 210ft-lbs so I just don't know anymore, maybe an overly stiff crush washer from factory? Who knows lol
I'll post updates soon!

 

That would be the incorrect way to remove the extra stack height. The crush sleeve is designed to hold the correct preload so the bearing don’t fail. To tight they fail, to loose they sing. You’ll have to buy a new crush sleeve now anyways. They are not reusable. If those other forms are not Mazda specific, they don’t mean squat. You already found out what happens when you over torque it. Just so I’m clear, are you reusing the bearings on the pinion from one diff that that are now being used in another diff that has different races? There’s a reason they say to replace the bearings and races in those.


~ GW

 

I have about 4 other correct crush sleeves to use, so no problems there. and no the carrier, bearings and races are all the same, the races for the pinion have stayed in the carrier and not been removed, and they are properly sat, I checked.
The only things I have changed are the new pinion and ring gears, which are a 4.30 set instead of the original 4.10, and the original torsen diff has been replaced with a 1.5 way clutch one from Kaaz.
The LSD side bearings, and the pinion bearings, are being reused.
All measurements have been made with those bearings in mind, and all spacers have been acquired or made with those bearings in mind.
The carrier is the same one, no changes.

 

Got it. That sounds fine. I have put together and currently run that setup.



So I missed this when reading this post. This is a possibility as well. Since you pre crush this when initially torquing to the 95 foot pounds. You can feel this or should I say should feel the crush sleeve collapsing, just like when you install a new spark plug. After checking everything out it’s possible that you don’t find anything and install a new crush sleeve then it goes together nicely.

~ GW

 

Yeah when I checked the crush washer I used vs the one I took out of the diff, they were exactly the same size and the same thickness of material, so its really odd to me that it would be SO stiff. That said I have three others to use so I will do a few checks and see what i get.
When I was torquing this one I could slightly feel it collapsing, but by the time I got to 210ft-lbs it only collapsed about 0.5mm at most when I compared it with a new one. Very crazy stiff for being an original parts acurate mazda part.

 

Hello friends, I finally have some updates to give, so I will attempt to be as detailed as possible;

I'm going to preface by explaining a few very important elements of this build with some context on parts used or re-used, so here it is:
- The original bearings and bearing races were re-used.
- The original diff housing was re-used.
- The original companion flange was re-used.
- The original pinion spacer shim, and smaller bearing spacer/washer were re-used.
- The companion flange nut and washer are brand new.
- The LSD is brand new.
- The 4.30 pinion and ring gears are brand new.
- All oil seals are brand new.
- All steps were followed with the original Mazda Service Manual for the FD3S, nothing was done "***** nilly", and the aim was to keep within procedures and within manual specs at all times, no compromises.

With all of that said, let me explain my entire work here, following from the rest of this post above;

The 4.30 ring and pinion gears are Mazda originals and the exact same sizes and dimensions as the original 4.10 set that came out of my differential, and I mean EXACT same important dimensions throughout. Knowing all of that, and the fact that I am re-using the original bearings and races, the thought came to mind that since I'm re-using all of the original important pieces, and the gears are so exact in size to the originals, there is a high likelihood chance that my pinion spacer is already correct, and now what I wanted to test was could I get the correct (consistent) preload on the pinion by re-using the original crush washer that came out of the assembly. My main plan here was to use that as a test: either the original crush spacer would be good to be reused, or it'd be too crushed already and I'd need to replace it, BUT it'd allow me to know how much crushing a new one would need to sit within the factory torque specifications for the pinion preload, which would allow me to pre-crush it close to that using my hydraulic press, and do the rest of the crushing with the pinion nut.
So, following that strategy and knowing the factory parameters for torque and preload:
Preload: 12-15 in-lbs (Aimed at 10 in-lbs due to used bearings)
Nut Torque: 94.1 ft-lbs to 209.7 ft-lbs
I proceeded with the pinion job. Once again: New pinion, original bearings, original space washers, original crush spacer.
This are the results:
By the time I reached the low end of 94.1 ft-lbs, the bearings had NO play, but still spun freely with no preload. So we continued torquing in increments of 10 ft-lbs at a time, slow and steady. Once I reached 140 ft-lbs I could feel the preload starting, nothing still had play, and I could feel that I was driving into the crush spacer properly, and preload was rising gradually little by little, which gave me confidence that yes, the crush spacer was behaving as intended. This was continued until i reached 170 ft-lbs and hit my target of 10 in-lbs of preload. Everything was checked and double checked, the pinion was spun about several times, everything I could think of, and everything remained within specs, all tools used were the best I could find on the market, nothing was cheapened out on those.
So I followed up from there on the manual steps, removing the companion flange, adding the oil seal, and repeating the process back to 170ft-lbs of torque. Preload was once again checked, and we were still at 10 in-lbs right where it needed to be. Perfect!
We ended up with a best case scenario on this, being; The crush spacer must have been crushed down to a lower torque spec at factory, which left room for it to be crushed down further and reach preload correctly. My lesson from this is to first try the original crush spacer and see where you land, and worst case scenario you will end up with a ballpark on how much it needs to be crushed by with a new one (As in, you'd crush a new spacer not to the height of the original, but leave a few mm that you crush with the pinion nut as per instructions, you're just pre-crushing it so you're not driving the nut so hard against it to crush from brand new size)

Moving on from that, it was time for the carrier and its assembly to go in, and check backlash and gear tooth engagement patterns;
Backlash should be within 0.09mm to 0.11mm at four quarter sides of the ring gear
Tooth engagement should be dead center between Root and Crown, and center of gear tooth from Heel to Toe.
The original carrier bearings were pressed in, and its races held in place before mounting it into the differential. And yes I was VERY careful to ensure that left and right side bearings, races, spacers and bearing caps were kept together with their respective side, nothing was mixed up.
I started with the original carrier spacer shims in place, in my case they were 6.00mm left ( ring gear side) and 6.15mm right (pinion side).
After hand tightening the bearing caps, I used a precision digital backlash gauge and followed the steps in the manual for checking at 4 different quarters of the ring gear. The initial backlash was measured at 0.04mm, not ideal. But I went ahead and used prussian blue machine ink to check where my contact pattern was at, driving the pinion with a power tool 5-6 times forward, and 5-6 times in reverse, all while adding load to the LSD using my hand and a rag. Again, pattern was checked under load; the tooth engagement was dead center between Root and Crown, with no run-off to deep or too shallow, meaning the pinion height was exactly where it needed to be, a good success! And contact was very large while biased to Toe on drive, and Heel on coast, which aligned with the backlash being too tight.
So I decided to only adjust the backlash, by changing the carrier spacer shims to move the assembly away from the pinion.
I moved forward with a small adjustment by changing the left spacer to 5.95mm, and the right spacer to 6.20mm, a total adjustment of 0.10mm, small changes.
Once again all the steps were followed and backlash was checked again, at four quarters of the ring gear as before. Now our backlash was at 0.09mm at all four points, exactly what I wanted, and within the factory manual specs!
Gear pattern was once again checked under load, same steps as before with the power tool; Drive engagement was still dead center Root to Crown, and moved a bit towards center between Heel and Toe, while still being more towards Toe, but with plenty of space between pattern and Toe's edge, and NO run-off. Coast pattern was dead center in all regards, no run-offs of any kind, you can check the final patterns on the photos below.
These patterns are perfect for a new set of gears and exactly what i was after, while every other parameter was well within factory specifications, preload, pinion nut torque, backlash.

So to me, I consider this a great success in the end, nothing was left at "good-enough" and the correct steps and specs were taken so everything is within what the Mazda manuals told me to have them in. I will be finishing this differential build now and the rest of my car build, and the goal will be to run the vehicle for 300-500 miles, then open the diff once more and check where everything stands, how's the preload, how's the gear pattern, any metal shavings present, everything, nothing left to chance.

And there you have it, I am quite satisfied with the results, and I hope this whole ordeal helps someone in the future if they end up going the same route as me on how I approached this diff rebuild and upgrade.

If you have any questions let me know!
I will post updates if any in the future once everything is up and running, might be several months.

Edit PS: The coast pattern looks much better in person than on the photos, it was just difficult taking a good picture of it, but I am confident in it.


Diff prepped with the correct carrier spacers, right before placing the bearing caps.
[img alt="Drive side pattern.
Nicely centered in height, biased towards toe, but with plenty of toe-end space, no run-off."]https://cimg6.ibsrv.net/gimg/www.rx7club.com-vbulletin/2000x1504/photo_2026_06_01_16_27_57_2__4e62e8e5dc0d8a418a53c 5da6b5a1ea0659cd297.jpg[/img]
Drive side pattern. Nicely centered in height, biased towards toe, but with plenty of toe-end space, no run-off.
[img alt="Coast side pattern.
Nicely centered in height and length, no run-offs."]https://cimg7.ibsrv.net/gimg/www.rx7club.com-vbulletin/2000x1504/photo_2026_06_01_16_28_00_2__51e52c2a7715d5dde2fd1 2ba62ec246dfd5984a4.jpg[/img]
Coast side pattern. Nicely centered in height and length, no run-offs.

 

Drive side Pattern.

Coast side Pattern.

 

Glad you got it all dialed in. Your process sounds proper. I don't necessarily agree with pre crushing the pinion spacer on a press. I have not ever had an issue following the Mazda procedure for seating the pinion. I have however felt differences in how each crush washer feels during installation. Some are very tight at first, but once they begin to deform the behave similarly. If you aren't experiencing any noise or issues after installation, I don't see a need to re open the housing for post installation inspection. I would maybe drain the fluid and check for metal, but other than that, not much to worry about after installation.