Great writeup! You referenced the RX-8 4.77 gear set...any idea how this would compare to the Re-Amemiya 4.77 set? I'm thinking of having this done, too. Based on your experience, would this install present a good tranny/diff shop any problems since it's for an FD compared to any other car ? And, lastly, can we get your driving impressions with the 4.77 gears ?

thanks

I have never seen the Re-Amemiya gearset.

Were I to purchase a new one, personally I would buy the 4.77 RX-8 gears because I know they are made to fit the diff, and that that diff is interchangeable with the FD diff. Also due to the fact that the MX-5, RX-8 and FD RX-7 all share the same part numbers for the shims, I would be pretty confident that the fit would be good.

With my gear set I am bumping the boundary of the shim sizes on one , with the thickest possible shim on the left side required to get the backlash right (re-read the thread if you like, but I think I spoke correctly here)


Driving impressions are for autox only - no track duty yet.

4.77 gears are pretty much required for autox. Based on my results against a dual driver C5 Z06 that I used to pretty much run neck and neck with, I can now beat them by .75 second.

There is no more hesitation coming out of a slow corner in 2nd gear.

I now have to be careful if I get a good launch out of the final corner because I now may run out of room to brake at the finish. That never happened before the 4.77 gears.

if you are planning on starting a workout, specifically toning your biceps, triceps, pectorals and any muscle above your waist. then this job is right for you.

have you done this install as well? Just how tough was it? Looks like murder...!

Take your time, get a tranny jack, build a simple stand for mounting the diff while working on it - it's not as bad as it looks.

Not sure if this matters or not. But, using the RX-8 4.77 gear set, could this be done without a press?

The pinion gear and rear bearing will require a press to remove and replace.

If the diff has never been out of the car, you will probably need to take the shaft to a shop with a hydraulic press for removal. The Harbor Freight bottle jack presses are probably not strong enough to remove them.

I didnt read the writeup, but its a pita. the hardest part is that the dif tends to want to rotate, causing the fluid to spill, and its heavy. get a jack to do most of the work. but you will have to to wiggle the dif into place at the end. drain it before you install it and then fill it after its in, thatll save you maybe 6 lbs and a mess all in your hair. stinky, heavy. you want to feel closer to your car, this is a good way to get to know her.

nice. great info

Really, filling the diff first is not that hard.
I did it twice without issue.

That is assuming you have a transmission jack.
I wouldn't do this job without the jack.

its not hard. it just tends to spill (unless you have a trans jack, which most dont) you can use a beefy floor jack but the dif still may tend to roll side to side and spill.


edit- unless you have the axles in ans just hunkered down.

Diff rolling side to side?

If your fingers aren't worth $150 for a tranny jack...

I've probably stated it in the writeup - if you're replacing the diff, get a tranny jack.
Doing the job without one is risky.

i did it without one. but yea. this is a good job for you and a friend if you have one. atleast so you can slip something under it besides your fingers if you need to take a break. i wish i had this writeup before, lol.

I can get some cheap fingers from the third world, PM me.

Thanks Op for making life easier!

My axle's back in the diff but i did it a little differently. I got the halfshaft in about halfway, then whacked the other end of the axle that's bolted into the wheel hub. After a few solid whacks, everything seems to be in place.

However, I didn't hear anything snap, maybe it's because the hammer was too loud, so I went under the car and started just moving the axle around. It looks like there's about 0.5 mm lateral play (moving out toward the wheel) as I'm turning the axle stub and just trying to move it around.

Is this normal? or did i mess something up?

I didn't measure the play, though if you look in the FSM it may show the allowable play.

If you can't pull the axle out of the diff it is probably OK.

I just put my diff back in, and I did what the OP did, give it a couple hard shove until I can hear and feel the axle snap in place. You have to position the C clip with the opening at 12 o'clock, not sure if that'll makes it easier to install but that's what the FSM said.

You should be able to see if the axle is all the way in from underneath, it should press against the seal pretty good, here's how mine look.

http://i133.photobucket.com/albums/q...ps58af6712.jpg

Thanks guys, I did some more moving around, and tried to put some pressure with a pry bar and it's really in there.

Cool, thats pretty much what mine looks like. The protrusions of the seal are going into this groove area of the axle so it looks right.

One of the more difficult parts of this differential adjustment is the measurement of the width of the torsen and carrier bearing unit, so as to determine the total width of carrier shims required (value "C" in the FSM). This is difficult because it is very hard to measure to within ~3 thou tolerance across two floating roller bearing races and a heavy torsen. "C" is important to get right, because it sets the carrier bearing preload.

If you have already purchased a range of carrier shims, it is my experience that the correct value for "C" is about 0.10 mm greater than the total shim width that can be pushed into place by hand. In other words, adjust shim width to find the greatest value that you can push into place by hand (with some difficulty, by rocking the torsen unit). Then add 0.10 mm to this value - that's the correct "C" for your assembly.

If you have not yet disassembled the factory diff and have a range of carrier shims on hand, you can test this yourself - pull out the shims selected by the factory and start putting successively thinner ones in place until you find the width that can be assembled by hand. I've done this test on two different factory assembled diffs, and there was a 0.10 difference (two shim sizes) on both.

Note if you are planning to do this test yourself, you will want to purchase a range of shims before making any changes to the factory assembly (e.g. installing new carrier bearings) that would make it hard to reproduce the factory-assembled state.

Some notes on parts compatibility:

The FD manual and automatic diffs have different (and incompatible) drive shaft flanges. The bolt pattern is wider on the auto, and the auto doesn't have a vibration dampener.

The RX8 differentials don't just have compatible pinions and ring gears - they also have exactly the same internals (drive shaft flange, bearings, shims etc) as the FD. Only the case is different, and some RX8 models don't have a torsen unit. The RX8 automatic diff has a bolt compatible drive shaft flange (at least the one I looked at did), but without the vibration dampener that comes on the manual flange.

Some notes on shim sizes if you are swapping OEM Mazda pinion and ring gears around (these comments almost certainly do NOT apply if you aren't using Mazda parts!):

In fixing my worn-out pinion (BTW, it really HOWLS when it goes, enough that I thought the diff was going to dump all its oil on the track), I ended up buying one FD and two RX8 used differentials, as I learned the differences between auto and manual diffs, and changed my mind several times about what ratio I wanted.

All the carrier shims on these diffs were in the range 5.95 mm to 6.15 mm.
Three of the pinion shims were 3.11 mm. And the other one I didn't inspect because I didn't need to change it.

Four (my original FD diff, and the three I bought) is a pretty small sample size, but I think it might be significant that all these diffs, spanning almost two decades of manufacturing, were in a relatively narrow range of shims - Japanese manufacturing is, after all, legendary for tight process control.

- If I were doing this again, rather than buying one of every carrier shim available, I would have been better off buying two each of those between 5.95 and 6.15 mm. A diff may very well need two of the same size shim (two of mine had identical sized carrier shims on left and right).

- If you are using used parts (say, from RX8 diffs) you might find that the backlash of that ebay diff is no longer in spec (since the pinion and ring gear wear). In my experience, bringing it back into spec meant moving the ring gear inward one shim size, possibly necessitating moving the pinion out one pinion shim size to maintain a good mesh pattern. Since all of my diffs used the 3.11 shim, there is only one thinner one - the 3.08mm shim. So I'd buy just the 3.08 and 3.11 pinion shims to start.

Many, many thanks to jkstill for such an awesome write-up. I would never have attempted to fix my own diff if I hadn't seen this.

A note on pinion preload
 

It is worth also reading the FSM instructions (pages M25-M26) for simple replacement of the differential companion flange oil seal, as it clarifies a few things about the process of setting preload.

These instructions make a little more clear that the pinion preload is a starting torque, not a continuous rotating torque.

Tapered roller bearings will have a discrete starting torque when preloaded, due to the design of the rollers. Examination of the inner race will show a ledge that holds the tapered roller in place even under a thrust load. The back edge of each roller bearing is designed to slide, as opposed to roll, on this ledge. When the bearing is preloaded, there is a discrete starting torque required to break this sliding contact point free.

Also note that the instructions for simple seal replacement do not specify what the preload should be, only that one should measure and match what was there already. Given that all the used diffs I checked had a MUCH lower starting torque prior to disassembly, it is possible the starting torque spec in the diff disassembly/assembly section of the FSM applies only when installing new bearings.

Is most of the advice consistent with Mazda differentials only? Or are most differentials set up the same way with the same quirks?
If you take your diff to the shop for new bearing install, would they need this info? Or would they know this by default because car differentials are sufficiently similar across model lines?

The basic process is similar for all pinion-and-ring differentials. The exact specifications (backlash, preload, etc.) are specific to the details of the diff design, but the overall process for setting those specifications are similar. Note several of jkstill's references are written for other manufacturer's diffs, but do a great job describing the basic process.

Yes, you would want to ensure a professional diff house was very familiar with the content in the FSM.

Just doing another diff, and wanted to add a couple things to this thread:

1) That method I mentioned in the post of 6/23/17 of seeing what shims can be pressed in by hand is actually just an easy way to measure what manual calls "A-B" (the gap between case width and width between the carrier bearings). I didn't realize this until just today. Duh.

2) I think the service manual has an error in the calculation of "C"(the sum of the widths of the two carrier shims). As jkstill noted, "The math for determining the shim sizes is pretty simple and in the manual. One funny thing about it is that you are supposed to add .002-3" to the diff width to determine the shim sizes. It's funny because the shims only come in .5mm (~0.016") increments." Page M-39 of the manual says to add between .01mm and .03mm to "A-B" to get C. But the shims only come in increments of .05mm as noted. I think that 0.01 to 0.03mm in the manual should actually read 0.1-0.3mm. This would explain why on the differentials I've disassembled from the factory, C has always been 0.1mm to 0.15mm greater than "A-B".