Diff contact pattern help

[twistinchassie]

Hey all,

Currently rebuilding the rear diff center out of my Hdj80. I’m currently on the smallest pinion shim I can get from Toyota 1.05mm
My back lash is 0.17mm
The gears are used and I’m only rebuilding as pinion seal was leaking and decided to put a solid spacer and new bearings whilst I had it apart.

does this contact pattern look ok? It thinking the only way to correct the pattern on the coast is to move the pinion further away from the ring gear. Ie surface grind the shim smaller?

Your Thoughts on contact pattern would be great.

Cheers
James

 

[Malleus]

Although contact pattern and backlash are both important, they really aren't directly related (I'm going to get opinions for that, but that's OK). They're just in the same place at the same time.

The contact pattern is the result of the positional relationship between the ideal gear cones, and the backlash (with respect to the pattern) results in the pattern size. That's a gross oversimplification, but let's start there. You need to get the pattern correct while the backlash is in spec, but you have to set the backlash first, because it's measurable and the pattern isn't. You'll only know what the pattern looks like, after you've assembled the set.

FWIW, the backlash is the only thing measured in the factory. Random samples are printed to check the backlash; not every set gets printed.

I assume you've mounted the gears in the carrier, torqued to spec, and that the case is rock solid in the vise. Any wiggle anywhere invalidates this procedure.

First, the caveats: you will not get a used gear to print the same as a new gear. Typically, the patterns fall to the root and toe as the gear wears, but that's only for properly (initially) set gears and those which have normal wear. If there's been a long term bearing wear problem, that likely will not be the case.

The direct answer to your question is: The drive side pattern (the second photo) is poor; the coast side pattern (the first photo) is unacceptable. Do not mount that gear.

One thing that bothers me is the significant difference between the lengths of the patterns. There is a serious angle problem between the centerlines. I suspect it's due to bearing mount position, but that's a WAG.

My guess, and again this is a serious WAG, because I'm not there, is that the gears are probably OK, just based on the little wear I can see from the photos. However, you can't mount those gears in that setup. If you're having that much trouble getting the backlash set, the pinion bearing's in the wrong place. There's a reason Toyota only offers a finite range of shims, the carrier is machined to set the cones relative to each other, you can't change that.

My first question, before you get any farther into this, is what did the gears sound like when you were driving? My second is what size shim was in there when you pulled it apart? Lastly, did you try to print the gears, BEFORE you took everything apart?

Need to answer these questions before you get any farther. FWIW, I used to build these, new. When it works, they come out right on target. When it doesn't, you can chase it for hours before you find the problem, if you don't follow the build procedure properly.

Finally, if it wasn't screaming before you took it apart, it's likely OK. You just have it stacked improperly. Sometimes it helps to visualize what you're doing by drawing two 45º cones, at right angles to each other, and think about moving the centerlines, in/out and at angles other than 90º, and the resulting contact between the cones. This works regardless of face angles of the gears, it's just a visualization tool.

Paper doll cones work, too.

---

Maybe this will help you understand why you need to do what you're trying to do:

As you can see in the lower figure, the contact path runs from the root, on the toe side, to the crown, on the heel side. The best position for the contact pattern is in the middle both ways. If it's too close to the crown, where the tooth is the weakest, you risk structural failure. At either extreme end, you quickly run out of room for the contact pattern, as the gear wears.

Although the coast side is shown here, the pattern is the same for both sides, as is the target contact area.

 

[Saito]

I’m setting up my rear w/ 4.88s right now. Your backlash is just at the minimum spec but that coast pattern is way too far down at the toe and I wouldn’t run it.

Once I got mine all reassembled and my pinion preload set, backlash was stupid tight at .002 . . .so I shimmed the carrier away from the pinion as much as possible and the most I could get was a solid .004 at two opposing points. Only two fat shims on each side from the factory. LH one is at a local machine shop so they can surface grind .002 off then I should be golden.

I reused my old pinion shim and it was a fatty. Didn’t measure it though. How big was yours ? I’d try putting that back in before anything else.

 

[twistinchassie]

Although contact pattern and backlash are both important, they really aren't directly related (I'm going to get opinions for that, but that's OK). They're just in the same place at the same time.

The contact pattern is the result of the positional relationship between the ideal gear cones, and the backlash (with respect to the pattern) results in the pattern size. That's a gross oversimplification, but let's start there. You need to get the pattern correct while the backlash is in spec, but you have to set the backlash first, because it's measurable and the pattern isn't. You'll only know what the pattern looks like, after you've assembled the set.

FWIW, the backlash is the only thing measured in the factory. Random samples are printed to check the backlash; not every set gets printed.

I assume you've mounted the gears in the carrier, torqued to spec, and that the case is rock solid in the vise. Any wiggle anywhere invalidates this procedure.

First, the caveats: you will not get a used gear to print the same as a new gear. Typically, the patterns fall to the root and toe as the gear wears, but that's only for properly (initially) set gears and those which have normal wear. If there's been a long term bearing wear problem, that likely will not be the case.

The direct answer to your question is: The drive side pattern (the second photo) is poor; the coast side pattern (the first photo) is unacceptable. Do not mount that gear.

One thing that bothers me is the significant difference between the lengths of the patterns. There is a serious angle problem between the centerlines. I suspect it's due to bearing mount position, but that's a WAG.

My guess, and again this is a serious WAG, because I'm not there, is that the gears are probably OK, just based on the little wear I can see from the photos. However, you can't mount those gears in that setup. If you're having that much trouble getting the backlash set, the pinion bearing's in the wrong place. There's a reason Toyota only offers a finite range of shims, the carrier is machined to set the cones relative to each other, you can't change that.

My first question, before you get any farther into this, is what did the gears sound like when you were driving? My second is what size shim was in there when you pulled it apart? Lastly, did you try to print the gears, BEFORE you took everything apart?

Need to answer these questions before you get any farther. FWIW, I used to build these, new. When it works, they come out right on target. When it doesn't, you can chase it for hours before you find the problem, if you don't follow the build procedure properly.

Finally, if it wasn't screaming before you took it apart, it's likely OK. You just have it stacked improperly. Sometimes it helps to visualize what you're doing by drawing two 45º cones, at right angles to each other, and think about moving the centerlines, in/out and at angles other than 90º, and the resulting contact between the cones. This works regardless of face angles of the gears, it's just a visualization tool.

Paper doll cones work, too.

---

Maybe this will help you understand why you need to do what you're trying to do:
View attachment 3288718

As you can see in the lower figure, the contact path runs from the root, on the toe side, to the crown, on the heel side. The best position for the contact pattern is in the middle both ways. If it's too close to the crown, where the tooth is the weakest, you risk structural failure. At either extreme end, you quickly run out of room for the contact pattern, as the gear wears.

Although the coast side is shown here, the pattern is the same for both sides, as is the target contact area.

Firstly thank you for taking the time for such a detailed response I appreciate it!

The gears are what I would consider good/ great condition as far as surface condition. upon visual inspection no major signs of excessive wear across any on the contact surface or root of tooth.

There was no noise upon acceleration or deceleration that warranted the strip down it was purely due to the leaking pinion seal.

Due to my haste in trying to get it stripped down I didn’t take a pattern sample for reference or check note the back lash however it was greater than 0.2 by feel.
(I am a fitter/machinest by trade so a somewhat educated observation lol)

The shim that was removed was 1.15mm thick.

The gear set I believe to be original, fitted to an early style arb air locker.

Where my initial problem with pinion depth may be coming from is the inner bearing I removed was a KOYO st4090-n
This particular bearing appears to be NLA and has been replaced with a koyo 30308Djr as supplied by multiple drive line rebuild kits in my care drive tech here in Aus.

Dimensionally the bearing is the same except for cone number and angle. I measured the heights of both bearings on a surface plate with a vernier height gauge and they came in the same.

One would assume with all things being equal I would have the same pattern and back lash upon re assembly. Unfortunately I don’t have enough data pre disassembly.

My initial instal with the new pinion bearings and original 1.15 shim gave me a worse off contact pattern towards the toe which lead me to think the pinion had to go further away from centerline and corrected the lash by adjusting the carrier position.
The result is as posted above.

I’m definently having trouble getting my head around why I have such a heavy toe on the coast compared to the drive side?

Fwiw I printed the pattern on 3 roughly even spaces around the ring gear all repeating the above.

Thanks Again.

 

[Malleus]

Not to worry about lack of data before teardown. It happens all the time. Sometimes it helps, sometimes it doesn't.

In theory, toe contact means you have the pinion too far "in". That assumes that all the other geometry is correct. Due to manufacturing tolerances, the drive and coast sides never line up, except when the set is correct. However, having the drive side in what I would consider a "fixable" position (using shims, that is), and the coast side so bad, leads me to believe the angle between the two is off.

If you can spin the gearset, you should feel it tight at one point and loose 180º out from that, if the angle is off. If you can't spin it at all, something else is wrong.

Since you're a machinist, I don't have to convince you that machining the carrier or shims is a bad idea. You already understand why. Your only recourse now is to do what I had to do when my line assemblers called me out of my office to show me one of these: start from scratch and build the carrier as though you are putting it together for the first time. At this point, you're chasing ghosts.

So, disassemble everything, and I mean everything. Make sure everything is clean, and bolts are properly torqued. Lock everything tight in whatever you're using to print the gears. Remember, we're talking about thousandths here, so every little bit matters. A thousandth here or there will more a pattern ridiculously far.

Consider getting a new bearing; you probably know the tapered rollers sets compress when installed properly so the loose stack height isn't a reliable measurement. Also, the cone angle does matter, since the compressed height moves at a different rate when loaded (high angles move faster than low ones). It might be best to get the new bearing from Toyota, unless you know for certain the number your supplier gave you is the same, or interchangeable with, the one Toyota would have supplied.

 

[flintknapper]

I’m definently having trouble getting my head around why I have such a heavy toe on the coast compared to the drive side?

It is not uncommon (indeed expected) to see a lower toe pattern on the coast side. But yours is REALLY low.

I'm with @Malleus......on this one. Start over, pay close attention to all aspects. Strive to get the backlash right then worry about the pattern.

Below is not a Toyota diff....but a Dana 44 I set up. You can see both the drive and coast sides. Note that the patterns (especially on the coast side) are nothing alike....yet this diff is solid and quiet as a mouse.

Typically setting up differential gears (optimally) is a tedious and time consuming ordeal. Just be prepared for that. Small changes in one area can have a big effect.

 

[Malleus]

I forgot how tiny those things are.

 

[use2bn83]

Rarely do I find well used 9.5’s or 8’s in spec for backlash when I take them apart.

As you adjust backlash, the pattern will move along the pitch line.

Large pinion shim and BL swings in used gears can cause pretty wild patterns, and that’s if the pattern looked good from the start (as mentioned above ).

I’d open backlash up to about .011-.012” or .254mm-.305mm and see if that moves the pattern closer to the center of heel and toe. this may also make your pattern appear less deep due to it being in the “groove” if you will.

 

[Malleus]

Fun fact: did you know that ASTM used the Dana 44 as the basis for the axle standard?

 

[flintknapper]

Fun fact: did you know that ASTM used the Dana 44 as the basis for the axle standard?

Did not know that.....but not surprised, it (Dana 44) being so ubiquitous.

 

[twistinchassie]

So I ordered the original pinion bearing and stripped everything. Stoned all mou ting faces. Re assembled with the 1.15 shim I intially removed and still have the same pattern? Set up as per FSM with 0.18 back lash. No binding across 360 rotation or change in loading.

I’m stumped

 

[Malleus]

Try setting it up with half the ideal backlash and see what the pattern does. At this point, you're chasing a lot of worn parts and some new ones trying to get their act together. You can get away with tighter backlash on a worn gearset. I suggest halving the backlash only because you know you have to have more than zero, and you want to be at the mean (ideally).

If the pattern moves in the right direction, but isn't where it should be, halve it again, closer to 0.018-in and see what that does. This is not a "standard" procedure, and you won't find it written anywhere, but it relies on the statistical progression of tolerance stack to find an acceptable position for related, but unconnected, parts. FWIW, this works on any assembly which is subject to varied tolerance stacks.

You really need to keep accurate records, both written and photographic, when using this method, it can get away from you really fast.

 

[use2bn83]

I’m catching a bit of the reverse pattern that’s looking interesting. Let’s see some more. This is the compound that is being transferred to other teeth.

Are you getting plenty of resistance in there? One of the ways I generate a bunch of resistance during setup is to wrap a doubled up towel around the pinion flange and twist it until it is wayyy tight. Then I put a wrench on the ring gear bolt to add enough leverage to move the ring gear back and forth creating the pattern.

Also open backlash way up to see if the patter becomes more centered heel to toe. Try .28mm for instance

 

[Darije]

Here is service manual file for suspension, axle, diffs etc. I use it for my 1991 HDJ80. I think it is up to 1995 year. Not sure is the later manual same as this one. Here you can find all details for diff settings

 

[twistinchassie]

Had a play with the back lash today, I did overlook on simple check of the axial runout of the ring gear and sure enough it has a run out of 0.08mm. I recorded the runout at 4 positions and referenced the backlash at each of these positions. They don’t completely correspond the way I would have expected however I need to strip the pinion and check its runout.




After this I backed the lash out to 0.28 at tightest and 0.33 at largest corresponding to my earlier runout/ clearance positions.



The greater lash did improve the drive side pattern however the coast side is still pretty rubbish.

Might be time for a set of nitro gears?

 

[use2bn83]

.08 is a bit high for a factory gear, but not worth condemning in my opinion. I’ve seen this a couple times caused by the fillet on the carrier interfering with the inadequate chamfer on the carrier bearing….. however that is typically an ARB/Aftermarket bearing issue. That last photo you have here looks fairly decent. Coast looks a bit shallow, and drive looks pretty close, having the drive slightly deep is not a terrible thing. I’d add .15mm or so to the pinion shim (err on the shallow side), reset BL to .28 and see what you got. What you have there is dang near runnable, especially for used gears. I’d have zero hesitation running .25 BL on a used set if that gets you the best pattern. In fact when I setup used gears I set BL at what it came in at to avoid them coming out noisy which seems to happen when you take a set that comes in at .012 inch and tighten it up to .006 inch.

Edit: if you don’t have the proper Toyota shim on hand, you can use some small squares or circles of shim stock to mock this up and the order the proper shim.

 

[twistinchassie]

.08 is a bit high for a factory gear, but not worth condemning in my opinion. I’ve seen this a couple times caused by the fillet on the carrier interfering with the inadequate chamfer on the carrier bearing….. however that is typically an ARB/Aftermarket bearing issue. That last photo you have here looks fairly decent. Coast looks a bit shallow, and drive looks pretty close, having the drive slightly deep is not a terrible thing. I’d add .15mm or so to the pinion shim (err on the shallow side), reset BL to .28 and see what you got. What you have there is dang near runnable, especially for used gears. I’d have zero hesitation running .25 BL on a used set if that gets you the best pattern. In fact when I setup used gears I set BL at what it came in at to avoid them coming out noisy which seems to happen when you take a set that comes in at .012 inch and tighten it up to .006 inch.

Edit: if you don’t have the proper Toyota shim on hand, you can use some small squares or circles of shim stock to mock this up and the order the proper shim.

So finally got back into the shed and swapped shim from 1.05 to 1.15 and increased tightest lash to 0.27(where it ended up with caps torqued.

Patern on coast side im alot happier with but the drive side now seems too deep. I do have a 1.10 shim i might put in but really need to get the 3rd back in the car. Whats your thoughts on the drive side being that close to the root?

Cheers

 

[Malleus]

Focus on the drive side; the coast side can be marginal and still be OK. By definition, it has the lower load on it.

Used gears are notoriously hard to set up properly. They can look like crap and still be usable; meaning they'll carry the load safely, but they'll sound louder than new gears. You have to be wiling to say, "OK, my drive side is acceptable, the coast side isn't unusable, I'll put them in and try them". If you're not willing to go to the effort to put them in and take them out three times, bite the bullet and buy new gears.

It really depends on how valuable your time is.

When we built axles for off-highway large AG/construction vehicles, we'd get differentials that we'd pulled for 100% inspection during initial build (we always marked these in the build records) returned to us, with customer notes that "they were too loud; they must be bad". We'd tear them down and print them again. Almost always they printed 100% acceptable. We broke them down and built the gears in new carriers, built those (returned) carriers with new gears, marked them both in the build records and forgot about them; they never came back again. Sometimes you build them and, by visual inspection, they are fine, but don't "sound like" they work. At that point you assume that it's something that is due to a very large number of components interacting deleteriously and move on with life.

 

[twistinchassie]

Focus on the drive side; the coast side can be marginal and still be OK. By definition, it has the lower load on it.

Used gears are notoriously hard to set up properly. They can look like crap and still be usable; meaning they'll carry the load safely, but they'll sound louder than new gears. You have to be wiling to say, "OK, my drive side is acceptable, the coast side isn't unusable, I'll put them in and try them". If you're not willing to go to the effort to put them in and take them out three times, bite the bullet and buy new gears.

It really depends on how valuable your time is.

When we built axles for off-highway large AG/construction vehicles, we'd get differentials that we'd pulled for 100% inspection during initial build (we always marked these in the build records) returned to us, with customer notes that "they were too loud; they must be bad". We'd tear them down and print them again. Almost always they printed 100% acceptable. We broke them down and built the gears in new carriers, built those (returned) carriers with new gears, marked them both in the build records and forgot about them; they never came back again. Sometimes you build them and, by visual inspection, they are fine, but don't "sound like" they work. At that point you assume that it's something that is due to a very large number of components interacting deleteriously and move on with life.

Thanks mate, i ended up seeing what the 1.10 pinion shim produced pattern wise at the same backlash, i feel it wasnt a huge change. Primarly concerned with achieving a better drive contact pattern however the used gears just dont want to play ball as id hoped.

In your opinion what is preferable contact toward the root or the top flank/face?

Ill run with the 1.10 pinion @ .28backlash pull for inspection after some use with all sounding ok.

Thanks again.

 

[use2bn83]

In my opinion, post # 15 is the best setup you’ve had so far. I’d run that. The drive pattern there looked pretty darn good. These recent ones look too deep. And ditto on the above about coast pattern’s.

Edit: looking at the pinion in 15 you can also tell that the pattern is a lot better. In the recent post you can see that the RG side is deep and the pinion side is shallow.