Differential carrier bearing preload

[18Fiatsandcounting]

Just thought I'd toss this question out to the forum to see if anyone has any experience. I recently got a 4.3 to 1 differential assembly for the early cars, also known as the "pumpkin" or "third member". It looks in pretty good shape, although it will need thicker thrust washers on both sides of the side gears in the differential case. But that's a separate issue.

I think I have a pretty good handle on setting the preload for the pinion shaft bearings, but what about the preload for the two side (carrier) bearings on either side of the differential? If I understand the shop manuals correctly, the process seems to be to adjust the pinion/ring gear backlash to around 0.004" using the adjuster screw collar on the ring gear side. Then, with no preload on the bearings, adjust the screw collar on the opposite side until the bearing preload causes the two bearing saddles to spread apart by about 0.005" to 0.006".

So here's my question: With the bolts for those saddle caps torqued to spec (36 ft lbs), do the saddles really move apart by 0.005" or so when the preload on the carrier bearings is increased? Given the beefy amount of metal, it just seems hard (for me) to believe that they actually spread apart by any significant amount. Maybe I'm wrong.

Another option for carrier bearing preload is to set the pinion bearing preload so that the rolling torque of the pinion shaft by itself is around 14 to 17 inch lbs, and then install the differential carrier, and then increase the preload on those carrier bearings so that the rolling torque of the pinion and ring gear together has increased another several inch lbs, perhaps to a total of 20+ in lbs.

Any suggestions or thoughts?

-Bryan

[Nut124]

Bryan, I spent a lot of time on this when I built a new pumpkin with the Russian LSD for my -78 last fall.

The saddles do indeed move apart.

A few things to consider:
- Make a wrench for the tension nuts. I used 1 x 1/8 SS flat stock and high grade M10(?) bolts, nuts. Three pieces, triangle with the log sides acting as the handle.
- Practice the procedure multiple times
- Practice the measurement to make sure it is accurate and repeatable
- I did not put full torque on the caps until after the pre-load was in. The outer race may not move if cap has full torque.
- Expect only to be able to move the outer race inward with the tension nut. If you go too far, loosen caps and start over.

I set mine on the tight side. 0.007" at the retainer bolts on top of the caps. Good so far.

In my failed pumpkins, there was evidence of the outer races turning under the caps although this could have happened after the failure contaminated the oil and perhaps got into the rollers while I was being towed home.

I used a slow curing Locktite bearing sleeve retaining compound to lock the outer races.

[18Fiatsandcounting]

This is very helpful advice, Nut, thank you!!! I actually do have a tool to turn the adjusting collars (tension nuts), and I can't remember where I got it nor what it is for, but it works pretty much as you describe.

So the saddles do indeed spread apart. Good to know.

And that's a very good point about the tension nut being able to move the outer race inward, but it won't just slide back outward if the nut is released. I hadn't thought of that.

I also think you are right that the carrier bearing caps shouldn't be at full torque when setting the preload, so do you have a suggestion? Half the recommended torque, which is 36 ft lbs as I recall? Or maybe only 10 ft lbs?

Bryan
PS: I also ordered myself a Snap-On dial indicator torque wrench, 0 to 30 inch lbs. Should be good for measuring rolling torque for differentials, steering boxes, etc. Early Christmas present to myself!

[Nut124]

I also think you are right that the carrier bearing caps shouldn't be at full torque when setting the preload, so do you have a suggestion? Half the recommended torque, which is 36 ft lbs as I recall? Or maybe only 10 ft lbs?

Test the outer races in the housing and see if you can move them with the tension nuts at full cap torque with your wrench. If they move, then full torque might be OK. If they will not budge, then you need to think it over.

I'd take time and develop a script to get it done using the same procedure with repeatable measurements, results.

BTW, I ordered the SKF bearings. Did not like the russian brgs sold by one of our vendors.

[Nut124]

PS: I also ordered myself a Snap-On dial indicator torque wrench, 0 to 30 inch lbs. Should be good for measuring rolling torque for differentials, steering boxes, etc. Early Christmas present to myself!

Sounds nice! I need to get me some xmas gifts on order!

For the pinion turning torque test, I built a gadget with a socket, bolts, nuts, cardboard and 3mm(?) spring steel rods. Kind of like a bending shaft torque wrench. The longer rod was the turning handle, about 16" long, with a graduated scale. The shorter rod was the torque pointer. I used a gun trigger pull gauge to calibrate the cardboard scale markings on the longer rod. At rest, the rods would be perfectly aligned, reading zero on the scale. Under torque, the longer rod would bend and the shorter one would point the reading.

This worked surprisingly well.

[18Fiatsandcounting]

I like your creativity! I've also heard of people using a fish scale and an appropriate length lever arm to measure small torque values. Put a 1 foot lever on the shaft to be measured, hook the fish scale on the other end, and if you pull 16 ounces, that's one ft lb. Or 12 inch lbs.

As for this particular differential, it's a spare, so just something for me to play around with and learn. Keeps me off the streets and out of the bars, as the saying goes. Plus, it being a 4.3:1, the ratio is too high (low?) for the engines in my cars. 4.1 or 3.9 would be better.

Since my workbench is now clear of other projects, I'll start working on the differential tomorrow and let you know what I learn. Appreciate the advice and tips, as always.

-Bryan