To my other GT6 pages

September 1, 2019

Differential

I once thought that automotive differentials were some kind of black magic.  All of the car manuals I'd seen warned that the exacting clearances and preloads necessary to rebuild them properly required special tools and skills.  The lengthy list of steps they offered seemed complex and confusing.

But today, with precicely one apparently successful differential rebuild under my belt, I am a little less aprehensive.  Still, tearing into a differential that I really want to be spot on is not something I take lightly.  It takes a lot of focus and concentration to keep from making stupid mistakes.

As far as I knew, the diff in the car was working OK when the car was parked 30 some years ago.  So my approach was to go slow, and to try not to fix something that didn't need fixing.

The first thing I noticed when I hauled the 50-some pound lump up onto the work table, was that the pinion seemed very free-turning.  Pinions normally turn in pre-loaded bearings that cause the rotation to be noticeably stiff.  The backlash also seemed to be a little generous.  I decided to open her up.




Since I had everything out of the cases, it was a good time to clean and paint them.  Sand blasting is a really good way to get rid of paint and rust.  Blasting goes better if the parts are degreased first.  I tried to plug all entries to the inside to minimize the grit on the interior.




Blasting didn't take long.




Having bare cases also presents an excellent opportunity to do a couple of small mods.  

This is the vent for the differential.  Differentials get hot when they operate, and the rise in temperature raises the air pressure in a sealed case. Higher pressures foster leaks.  The vent is just a small hole to let air pressure equalize.  As is pretty common with simple vents of this type, the hole had a loose cotter pin in it.  The pin is presumably there to rattle around and prevent the hole from becoming clogged with road grime.  This often didn't work.




There are better vents.  This one, made especially for differentials and transmissions, features a cap to keep crud from falling into the vent.  It also has an internal filter material.




The new vent takes a 1/4 NPT thread.




While in a drilling and threading mood, I decided to also do another mod that is pretty common.  This is the bottom of the rear case half.  There is a boss there at the lowest point that was clearly intended for a drain port that was never installed.  Drill and tap for 3/8 NPT, and we have a differential drain.




With the cases cleaned and modded, it was time to paint.  A nice 2-part epoxy primer and topcoat should be pretty durable.




New rear bushes and pop in the vent.




By this time, I'd decided to try to re-set the backlash and pinion preload.  I thought it would be silly to dig that deep into the diff without installing new bearings.  This, of course, ups the cost of the refurb quite a bit.




After removing the old bearing cups and doing a thorough cleaning of the inside of the cases, the new bearing cups for the pinion shaft have to go to their forever home.  The threaded rod gizmo pulls both cups into place at the same time.




The first ctitical setting in this diff was the pinion height.  The pinion has to be in a certian fore-aft position relative to the crown gear to ensure proper mesh of the teeth.  There was apparently a factory tool to help set pinion height, but they would be like hen's teeth today.  I believe it worked by measuring from the top of the pinion head to the flat surfaces where the differential carrier caps seat.  It was complicated by the fact that pinion heads apparently varied a little in thichness, and in fact each factory pinion came with a marking indicating how much it differed from some set standard.

I took a slightly different tack.  There is a specification in the factory manual for the "Distance from head bearing abutment face on pinion to center of crown wheel bearings".  I took this to mean the distance from the backside of the pinion head (where it meets the pinion shaft) to the flat surface where the differential carrier caps seat.  To check my interpretation, I set the rear pinion bearing cone in the cup, and added the original 0.081" shim.  The top of that shim is where the pinion's "abutment face" rests.  Measuring this distance, I came up with something within less than 0.002" of the manual spec (which is 3.03125").

This was my setup for the measurement.  on top of the 0.081" shim is a presicion machinist's block that is 3.000" tall.  Across the flats at the center of the crown wheel bearings is a precision ground parallel that has very good flatness.  By the spec, there should be 0.03125" between the top of the block and the underside of the parallel. With feeler gauges, I measured 0.033.  The pic on the right shows how I managed to hold the block rigidly in place for the measurement.  This would also emulate the preload on the bearing.




The pinion height is adjusted by the shim under the pinion head.  Thicker shims will raise the pinion, and my pinion needed to go up a little less than 0.002".  The next larger shim that I could find was 0.084", so I ordered one of those, plus a new 0.081" shim.  It turned out that the new 0.081 shim got me closer to the spec.  It appeared that the original 081 shim  was slightly worn.

At that point, with pinion height set, I moved on to set the pinion bearing pre-load.  Correct preload is gained by precisely setting the distance between the front and rear pinion bearing cones.  This is done with a steel distance piece and thin shims.  When the pinion nut is tightened to its spec torque (90-100 ft-lb), the bearings are under a load determined by the spacing.




I couldn't immediately find the proper shims (1" ID, 1-1/4" OD), but I found an assorted set of shims with 1" ID and 1-1/2" OD.  Luckily, ODs are easier to trim than IDs.




Preload is measured by the torque it takes to turn the pinion.  I think it took three iterations or so to get something in the 12-16 inch-pounds window.  Since the pulling point on my setup was actually four inches from the axis, the reading on the scale had to be multiplied by four to get the actual torque.




With the pinion in good shape, I turned to the differential carrier.  This assembly's job is to allow the wheels of the car to turn at different speeds while both are driven buy the engine.  It consists of two "sun" gears connected to the wheel axles, and two "planet" gears that act as idlers between the sun gears.  Ideally, the backlash within the sun/planet gearset should be as small as possible, consistent with free movement.  




My backlash was about 0.010", so it needed some work.




Backlash is set by varying the thickness of the spherical copper washers behind the planet gears.  Unfortunately, no one I could find supplies the washers.  It was frustrating to realize that wahsers for the TR6 are widely available, and are the same ID and OD, but are quite a bit thicker.  I read online about one ingenious fellow who made washers from sheet copper by clamping them between a planet gear and the housing.  I looked at this, but didn't have any copper of the right thickness.

My final solution to this was to electroplate new copper onto the original washers.  I worked out a current density and time that would plate about 0.001" of new copper all around (washer thickness increased by about 0.002").  After each plating session, I assembled the carrier and measured the backlash.  It took four cycles to get to the point where backlash was very close to zero, but the mechanism still turned pretty freely.




The next step was to set the pinion-crown wheel clearance.  This is done by moving the position of the crownwheel left or right by shims placed between the differential carrier bearings and the case.  The total amount of shim is determined by assembling the carrier into the case without the crownwheel and measuring the distance it can float from side to side.  The total shim pack is set to this distance plus 0.005".  The extra 0.005" places a pre-load on the bearings.  This total shim pack is divided into two parts, with one part going behind the bearings on either side of the case.  The crownwheel position is set by moving individual shims from one side to the other.  Compared to the shims I found in the diff, I had to add 0.003" to the total shim pack, and move 0.008" from the off side to the crownwheel side.  I also could not find much variety of these shims to purchase, so they were mostly home-made.




A case spreader allegedly makes installing the carrier with its shim packs easier.




As a sanity check of proper shim distribution, the backlash between crown and pinion should be between 0.004" and 0.006".




The final determination of proper setup is the contact pattern on the crown teeth.  This is seen by coating some of the teeth with a dye, and meshing them repeatedly.  The contact areas will wear the dye away--eventually.




With no load on the teeth, the contact patches are very slow to appear.  I could begin to see them about the time my arm got tired of moving the pinion back and forth, but the camera still couldn't capture them. They looked very close to the desired pattern in the shop manual.


At last, I could button her up.




The last task was to do the inner axles.  Each one consists of a splined shaft with flange, a seal and seal carrier, and a bearing.




A little cleanup, new bearing and seal, and some powder coat...




...and these dudes are ready.




Final stretch...




This space is begging to be filled.




It was a piece of cake with the frame upside down.




Even though this differential went faster than my first one, I still would not want to do it for a living.  Cost was under $200, mostly for bearings.

Comments to Ed at elhollin1@yahoo.com

To my other GT6 pages