8.8 pinion preload problem

[sonicblue]

I just had the pinion flange nut backoff about 2 years after changing the pinion seal, so this time I wanted to change to a solid spacer to end this problem. I got the Raytec IK4105 kit and noticed the solid spacer ID is a pretty sloppy fit on the pinion, but I talked with a shop and they said that was normal. I measured the old crush collar and ended up with a stackup of about 0.010 less to get to the correct preload. I figured this was probably do to some springyiness in the collar since it's now unloaded. The Raytec kit comes with shims for 0.020 through 0.050 in 0.002 steps, but I only needed about 0.010 so I couldn't get into the normal shim range. I machined the solid collar down 0.012 and got a good reading using 0.022 shims, about 8in-lbs. I put it all back together, drove it on the freeway for about 10 miles, came back and was shocked to find I could hardly turn the pinion it was so tight! I pulled it back apart, redid the shimming, and now I need an 0.030 to get 8in-lbs of preload.

WTF??? How could this change? The axle was very hot when I started reshiming and cold when I was done but this didn't seem to have an effect on the preload which is what I would expect. I then drove it again on the freeway for 10 miles and it seems fine this time, but since I can't figure out what happened, I don't have a lot of confidence in my setup.

Sorry for the long post but this has be stumped. Has anyone else ever seen this issue?
BTW, I have setup dozens of axles over the years and this is the first time I have ever seen anything like this happen.

One other note, I'm torqueing the pinion nut to 200ft-lbs. There is no torque spec for this nut since you are usually crushing a crush collar, I thought this was reasonable. Maybe it's too much?


Confused,
Darryl

 

[boominXplorer]

Were your bearings good and lubed before you installed? Make sure when you are tightening the nut you need rotate it counter clock-wise between turns. The range for the preload is 14-19 in lbs also. Hope this helps

 

[sonicblue]

The bearings seemed to be well lubed as this is a running axle not a new rebuild. I always turn the pinion a few turns after tightening to run it in and always got a consistent reading on the torque wrench. I was setting the preload to around 7 to 8in-lbs since if I dropped the shim back 0.002 it would jump to over 20in-lbs and I felt that was way to much for used bearings.

One thing I have been reading is Ratech suggests 125 ft-lbs on the nut for final torque. I have been using 200 ft-lbs. Also when I machined the collar down, it didn't seem to be all that hard of a material and was actually made from welded seam tubing. Is it possible I overtightened the nut and the spacer actually did crush or distort a little on that first test drive. That would certainly explain why the preload went up. It took 0.008 additional shims to get it back in line, so maybe the spacer shrunk this amount? I would call Raytech but can't find a number for them.

Darryl

 

[4x4junkie]

I would think a spacer should be hardened. Not sure why it would change unless it wasn't hardened and it did infact crush (in which case it's garbage). Also, I think 250 ft-lbs on the nut is a more typical torque when using a solid spacer (D44 and D60 pinions I know are around that range, give-take a few).

FWIW, some Loctite on the pinion threads (be sure the threads are clean) should prevent the nut from backing off if you reused it. Also, a very tiny tack-weld on it would do the same (Toyota guys do this a lot as backed-off pinion nuts are a frequent occurrence on their trucks).

 

[sonicblue]

The Ratech collar I got definitely wasn't hardened. When I machined it down the 0.010 it cut like mild steel on a lathe. If it was hardened I would have had to grind it down. This is why I'm thinking they specify the 125 ft-lbs, but that's just a hunch.
There is another solid spacer kit I found the SK CSF8.8 from places like Jeff's Bronco Graveyard. Has anyone used this kit and found it to be better quality then the Ratech?

Darryl