Originally Posted by Don's GC
The mechanic probably did not mark and count the turns he had to turn the flange nut when removing the pinion seal. If he got it to tight the crush bushing is compressed causing the backlash to be wrong. I always (years ago) marked the pinion and nut and counted the turns to remove, and reinstalled with the same amount of turns to keep the backlash correct.
It doesn't change backlash backlash is adjusted with the threaded adjusters on the carrier, its bearing preload that is the problem. When the bearings are preloaded so tight that the bearings cannot ride on the lubricant film instead it squishes it all out you get bearing failure.
Also they way you suggested is how I was taught how to do it, I was then taught at school that this is STILL not correct. I even tried it for myself while setting up a differential in class, setting up the turning torque removing the nut and reinstalling it to the marks i put on the pinion and the nut it further crushed the crush sleeve and my turning torque went way up. the "proper" way to do it (which VERY few people do, at ANY shop) is to remove your wheels, followed by your brake calipers and rotors (due to the ebrake in the hat), then measure your turning torque at the pinion flange (drive shaft removed naturally), then you match this number while going back together.
Considering the noise happened after the pinion seal repair he likely has some dead pinion bearings as already suggested, the likely hood a wheel bearing gave up the ghost right after a repair that can commonly cause this concern is slim-to-none, possible but not probable.