Played with the 19818 locker over the weekend.
I pulled the ring gear off the AMC-Unique carrier. The ID chamfer on the ring gear pilot bore is ~0.100 deep, which make things a little more difficult.
I took the three units into the QC lab to measure the flange offset using a height gauge.
3.54-: 1.820 (Spec from RockAuto is 1.810)
3.73+: 1.938 (+0.118 from 3.54-; spec is 1.943)
AMC: 2.149 (+0.211 from 3.73+; don't have a spec)
On both Eatons, the ring gear locating journal looks 0.320 long.
Soooooo... 0.320" (journal length) - 0.211" (difference in flange offset) = 0.109 (overlap of ring gear on locating journal)
0.109 overlap initially sounds good, EXCEPT that the ID chamfer on the ring gear is 0.100 deep. So maybe there's 0.010 overlap, maybe there isn't because +/-0.010 is about the best confidence I have in some of those measurements, as I was eyeballing edges next to the calipers.
HOWEVER, the AMC carrier was installed with 0.030 & 0.007 (0.037 total) shims on the gear side and 0.028, 2x 0.010 & 0.005 (0.053 total) on the far side. If I drop the far side shims down by 0.040 to 0.013 or so and increase the gear side shims by the same amount, that will move the ring gear flange 0.040 closer to the ring gear, which increases the overlap from 0.010 +/-0.010 to 0.050+/-0.010... which is workable.
That means that I can make a 0.170-0.180 ring gear spacer, which can be ground from a 3/16" laser cutting.
Because I'm adding a ring gear spacer and the stock ring gear bolts are 7/8" long, I can put the assembly back together with 1" ring gear bolts, which are hella easier to find than 7/8" bolts.