OK, Home From Work Now... Time For a Debrief.
Alright, I started to do a multi-quote of all your observations and suggestions, but that would result in a novel that nobody would want to read.
So...
Those of you who understand scientific method know that I have done this in as disciplined a way as I know how. This is the thread where the pros hang-out, so I had better not enter here with any ignorant BS. I understand all the known failure modes of tubes in a clincher tire assembly, but I have never seen this before. I thought that by changing the rim-tape/tube brand/talc-lubrication variables in an organized experiment I would find a solution, but no.
My best take-away here so far is that I am indeed running too wide a tire for this particular rim. I'm comfortable with that conclusion, and have already replaced both wheelsets with same-hub & CR18 wheelsets. These are working perfectly for my applications so far.
But that damned dimple pattern - and no, there hasn't been a case where both dimples leaked - keeps mocking me. I've checked this with a digital caliper, and the c-c distance between the dimples in the tube is always at least 1-2 mm larger than the diameter of the spoke hole. If this tube were somehow poking into such a space, you would expect it to be smaller than that diameter when the tube was deflated. 
Is it possible that the tube is entering two adjacent spoke holes and expanding until one fails? Then when I deflate the tubes, does that distance between the "nipples" shrink that dramatically?
If that were the case, surely I would find evidence of that around the rim-tape. 
(edit) I had written some nonsense here about sending wheels out to people, etc. What was I thinking? Do I really need that kind of hassle right now? No. So here it is, four-thirty in the morning, and I'm retracting that offer... sorry. Now maybe I can get back to sleep... Have a nice weekend everybody.
__________________
Roulez pour la joie, jamais pour la douleur.
USMC 1981-1991 Semper Fi!
Last edited by DQRider; 11-04-17 at 03:30 AM.