"Start at the beginning" worked this time! I re-tested the power supply voltages after the recap and found them still out-of-whack...and as it turns out the electronics of the SP-10Mk.II are EXTREMELY sensitive to out-of-spec voltages.
First test results (no load, nominal/actual):
5V / 5.39V
32.5V / 33.12V
140V / 151.5V
For further testing, I connected the trusty ol' Fluke to the power supply, hooked up the table to test things under load, and adjusted the 5V supply so that it was dead-on at 5VDC. 33 & 45 OK, 78 still slow and physically wobbly. Not a logic problem, obviously.
After that I set the Fluke on the 32.5V motor drive line, and dialed it in with the table running. At first the pot didn't seem to want me to get anywhere between 32.42 and 32.58, but after working it back and forth a bit I got it right at 32.5. 33 and 45 good...78...also good! No more wobble!
The 140V line is unregulated and only drives the neon strobe light. This French page has a pretty cool LED replacement in post #5: http://www.ls3-5a-forum.com/phpBB3/view ... ?f=15&t=94