So... this afternoon I finally got up the ambition to survey the damage; the x-axis felt weird at approx 20-60mm travel, like it was binding or something. Between that and the weird behavior, I was fearing the worst; like the stepper had grenaded (even tho it was stone cold last night) or worse, the TMC2208 had kakked from repetitive overload caused by the binding. Well, at least the troubleshooting session gave me a chance to practice my cold-pull technique.
Of course, the culprit (as usual) was something much more mundane; the pinion had spun loose on the shaft. Usually this gives you some warning; the screw hitting against the flat of the shaft will hammer like a summitch, and you'll notice it right quick. But not in this case, even though I DID check all the set screws when I assembled the kit.
Turns out that there's a bit of a fault in the geometry of this stepper in this application; with the pinions they chose, there just isn't enough depth on the flat of the shaft to be able to center the pinion on the belt. So, rather than fix the design flaw by choosing a different stepper or pinion, they just tightened the grub-screw against the round of the shaft. Well, the half of it they could hit with the pinion located properly.
So, a little Dremel work got me the flat on that shaft extended ~3mm; reassembled everything and now the X-axis geometry is correct and should stay that way.
Interesting side-track; it took me 3 more tries to actually get a successful start on the print; first one curled at ~3mm progress, so following the conventional wisdom with PLA, I turned the bed heat down. That caused a print fail clusterflop where the part broke loose in one piece and glued itself to the nozzle. Digging the PLA from around the nozzle without ripping the insulation on the heat-block proved quite the job, even with tweezers.
Finally I noted that the edges (where this print was going down) were actually enough cooler that I could feel the difference; even though I wait a good 10 minutes every time for the glass to come up to temp and stabilize. I had to crank the heat up to 70° to get a successful start; and yes, I have a good 4mm brim on both part and support, because about 60% of the footprint of this thing IS support.
Now just hoping it doesn't start to curl up again...
mnem