actually thinking of it yeah ibm used the 8p5c version (Rj45 looking connector) and the best way to not have blue smoke would be to trace the signals to a known chip and see where power and ground land. like that big NEC chip, seems to be an intel D8748 replica (pretty common keyboard controller back in the day from what i understand) https://pdf1.alldatasheet.com/datasheet-pdf/view/80208/INTEL/D8748H.html
Thank you for that. I was almost entirely certain that I had done my due diligence, but I still managed to cook a resistor with it wired up to a Soarer's converter.
The yellow wire leads directly to various negative leads on components, has continuity with one of each of the contacts on the switches, and has direct continuity with the appropriate ground pins on every chip I bothered to check. Continuity goes cold quickly on the red lead.
It leads directly into pin 1 of this voltage regulator, labeled in the spec sheet as "input". The yellow wire has continuity with pin 2, labeled ground in the spec sheet. The white and blue wires are bridged, and connect directly to a bolt that connects to the switch plate. R2 is the resistor that gave up the ghost.
Pin 3 of the aforementioned voltage regulator eventually leads to, and has continuity with, the VCC pin on the SN74LS273N, which has continuity with the VCC pin of the D8748D. Both ground pins on these chips have continuity with every negative lead, including straight back to the yellow wire, that I could identify.
I just guessed on the clock and data wires, figuring mixing those up wouldn't damage anything, but maybe I was wrong? I'm certainly no electrical engineer. I think I did trace the black wire back to something on the D8748D that the spec sheet said could be used for clock, but I didn't necessarily run that to clock on the pro micro. The D8748D should be 5v, but maybe the original power from the terminal was something else entirely?