![\"\"](\"https:\/\/blog.system11.org\/wp-content\/uploads\/2024\/01\/iskraext2.jpg\")
<\/a><\/p>\nYou can use board swapping if you have a known good unit to isolate which PCB is faulty, but from there?\u00a0 It’s difficult.\u00a0 I spoke to Sergei Frolov who runs the Soviet Digital Electronics Museum (link here: http:\/\/www.leningrad.su\/museum\/<\/a>) about this problem a while ago and he showed me a video using a transit board – essentially a straight extension that lifts the problem PCB up so you can get an oscilloscope on it.\u00a0 Logic probes are no good here as all the chips operate on -27v!<\/p>\n I had a brief search but couldn’t find anything that would work, so I finally on Christmas eve, sat down with KiCAD and designed one.\u00a0 The pitch of the boards is 4mm which is no longer a common type and edge connectors with 58 pins in this pitch are completely impossible to source.\u00a0 However, there’s a 3.96mm part from EDAC which fits the boards snugly enough to allow that 0.04mm per pin to not matter too much, so I made it 4mm on the calculator side, and 3.96mm on the board side.\u00a0 It’s important that the PCB is the correct side to fit in the cage and guides, and fortunately my first attempt was correctly measured.<\/p>\n Here it is in place, the machine in the photo runs so badly that pressing any number causes it to crash.\u00a0 It didn’t used to be that bad and only crashed on certain calculations – so my hope is that the chip in question has become bad enough to be easily spotted now.\u00a0 Previously I’d changed 3 chips which are common failures, with no change whatsoever.\u00a0 Maybe I will have this one fixed soon.<\/p>\n<\/a><\/p>\n