PDP-11/45: LA30 repair III

Digging in on the flip-flops identified as potentially problematic in the previous post, found that E5 had failed. Pulled, socketed, and replaced; character generator now correctly clocks all five character columns:

After this repair, characters were printing full width, but two problems remained: about half of the characters printed in response to typing on the keyboard were the wrong character, and the top row was not printing at all on any character:

Looking at the incorrect characters problem first, it was clear that bit 4 was not being received by the character generator correctly. I was a bit worried that the SMC KR2376-17 scanner/ROM on the keyboard assembly might be at fault, since Mattis had had some trouble with his. This is a pretty cool part; a combined scanner and code translator, with an internal oscillator, rollover logic, debounce delay, and flexible interfacing:

...not to mention the very cool vintage ceramic/gold packaging (see below.) Fortunately, inspection with an oscilloscope showed that the outputs from the scanner were just fine; chasing downstream, the problem was found to be just a loose pin (SS) on the keyboard cable Berg connector. With that sorted, we now have this:

For the final issue with the top row not printing, verified that the problem followed a particular G380 solenoid driver card when swapping them around, and that with a functional G380 in the appropriate backplane slot pin 1 fires and prints correctly. Inspection of the problematic G380 revealed a failed power transistor and blown associated micro-fuse; replacement parts on order.

For the ribbon advance issue, I pulled the ribbon motors and disassembled their top-side reduction gear cases in order to gain access to the upper rotor bearings. Cleaning and lubrication of these bearings, plus a few more taps with a mallet after reassembly, achieved an improved bearing alignment. With the increased output torque, the ribbon now advances reliably.

Other minor items: Replacement vibration isolators arrived, and were installed. Threaded inserts in the fiberglass top shell that had pulled out were reattached with epoxy.

Have some more travel coming up for work, so may not be able to get back to this for a bit. Next steps will be repair of the failed solenoid driver channel, calibrations, then any debug necessary on the M7910 interface card for the PDP-11.