Found two failed components on the 3819 card

OPERATED THE CARD ON THE BENCH TO CHECK OUT ITS FUNCTION

I plugged the card into my SLT card test jig and hooked it up to its voltages - +3, +6, +48V and then +12V through the indicator lamps it will drive. I quickly noticed that the +3 and +6 supplies were being fed voltage from the 48V source, so I shut down and proceeded to check parts without power.

By comparison, I tested another 3819 card from the machine and it worked just fine. No strange flow of 48V to the other power rails and it worked properly. I had a spare 12V bulb I wired to a +12V supply and hooked to an output pin. When I pulled the input pin to ground, the bulb illuminated. High resistance between all the rails, as should be the case. 

FIRST DISCOVERY - ONE OF THE EIGHT POWER TRANSISTORS WAS A DEAD SHORT

I was testing diode voltages of the transistors when I came to this one which was completely shorted between all three leads. I removed the offending transistor, which was made by Fairchild and is marked 284-6516 - obscure IBM internal code. The schematic also lists this as an IBM 284 transistor, which is equally useless in purchasing parts since those are private IBM numbers. 

I didn't find any of those transistors in my stockpile of old SLT cards, but fortunately this particular 3819 card is only connected to six outputs, so that two circuits were spare. I can grab one of the spare transistors and move it over to the location where the failed part was removed, thus this card can be used in slot D5 going forward (once all other faults are found and repaired).

SECOND DISCOVERY - DEAD SHORT ON TANTALUM CAPACITOR FOR 48V RAIL

The capacitor for the 48V supply is a dead short to the ground pins that go back to the indicator lamps on the console panel. I remembered that a team restoring an 1130 in Europe encountered exactly this failure. The result is that they saw 48V out on the lamps not the 12V that should be observed, and they found that some ground traces on the backplane at the socket for the bad card had been evaporated by overcurrent. 

I have many spare tantalum capacitors from my old IBM SLT cards, so that is easy to replace. I will have to carefully check the traces on the card itself as well as on the backplane at socket D5 and make suitable repairs. 

ONCE REPAIRED, WILL BE FULLY BENCH TESTED BEFORE REINSERTION

I will use my bench supplies with their overcurrent protection capabilities and ensure that everything works properly with this card, circuit by circuit for the seven circuits that will be useable.