Spurious parity error problem fixed! Installed entire 1132 controller logic at same time

AS I HAD REALIZED, PULLUP RESISTORS MISSING FROM A-A1 SLOT M3

I did find an SLT card in the pile of removed cards for the 1132 printer controller that was the only card in that compartment that was NOT associated with a specific peripheral controller. It was type 410 that I replaced in slot M3. 

The keyboard entry worked perfectly now without any phantom parity errors. I am surprised it ever worked at all, rather than only triggering errors with certain situations. What matters is that it is working reliably now. 

It may have involved charging up the Miller capacitance that exists on the base of germanium transistors, which increases with higher leakage current. One of the transistors may have had enough leakage to charge while the bit from a previous character was latched in from the read part of a memory cycle, thus appearing to conduct when the other bit turned on with the new keypress. 

INSERTED OTHER SLT CARDS FOR THE 1132 PRINTER CONTROLLER LOGIC

Compartment A-A1 contains cards for the 1132 printer across rows 2 and 3. The pullup resistor card was in slot M3, which I mistakenly took to belong to the controller along with every other card in those rows. I decided to reinstall the cards for the 1132 as that would give me the chance to test portions of the controller logic without having to cable up to my 1132 printer. 

CPU TOOK A LEVEL 1 INTERRUPT IMMEDIATELY; REQUESTED BY 1132 CONTROLLER

I saw the keyboard access loop jumping to a random address when it took an interrupt for level 1 immediately upon pushing the Prog Start button. I opened the ALDs and found where the controller logic asserts the request for IL1. There are a few reasons, among them completion of a Space or completion of a Skip to carriage tape channel operation. 

The flipflops to set up a Skip or a Space are triggered by inputs with a + polarity, meaning that the unconnected printer wires for those buttons will appear to be high. As previously mentioned, an open circuit is the same as a logic high to many SLT gates. 

I installed a jumper that blocks the interrupt request for level 1, after which the machine no longer tried to take the interrupt.