I've had a quick look at my 11/45 printset.
It appears that the interrupt handling is partially done by the processor
microcde (page 12 of the microcode flows in my printset). And that there
is no special path for the vector, it's goes via the normal data paths,
into nromal intenral CPU registers, etc.
So, I would propose the following tests :
1) Make sure you can read/write all bits from external memort. Just write
1, 2, 4 ,8, 16,.... to successive locations using the panel and examing
them. Jsut to eliminate a silly fault like a dead data buffer.
2) Run the basic instruction diagnostic if you've not done so already.
This should veryify that the data paths are OK.
3) If possible, try some other interrupting device. An obvious one is the
console receiver. Wirre a program to : Load all the vectors with .+2, HALT
as you did for the line time clock test that's failing.; Read the Console
Rx data register to clear the data received bit (if necessary); then
enable the receiver interrupt; and go into an endless loop. Then you press
a key on the conosle. The processor should halt, but _where_ does it
halt. If 4, then it appears the processor is ignoring all vectors.
Hopefully this will point us in the right direction to look for the fault.
Show replies by date