[Simh] Non-numeric character in IBM 1620 P & Q addresses

[Robert Armstrong]

> an invalid digit in MAR generates MAR ERROR, lights the MAR 
>CHK LAMP, and stops the system.

  I thought that might be the case.  Thanks for researching it.

  BTW, I was wrong about the check stop switch bypassing the MAR CHECK - I found several references in the manual that make it clear that the PROGRAM/STOP switch in the MAR CHECK, MBR-E CHECK and MBR-O check column does NOT have any effect on the MAR CHECK.  It only affects MBR-E/O parity error stops.  MAR CHECK always stops the machine.

  There is a switch on the CE panel inside the machine which bypasses the MAR CHECK, but that was for field service.  It wouldn't have been used in any normal conditions.

  So it looks like an invalid digit in the MAR was considered a hard error and would always halt the machine instantly.  Since the original problem was a bad digit in one of the address fields (I forget which) of a TD instruction, it seems like that never would have worked on a real 1620.  That means either the original program image is wrong, or there's a bug elsewhere in the simulator.

Bob A