PDP-11/45: Diagnostics with PDP11GUI

Sun 03 July 2016 by Fritz Mueller

Okay, now that serial is straightened out, on to running diagnostic tests via PDP11GUI. PDP11GUI itself, as well as a thorough and useful database of DEC diagnostic programs, are available at http://www.retrocmp.com/tools.

Since my home computer is a MacBook, I had intended to run PDP11GUI under Wine. I ran into a problem with this where PDP11GUI under Wine could not generate its pre-processed machine description temp file, seemingly because of some incompatible behavior wrt. multiple backslashes in pathnames. Rather than fight with this too long, I just sprang for a Windows 10 license and installed a Windows 10 VM; it will come in handy for other Windows-only tools that have been tweaky to use under Wine as well.

With PDP11GUI up and running under Windows, results of the initial set of 11/45 CPU diagnostics were very encouraging! In summary:

DiagnosticBELDescriptionStatus
CKBAB0.BIC002562SXT instructionpass
CKBBB0.BIC003604SOB instructionpass
CKBCB0.BIC007410XOR instructionpass
CKBDC0.BIC007262MARK instructionpass
CKBEC0.BIC002212RTI/RTT instructionspass
CKBFD0.BIC002272stack limitpass
CKBGB0.BIC001446SPL instructionhalt 001404
CKBHB0.BIC00376211/45 registerspass
CKBIB0.BIC013746ASH instructionpass
CKBJA0.BIC014722ASHC instructionpass
CKBKA0.BIC014430MUL instructionpass
CKBLA0.BIC011574DIV instructionpass
CKBME0.BIC01600011/45 trapshalt 005320
CKBNC0.BIC004702PIRQ instructionpass
CKBOA0.BIC01364011/45 stateshalt 000610

Note that these tests are written to output an ASCII BEL to the console on each successful pass. The terminal built in to PDP11GUI doesn't sound when given a BEL, however, so it is convenient to patch the BEL literal in the programs to a visible character (e.g. 000052, ASCII '*') before running them. This may be done in the memory loader window in PDP11GUI after "Load" but before "Deposit all". The patch address I identified for each diagnostic is listed in the table above as well for convenience.

So, out of this initial set of tests, only three halts to investigate. I'll be posting more information here as I look further into these.

Regarding the backplane SPC issue discussed in the previous post: Marty from the vcfed.org forum did some investigation of his 11/45, and reports that he does have +15V (actually +12V on his system) distributed to pin CU1 on slots 26-28. It seems to be wired over from slot 15 on his system, but it is not clear whether this was a factory wire, an ECO, or a user mod.

Marty's 11/45 also has no power distributed to CA1 on this slots, so it's really looking to me like the reference to that on page 5-10 of the 11/45 maintenance manual is a misprint. Thanks to Marty for checking all this out! I'd be curious to hear from any other 11/45 owners out there regarding wiring of CU1 and CA1 on slots 26-28 in their systems.