I just finished wiring up my controllers and wanted to test them using the /opt/pidp1/test.sh script.
In the process, I decided to document ALL the switch positions as seen on my PiDP-1, which I will assume is correct as everything has been working as expected.
Attached is a PDF one page printout of my spreadsheet that I used to document each switch action as well as various combinations. Every combination is the simple adding of whatever switch value each switch has - for example if switches individually give 040 020 010, then flipping all 3 gives 070 (abbreviated here).
My numbering scheme follows the labelled sense switches, starting with 1 on the left and increasing to the right (so rightmost test switch I name 18). I did number the first Extension switch 0 to keep the other numbers grouped nicely.
SO ALL NUMBERS START WITH 1 AS LEFTMOST SWITCH unless otherwise printed on the PiDP-1 panel.
The only oddity I noted was the two rightmost address switches (I numbered 11 & 12). When flipped alone, the bit pattern replicated across the whole set as you will see in the pdf. However, if any other switch was also flipped, then only the primary pattern remained. It may be a glitch on my system but again seems to have no ill effect so I'm not worried.
Oscar, feel free to use the pdf if you find it useful. It's interesting to note where certain named switches find their grouping.
-R