And after thorough testing, I can report that the whip was not to fault, neither was the IO board.
Instead, there seems to be a distinct problem with the IO-interface part of the mainboard itself.
I tried different (both unbalanced and balanced) whips, and also swapped I/O board for an identical I/O board from a (confirmed functioning) Pulsar II.
MIDI does not work well, lots of wrong and/or hanging notes.
It appears that there is a general IO problem which means MIDI data being corrupted, as per the original symptom description.
Additional testing showed that ADAT transfers is also at fault, resulting in faint clicking noises in audio.
This applies to both send and receive, as it is evident both ways, and looping data through the card doubles the noise.
I have not tried the A/D & D/A converters, but I expect them to have similar problems.
The fault is light, meaning that audio-data such as that transferred over ADAT is not heavily impacted, but MIDI is impacted severely as it is non-recoverable errors (eg. MIDI-off messages lost resulting in hanging notes).
Luckily, the board still functions effectively as an SRB.
It is currently housed in a machine with 2 PCI slots, with a Powersampler doing the all audio I/O - this seems to work ok... (No clicks or pops, and no error messages.)
Although there is no visible deterioration (bulging tops or leakage), I am considering replacing all the caps on the 15 DSP board (not too expensive, and doable without SMD de- and re-soldering), just to see if it does the trick.
Although it could be any component failing, at least the CAPS have a good 20 years on them now, and their probably specced to 3-5 years originally.
If and when I get to it, I'll share how it goes.
Meanwhile, happy to hear of (constructive) suggestions for further qualification
