Hi Lapponia,
Really interesting, I think you are on to something. I have just taken out the contact from my suspected faulty DME relay, the one where I noted the purple wire in the previous picture, and as you can see:
one contact, the 87b contact, is badly burned and pitted. The contact on the other side is perfect, as are both contacts on the other relay. Is this the contact you suspected?
I am sure this burned contact must be the source of my problems. There is a ring of soot around the burned pit, which looks just like the pit burned by a current arc, so I am sure that your conjecture that arcing is tacking place when the engine is switched off, and the relay contacts have to break the current through an inductive load, is correct. (Just like the points in an old style mechanical distributor, as you say.)
This is clearly an inherent fault in the circuit design, because this arcing must
inevitably, sooner or later, cause enough damage to the relay contact that it becomes unreliable, and leaves the car stranded, as I now know many 944 owners have been over the years. However, I must say that the DME relay that failed on me is the original one fitted to the car over 20 years ago - I hope the new one lasts as long!
Swapping to a higher current capacity relay won't cure this arcing, surely it needs a zener diode or capacitor (or both) to soak up the voltage spike and prevent the arcing? This would be a nice mod that could probably be built in to the existing housing? Any idea what capacitor or diode spec to use?
Regards - David