An update: I've extracted the 1345A display unit and examined the boards we spoke about yesterday; I powered it using a bench supply I designed and built myself years ago which I use for all this kind of work, so the unit was independent from any other elements of the 8757A analyzer, to produce the "pikes peak" test pattern.
?
I saw some suspect solder joints on both of the smaller boards on the side, cleaned them and re-soldered them. I also dismounted the PROM and the two static RAM chips and re-seated them; the RAM's are fine (checked on another unit I have).
?
The problem persists, I ran out of time last night and have been busy with other matters today, so my next steps are to check both oscillators (20MHz and 19.66MHz) and the divided-down derived frequencies and also examine the "Vector Done" signal path.
?
I am also wondering if the tantalum capacitors (those yellow ones on the boards) may be suspect - I have spare caps which I can replace them with but I cannot think of a way they would be causing this problem, the supply rails are exactly as they should be so they are certainly not short-circuit, but maybe open-circuit or high ESR.
?
This is taking a lot of time I could be using on other things so I'm also thinking of finding a replacement 1345A to buy, but that would not be a "real" solution!
?
Trevor Gale