As I've posted previously posted
here, I bought a Gould Delta 9500A which I was able to make run after replacing the battery. I also did some other maintenance on it, like replacing the fan to make it quieter, but I'll leave that for another topic.
I've tested the scope and it mostly works fine, but Channel 1 starts glitching after ~5 minutes, less if the device had been previously running, which points to it being a thermal issue. I've tried the inverted computer duster freeze spray trick all around and it didn't make any difference.
I am absolutely sure it's not an analog board issue: I've scoped out the output of the analog frontend with another scope, and no glitching appears. I've also tried swapping the CH1 and CH2's coax lines, and the issue always appears on the digital CH1, regardless of the analog channel connected. The issue is probably not on the main board (which is on the opposite side of the case), because it handles all the channels.
There are no service manuals for either the Delta or (similar) Classic series of Gould scopes. Even the last supplier of their parts (H.J. Helmbold) doesn't have it. Let me present some images to clarify the architecture.
Here we can see the underside of the device. It has two overlapping boards: the top one is the analogue one, and the bottom one is the digital acquisition board. The conditioned analogue signal is funneled from the analogue board into the digital one through coax cables you see in the photo. The boards are also connected with 2 flex cables. The bottom board is also connected to the main ('computer') board on the other side of the chassis through a board-to-board connector.
On the analogue board, you can see by the silkscreen that the leftmost of the identical circuits is the Channel 1. On the digital board, CH1 is the bottom one. Now, here's a potential issue: when I first opened the scope, I seem to have knocked off the black heatsink (bottom left). I then unknowingly proceeded to run the device for 10 minutes, with the artifacts soon becoming apparent. When I figured it out, I immediately stopped, and used
thermal glue to replace the heatsink which was previously attached by a thermal double-sided tape. I stupidly forgot to read off the chip markings before gluing the heatsink back in place. This didn't help the gitches.
Let me post some close-ups:
Now, here's the other side of the digital board; CH1 circuitry is now on the top left:
And some close-ups of the CH1 part:
OK, now, my naive guess is that the chip under the black heatsink, the neighboring ASIC and the 'ANITA' chip digitize the signal, storing it into 8 SRAM (as7c256-12jc) chips you see on both sides of the board. There are also 74ACT16245 bus transcievers and 74fct162827ctpv bus drivers involved, and the data from the acquisiton RAM is then somehow sent to the main board.
Now, let me show you the artifacts:
This only happens after a few minutes: the signal occasionally gets distorted in a very digital way. Moving the position of the signal affects whether it appears or not and its amplitude, and the waveform is occasionally completely distorted (but this is also affected by the position parameter, also the coupling mode). Sometimes the position doesn't affect the way the signal is distorted, just moves it around.
Maybe it's some kind of acquisition RAM issue? Doesn't seem to be clearly indicative of either a stuck address bus bit or a stuck data bus bit, or does it?
Otherwise, what else could go wrong that's actually fixable. I know I could get another broken scope (but these are very rare) and scavenge one channel's chips, but I'd rather try to fix this one if the issue is not in one of these unobtainium chips. Anyone have any ideas?
Home
Help
Search
Login
Register
