A few months ago, I was gifted a non-operational DEC VT320 serial terminal. This is a 80s/90s text-only CRT monitor, which was only displaying a thin vertical line rather than a proper picture. I've done some electronics repair in the past, but nothing this old or involving a high-voltage tube. But I figure that it would make for an interesting project over the summer, whether or not I can recover it. If nothing else, it is quite a retro talking piece.
Fortunately, DEC -- aka Digital Equipment Corporation-- made a lot of these things, and also published troves of service information, which was the norm back then (#RightToRepair). Among the documents that can be found online included the schematics, which would be of incredible aid. That said, the quality of the scan is pretty poor, and it took quite a bit of cross-referencing of component identifiers with the parts list to confirm what couldn't be read.
Before opening up the monitor, I did some preparatory research, to identify areas which would be worthwhile to investigate, and also to make sure I'm not going to shock myself in the process. My findings showed that because the CRT was able to display something beyond a single dot in the center of the screen, the issue would not be in the high-voltage generation circuitry, but rather the downstream circuits. Specifically, in the horizontal deflector path; this explains why the only image is a vertical line, because horizontal control was lost.
Using the field replacement guide, I got the monitor open and then discharged the anode using a screwdriver attached to the ground strap. There was no spark or sound, which can happen if the bleeder resistor was still intact. A good sign as to what's probably still working in this monitor, but I take no unnecessary risks around potentially high voltage.
The innards were reasonably laid out, basically existing as components that adorn the CRT display itself. The PSU, main board, and "arc protection" board were all easy to identify, although the latter was more like a connector than a board. I quickly ruled out the PSU by checking its output voltages, so the issue must be on the main board.
It took a while to extract the board, since I didn't want to break any of the 40-year old plastic clips. But once out, I began matching the board to the diagrams and examining for any obvious damage. No obviously blown caps, no evidence of thermal events, no components rolling around on the bottom.
At this point, I stopped to do some very thorough circuit analysis of the circuit diagrams, to absolutely understand what I was going to do. This actually took two tries, since each attempt revealed faults in my understanding, and I had to go back to the thinking chair. This part took a few days, until I finally internalized the circuit's behavior. As it turns out, this wasn't necessary, and I'll probably include it as a later comment, just for posterity.
When I returned to the board, I made a plan to solder some trace wires, to verify my expectations when powered on. And indeed, after reinstalling the board, my oscilloscope confirmed that the H sync signal was intact and the power transistor was functioning as expected.
Narrowing the search, I took the board back out and started tracing the lines on the PCB surface -- it was fortunately only a two-sided board -- and then compared my observations to the schematics. This revealed a difference, where the Horizontal Linearity inductor (H-LIN) was not showing connectivity, despite visibly being attached to the trace.
Closer examination revealed that of the inductor's three legs -- two for electrical connections, one extra for support -- one had developed a hairline fault. This physical damage broke continuity, likely from the shock of impacting something. I made a repair by constructing a wooden splint for physical support, and then soldered over the leg for electrical connectivity.
For good measure, I completed the exercise of verifying all other components, which showed that all other resistors, capacitors, and connectors were intact and likely working. And with that, I reassembled and powered on the monitor to see.
And it works! In total, I probably spent a week on-and-off working on this. I will say that this is a strange machine to have, since I've been able to hook it up to a modern Linux machine and use it as a serial TTY. I even wrote most of the text of this post in vim.
I skimmed this article, but I'll revisit it later because MPU protections have always intrigued me, in terms of being a much smaller subset of the protections afforded by an MMU. That said, for work projects, I've only ever been able to make use of MPU regions for coarse-grain protection (eg prevent the DMA-dedicated regions from being accessible by everything) and not for per-task protection. So I'm always interested in ways that more isolation can be applied, precisely because it keeps code honest, which isn't a bad thing. It's no different than having software asserts.