AppleVision 1710AV Service questions

[NeXT]

I have three of these monitors. Two are crumbled messes of yellow plastic but I have another one which I bought about 14 years ago that's still completely unyellowed and was trouble-free until yesterday.

For those who don't know the 1710AV was a 17" Trinitron monitor sold by Apple with built-in stereo speakers, microphone and ADB software control and an integrated ADB hub. It also suffered from reliability issues. The most common of them was arcing in the high-voltage would corrupt the EEPROM that stored deflection, convergence, color and geometry data and in the worst case the high voltage section would completely shutdown. If you were within a certain serial range (SG522xxxxxx and less than or equal to SG628xxxxxx) you got an extended warranty due to them being a recall nightmare.

https://www.applerepairmanuals.com/the_manuals_are_in_here/AppleVision_1710AV_Display.pdf

Mine is serial SG640xxxxxx from October 1996 which was quite late for the monitor series and thus had a number of the factory fixes already applied to it. Aside from the odd time it decided to twitch the deflection it had been flawless until finally the monitor will now work for the first few minutes when cold and then chrip, arc, the EEPROM values get scrambled and eventually the high voltage shuts down. If I leave it for a few hours it work again for another few minutes.
Since it's a Trinitron from a known problematic family I cannot tell if I am looking at a problem with capacitors, dust, flyback or another component failure. For how much some people desire these monitors and the lust others have for the Trinitron there's little to no discussion about their problems and the many fixes beyond reading the service manual and people mindlessly saying "just replace the deflection board". Come on now, You cna't tell me everyone who ran into problems with these monitors just threw them away. At least not in this decade.

Edited: As I was typing this I left the monitor running and it again kicked out after around five minutes, so I let it rest for another five, powered it on and it's been mostly behaved now for the last 20 minutes. The failures are completely unpredictable but each other time it started acting up either letting it sit for a while or turning the monitor off and then on again as the service bulletin says was all it took to fix it.

 

[NeXT]

So the answer to what was going on here was not that it was a flyback failure or an EPROM problem. It was dust.

Disassembly of the monitor as per the manual isn't possible anymore due to the fragile nature of the plastics. Instead I placed the monitor on its tube face (use a rag to protect it!), remove the screws holding the speaker base and stand off the underside, unplug and completely remove the base. The manual wants you to dismantle the monitor, starting with the front. I can assure you the many clips holding the front bezel will not survive this, so we can go in through the back.
Locate the two plastic clips on the underside that hold the back on, break them off (because they will not unclip on their own) and then being mindful there is still a CRT inside with a fairly large neck deflection board, grab the back of the monitor and use the weight of the monitor to rip it off the back of the monitor.

Remove the bits of broken plastic rattling around inside the monitor and then clean the analog board and interior of the monitor with compressed air. You can at this point plug the monitor as-is back into a mac and power it on because the base is just an ADB hub and audio section.

I've had the monitor sitting like this on my desk since last night and so far it's running fine now that it's cleaned out. I have however found that the pot for upper convergence adjustment has gone bad. Classic pot is linear to a point, then goes open. Oh well. No point mumbling to Michael Spindler in spite at this point because he's already dead.

 

[NeXT]

So this was interesting. After letting the monitor sit for 8 hours cycling test patterns and not hiccuping once I put the monitor back together and corrected the remainder of the convergence issues in software, only for five minutes later the monitor once again began chirping and losing convergence, geometry and color balance.

Pulled it apart again and like watching a pot until it boils the monitor it wouldn't fail, so I started thinking of what else it might be. Thermal breakdown? It's working with the back off and failing with the back installed.
There is a ribbon cable that runs from the interface board for the base that runs up to the neck board. Presumably this passes data from the ADB monitor interface to the neck board which then handles most of the geometry, gun balance and convergence before exiting on the bottom of the neck board and into the analog board. This ribbon cable is physically in contact (either from age and things settling into place or poor design) with the high voltage anode wire which is clean and should still be insulating the 15000v or so, but in theory if there are transient voltages building on the exterior of the wire it's able to discharge into the ribbon cable which isn't likely to have insulation rated for more than 600v and would thus fire spikes into the data path, causing all sorts of problems.
I can't relocate either the wire or the ribbon cable so instead I tried pulling the anode cable away using a ziptie, then I put the back on the monitor and tested it again for the evening so the base and its electronics (which should just be the ADB hub and the audio amplifier) are out of the equation but the monitor was allowed to warm itself up to where it was failing. The problem has gone away again. I'll let it run for a few more hours tomorrow after an overnight cold soak and it it doesn't act up I'll reassemble the monitor again and test it.

 

[NeXT]

Nope. Still failed.

Frustratingly it started getting worse and would fail even with it completely opened up, FINALLY though a smack on the side of the monitor caused it to glitch out and confirm we had a mechanical failure. One that as previously observed would happen once the monitor warmed up.
Turning all the lights out and waiting for it to fail again I peeked around and eventually spotted a small blue arc underneath the monitor. On closer inspection I found the HOT was coming loose. Collector was still soldered in, but base showed signs of failing and Emitter clearly had gone cold.

Assuming this is the smoking gun, this explains a lot. If the HOT starts playing up the flyback immediately starts acting up as well which in turn disrupts the B+ feedback loop, and when it wasn't causing HV shutdown would cause both spikes and dips which would cause the microcontroller for the screen adjustments to corrupt its values.
What I find interesting is that this entirely from thermal cycling. The monitor has not moved in over a decade. This means that it will unfortunately be a problem with other Applevision monitors.

 

[NeXT]

Well it's been over a week now and with on and off usage the issue has completely gone away, so I can conclude the initial fault has been located and repaired.