HP 3561A display cutting out after successful POST

[lcd]

Hi! I've been lurking around here forever and appreciating it, but this is my first post. I hope I get it right!

I just got a nice HP 3561A Dynamic Signal Analyzer "For Parts/Not Working" on the Bay. Looking at the chips, it seems to be about a 1991 model. Just as the seller stated, it boots up but the screen will only output for a short time. If I wait a few days, it will fire up, pass its POST, then go into the regular screen. It will last about 90 seconds and then fritz out. Subsequent power on tests result in that 90 seconds going to 60, 30, and so on until the screen barely comes on unless I wait another few days or something. During the time the screen is off, the thing seems to operate normally. No error codes on the screen or via the LEDs on the front. Wish I could hook up HPIB to look around, but I don't have that capability right now.

I've got a PDF of the manual, and I've gone through a good bit of the power supply testing. All the rails look in spec with the exception of the 24V one, which is about .3V under allowance. I think that rail runs the input board. Doing the troubleshooting procedure, removing the boards in sequence to see if they are causing the issue doesn't reveal anything. I've not yet looked for ripple.

Cosmetically, it's in beautiful shape inside and out. No goo or extra dust in there. Someone's been in it before me, because I did see the screws for one of the shields taped to the shield instead of screwed in. I can't tell what they did, though.

I took all the cards out, cleaned the contacts and reseated them. They didn't look dirty, though. I've not looked very close to the CRT, because I don't want anything to do with 8kV.  I've fixed up several pieces of equipment, but stayed away from high voltages.

I've got a searchable manual on the way from QService, and I'll go farther into the troubleshooting in there. But I wanted to post here to see if anyone had an idea off the top of their head. I've Googled around quite a bit and haven't seen a lot. Here's a video of the startup so that you can see it happen: https://youtu.be/WRjR0Wevgvc (It happens first briefly at 1:02 and then gets going finally at 1:18).

Thanks in advance!

[KE5FX]

Interesting failure mode.  Normally you'd tackle this with a can of R134a and/or a heat gun, but a time constant measured in "days" suggests that charge on a capacitor rather than temperature is part of the causative mechanism. 

I had the opposite problem with an 85662A display that drove me nuts until I realized what was happening.  It would fail to power up if it had been off for a few days.  Subsequently, after a power-up failure, I could get it to work (and continue working) by power-cycling it.  The problem in that case was that the CRT shield ground was broken.  Without a grounded shield, the large E-field spike when the HV supply came on was radiating through the chassis and latching up some control logic.  Once the CRT was already charged, there was no spike, hence no failure.

So, I'd look for a fault condition that depends on a capacitor charging up slowly and staying that way.  Probably not the CRT HV area this time, but still, how many other caps are likely to hold their charge that long?  I would look for something in the video sync area.  Apply heat or cold to that board while the display is still running, to see if you can either force the failure to happen sooner or not at all.  It won't be surprising if that doesn't shed any light on the problem, but it's still a good place to start IMHO.  Then, follow up by measuring (and discharging) voltages across various caps after disconnecting power following a failure. 

[lcd]

Fascinating. OK that's good thinking. Would be easy enough to hunt for charges and bleed them off with a big resistor, then try again to see if the effect resets. I will say that my main trick with fixing this old gear is electrolytic replacement. So often, they are SO well designed and built that they are still in calibration, etc and just need new filter caps. So it would not surprise me at all if that were the thing here. There are not a lot of electrolytics in that path and this thing is well laid out. Once I wrapped my head around the service manual, some things started to become clear. I will do as you suggest and report back! Might take me a bit of time. Thanks!

[lcd]

Hm. OK well, the electrolytics that I found seem to discharge over a reasonable period of time. Even the big 250V (1300uF) ones. And with my little Peak ESR70 they seem to be in spec. However, I've been fooled by that before. Discharging them all and then starting up seems to have no positive effect. Same with heat on the video board.

The manual says that there is a CRT Heater, which should glow orange during operation. I don't see that. So, maybe something related to that. Well, more troubleshooting to do.

[KE5FX]

This won't have anything to do with the PS filter caps, I suspect.   Don't mess with them without a clear indication that they are faulty.

If the CRT heater isn't glowing, you won't see anything on the screen.   And if it were failing after power-up, it wouldn't look like that.

[lcd]

Point taken about the caps @KE5FX. I bet they are fine. There is one on the main board, C210, which is a 3300uF 25V one, that I can not measure while it's in circuit. I may pull that one and have a look at it.

To your point about discharging over days - that does sorta fit CRT. I wonder if there is an issue where once the CRT is fully charged up over that initial 90 seconds or so, some short or other thing develops. Guess I could test that by discharging the CRT. I'm not comfortable with that voltage but I do understand there are specific tools for that and safety procedures.

[KE5FX]

It won't kill you.  It'll just hurt.   

[lcd]

OK I have confirmed that I DO see the glow from the CRT heater even when the screen is not displaying information. So I think what I need to do is get back into the manual and start the trouble shooting of first the A15 Digitizer and then the A90 Analog Display Driver. Once I have more information, I'll post back here. Might take some time. 

[GigaJoe]

wait ...  crt ?
so it high voltage that accelerate electrons from  the el.gun (glowing thingi),  the higher voltage then more energy then brighter luminophore
it can be 2 voltages like 500V 10KV,  or just one 6-20kV depend on the size,   as high voltage generator usually do electrostatic potential, it easy leak somewhere, decreasing potential and therefore brightness. it migh be your case, look at high voltage parts.  (super careful , remaining discharge of few kV veery powerful to knock out human or el. equipment)

[lcd]

Thanks. I have a manual and am super mindful of the voltage potentials involved. This one runs at 8kV. I'll be methodical in stepping through their intended tests (of which there are many). If anyone sees the video though and thinks "Ah! I know that exact behavior!" then I appreciate it!

[TheDefpom]

I did a repair series on one of these, it might be of interest to you, I did not have any problems with the display though.




There was (maybe still is) a LCD replacement available for these units on eBay, so even if the HV system is dead, you may be able to convert it to LCD to keep it going.

[lcd]

@TheDefpom thank you! I have watched that entire series (some of it more than once). I appreciated it and your other videos. I'm looking at that LCD replacement. It looks nice and I am a little divided on it. On the one hand, it sounds like a great way to go, and on the other, I am curious to try to drive this to the end.