VOGONS


"TRL" CRT Horizontal Position issue

Topic actions

Reply 20 of 25, by Deunan

User metadata
Rank Oldbie
Rank
Oldbie
x86_guy wrote on 2024-04-08, 20:11:

The issue now - at cold startup the image is vertically shrunk and vertical linearity isn't ok. Those issues self sorted after a minute or so of the monitor running - Any directions?

There is no bottom side photo of the very front of the PCB (it's obscured by monitor case) so I can't be sure, but the vertical driver is that thing under the heatsink and either the oscillator part is also in there, or perhaps the 555 in the corner is the oscillator. In that case my bet is the round foil capacitor in the corner is setting the free-running vertical frequency.

The cap is not really suspect because these don't go bad, and if they do it's not really temperature dependent. But the soldering could be cracked so check that. Without a scope it's just a guesswork, you can't tell if the problem is in the oscillator or the yoke driver. Since it fixes itself with time, and pretty fast too, I'm tempted to blame the driver because it will get warm much faster then 555 chip - but then again the 555 are all clones of better or worse quality and those can die or break in weird ways with time.

There is also an electrolytic cap on the bottom of the PCB, under the driver, where it is hot, so do check or even try replacing that one.

This is perhaps a good case for the freeze spray - if you have one - you could try cooling both parts once the linearity gets good to see which one will break it again. But make sure you don't hit the CRT, the glass in the neck is thin and could easily crack due to temperature shock. I'd also avoid touching the heatsink with bare fingers, not because it's so hot but there might be unsafe (or at least unpleasant) voltages present. The 555 should be safe to touch.

Reply 21 of 25, by x86_guy

User metadata
Rank Newbie
Rank
Newbie

I replaced the capacitor under the driver (the one with the heatsink) and also cleaned the rear v-size pot (the one on the back panel) using contact cleaner. Non seems to help.
This driver controls also the amplitude of the sawtooth wave of the deflection system? I may replace it, if it is the main suspect.

I'm a bit afraid to use freeze spray because of your warning that the glass can crack as a result.

Below you can download a video that shows the monitor at cold startup:
https://file.io/9xbyLirYYHBI

Reply 22 of 25, by Deunan

User metadata
Rank Oldbie
Rank
Oldbie
x86_guy wrote on 2024-04-13, 23:19:

This driver controls also the amplitude of the sawtooth wave of the deflection system? I may replace it, if it is the main suspect.

The driver usually only follows the shape of the signal from the oscillator. Any correction it can make to help the current through the yoke coil be constant (which improves linearity) is very limited. But obviously this only applies to parts working properly. As I've mentioned without a scope it's not really possible to tell if the driver is bad or if the oscillator produces bad signal.

Since this seems to actually get better/good as time passes, rather than worse, you could try heating the PCB instead of cooling. A hairdryer would work just fine for that purpuse and you can use on the back of the PCB, rather than from above to minimize any CRT neck heating. Do note some dyers are pretty powerful so test it on your palm first, you don't want to damage the PCB or overheat the chips.
Start with cold PCB (well, within reason, give it an hour after a previous test) and try heating the two spots I've marked. See if linearity improves faster when one of those is heated and you'll narrow down the possible cause.

BTW the file you linked seems to be deleted already.

Reply 23 of 25, by x86_guy

User metadata
Rank Newbie
Rank
Newbie

I forgot to mention earlier, I also checked the soldering joints in the areas you mentioned, and also in the rest of the board. I couldn't find any dry joint but I may need to look into it deeper when I get the time for it.

For the heating test - The driver (the one with the heatsink) sits really close to the CRT glass, I'm still a bit afraid to try it (I don't have a hairdryer but a heat gun). I have to think of a way to avoid the heat from reaching the CRT, then it will be possible to test it this way.

I have an oscilloscope, is it possible for you to instruct me what to measure to try and determine where the problem is?

I'm trying again to share the cold startup video:
https://ufile.io/x6l9dixq

Reply 24 of 25, by Deunan

User metadata
Rank Oldbie
Rank
Oldbie

Got the file this time. I see the initial linearity is very bad but it quickly improves. Things to try:

1) Inspect the flyback area for solder cracks. Or just skip the search and resolder it, using flux or adding fresh solder (but don't make big balls of solder, remove any excess solder or flux afterwards). When CRTs were still in use this was the very first thing anybody would do, it's quick and can fix weird problems. Sometimes cracks here can be inside the solder blob and not really visible.

2) Look for any electrolytic caps connected directly with one lead to the pins on the yoke header. You are mostly interested in the 2 side pins on the denser part of the header. The pins that are more spread out are for horizontal deflection and this seems to work just fine, so you can skip those. Mark those caps and consider replacing them even if they seem to test OK.

3) I never had issues with it myself but that brown glue that was used hold the parts for soldering (and initial transport) can apparently turn conductive. I would not expect it to cause the issues you have though but when you have some time see if the brown stains on the back of the PCB can be rubbed off with IPA - it might take some time as it doesn't dissolve easily. But try not to use any abrasives or you will damage the solder mask. If water and/or IPA don't make a dent in it, leave it be for now.

4) Look for pintout of a 555 timer chip. Pin 1 should be GND. You can connect your scope there and probe the other pins, look for ~60Hz (actually a bit lower without input signal, 56-58Hz I guess) triangle waveform there. You can solder temporary wires (but not too long) to the pins to allow easier testing. Just don't solder a bunch at once or they will short out.
And before you do any of that, make sure that pin is connected to the same ground reference as the GND pins on the video input connector. This is just in case the chip is not grounded but somehow floated, you don't want to short the PSU in the monitor via the scope. If in doubt wait and ask.

5) If you don't feel confindent heating or cooling the PCB with the tools you have then don't. The CRT well-being is much more important than any of these tests. I sure would hesitate to use heat gun on anything electronic, other than to do quick and dirty desoldering job.

Reply 25 of 25, by x86_guy

User metadata
Rank Newbie
Rank
Newbie

I heard about this glue issues somewhere before, good that you mentioned that.
I went ahead and tested conductivity on this glue with a multimeter. I put the multimeter on resistance measure and placed the leads on various points on the glue - this thing is really conductive! The closer I placed the leads one to each other, the less resistance it measures (the scale was in Mega ohms).

As this glue really placed over various contacts on the board, I decided to remove it (as you suggested).
I used IPA, heat gun and a brush. During the process I had to remove some caps (glue was under it), and some traces wasn't look very well so I re-soldered what I thought needed.

Than, I assembled everything again and tested the monitor and - problem solved!
From the beginning of a cold start, the vertical size and linearity is great and doesn't change as time goes.

I can't really tell what solved the issue, is it the glue cleaning, or the re-soldering of the joints that the glue made dirty, or anything else I didn't payed attention to.

Thanks a lot for the help!
(Another project is about to come soon 😀 )