Yet another PVM RGB lines problem, but it's not the caps!

[tms9918]

All of a sudden, I got the three lines on the top problem. I changed the two usual caps, but that made no difference. Indeed, measuring the ESR, they were fine.
I then recapped all the board, but that brought no improvement either.
With the service menu I can bring those lines up, but even 255 is not enough for pushing them out of the screen.
I attach a picture of the problem. Top row, underscan enabled. Bottom row, underscan disabled.
The menu seems to be shifted (without underscan, the last row of text is clipped out) but the monoscope is clearly not.
The yoke coil should be fine. If I make the image exaggeratedly tall, they are able to draw up to the screen border.
The problem seems to be different from https://sneeknet.io/index.php/2019/02/25/pvm-20m4a-power-supply-fault/ because, as you see in the pic, the monitor still shows the signal on screen.

Any guesses on what the reason could be?

 

[hatmoose]

Another PVM fan! I have about a dozen ranging from "museum quality" to "cant even pay them to take it away"

Whats ya model number?

 

[tms9918]

PVM20M4J, which is identical to the PVM20M4U,

 

[hatmoose]

One other thing to check is the value on C584 - this was installed in the factory as 2.2mf/160v and the service manual notes it as 2.2mf/160v.

Many people on the internet (i know, I know) say that C584 is the wrong value and should be replaced with a 1mf/160v cap. There are various rumors of a service note about this, but I have not seen it with my own eyes

But if you are looking for a place to start researching C584 could be a begining

 

[tms9918]

I tried both values. I too had a mismatch between the mounted value and the service manual value. It made no difference.

 

[Mukuro_W]

You have to replace all caps near the V. Blanking ic, those are just two caps, one 1uf 160v high ripple and just near the 4.7uf 160v those often fail due to heat in this area