CRT Horizontal Sync / Image Offset Problem

[fixedsys]

Bought an old CGA chassis (Nanao KZ20EN) that has a problem with the horizontal sync. The image is offset and I haven't been able to find the root cause. The white raster position / size and adjustment control is OK.



The schematic is here (pg 45): https://www.arcade-museum.com/manuals-monitors/Sega_MC-2000-S_20in_Color_(420-5028)_4th_Printing.pdf

I don't exactly know what the flyback pulse input RC circuit is supposed to do to the flyback pulse (which measures at 5.6V), but it doesn't appear to adjust the signal via the trim pot (H-PHASE) in any significant way (I expected it to shift the phase, and maybe fix the issue).
Also the voltage seems low on pin 13 at 0.5V, as the schematic specs 3.5V.
Any ideas?

BTW, what are the encircled M and PP symbols used in the schematic? Mylar / Polypropylene?

[inse]

Thats really strange, is the monitor synced to the video source?
Does the video format match to the monitor specs?

[fixedsys]

Does the video format match to the monitor specs?

I was thinking that too. I'll check with the video card mfg.

I still think it's related to the seemingly non-functioning H-PHASE circuit but there doesn't seem to be much that could go wrong in there.
Could the flyback transformer have an issue where it's generating the incorrect pulse, i.e. caused by winding insulation breakdown?

[inse]

Unlikely the result looks that ‚perfect‘

[fixedsys]

OK, so the pulse would be erratic if there was a flyback insulation problem. The image is stable however.

[fixedsys]

Thats really strange, is the monitor synced to the video source?

Only the one card is connected to the monitor (with R/G/B/Gnd/H/V).

[m k]

Pin 13 is low, h-phase is not detected.
The chip needs to know how to position the horizontal data on screen.

Shrink the height.
Is the starting point of the first line where the faint vertical line is on the right?

On paper you have a string of text.
After certain amount of characters you have a LineFeed character.
If those LF counts don't match the paper their combination is diagonal(/), that's h-sync.
Obviously moving the paper is not going to help, that's h-pos.
(analog is a bit off here since cathode ray is still advancing when printer is just returning)
When LF count is correct their combination is vertical.
But nothing is defining where the '|' should be, that's h-phase.

One is still missing, h-hold.
That is not needed, the chip does that, and in the analog it's a paper property anyway.

[CaptDon]

Are you sure you need to be looking at the horizontal circuit? Is it possible that particular display has the deflection yoke turned 90 degrees from normal and horizontal sweep is across the narrow side of the tube and vertical across the long section? Some radars are built that way. Some of the old 'Paper White' document display C.R.T. systems also had the yoke turned 90 degrees.

[inse]

You can see the scan lines…

[CaptDon]

O.K., closer look I see the scan lines. How about the H-Sync polarity is wrong and your picture is offset by the width of the sync pulse?

[fixedsys]

Is the starting point of the first line where the faint vertical line is on the right?

Yes.

But nothing is defining where the '|' should be, that's h-phase

So it is an H-Phase problem then?

[fixedsys]

Are you sure you need to be looking at the horizontal circuit? Is it possible that particular display has the deflection yoke turned 90 degrees from normal and horizontal sweep is across the narrow side of the tube and vertical across the long section? Some radars are built that way. Some of the old 'Paper White' document display C.R.T. systems also had the yoke turned 90 degrees.

Not 100% sure I need to be looking at the H circuit, but I am sure this is a video game monitor and chassis and the voke is the correct orientation. I know this because the card is set up for vertical CGA resolution, and I'm getting a vertical image.

[fixedsys]

O.K., closer look I see the scan lines. How about the H-Sync polarity is wrong and your picture is offset by the width of the sync pulse?

Fortunately the H-Sync polarity is switchable on this chassis. When I reverse the polarity, the offset gets worse - seems to offset ~50%.
I also swapped the H and V just for good measure. This messed it up beyond control, and so confirmed it's correct as it was.

[fixedsys]

Update: Replaced the three caps in the H-PHASE circuit and that seems to have fixed that particular circuit. The pot now shifts the phase of the flyback pulse signal a little while the signal looks much stronger and cleaner.
However, this has not fixed the offset issue. Adjusting the phase pot will cause it to lock in either H- or H+ offsets (~25% offset / ~50% offset).
Seems the phase lock is incorrect. How is it adjusted? Pin 12 - C508 which has 2% tolerance?

[m k]

Detection moment is just a level.

H-phase pot is between 6.8k and 8.2k resistors, check them also.

How is brightness?
Q506 voltages are given.
You should have plenty of adjustment area from where picture start appearing to where returning beam start being visible.
If not then voltage levels are not correct.

You bought a known problem machine and don't know what previous owners have done.
Maybe you should check the whole setting procedure.

BTW,
H-hold and V-hold are still adjustable, even their signal examples are there.

[fixedsys]

Brightness is good.
Q506 voltages are about 15% over. I'm not sure what the tolerance is, but the image is offset by about 15% also.

[m k]

15% is quite high.

If it's all around then initial regulated voltage of 105V at TP2 would be 120V.

[fixedsys]

I did set the B+ to spec 105V. But I think I need to go through all of the setting procedure, as you've suggested.
I just checked the input voltage and it's ~15% over, and therefore over spec. I'll need to change the isolation transformer.