Amstrad CM14 CRT computer monitor - blank screen

[PenroseHouse]

The eBay seller who makes and sells the cable that converts VGA to SCART RGB - specially made for the MiSTer - has this to say:


If white is purple, I would say it is most likely the convergence on the CRT TV/monitor. I.e. one beam either red, green or blue is not aligning well enough to produce a solid white colour. It would be more evident with white borders vs. a full white screen. You'd be able to confirm with a jewellers loupe or possibly a phone with camera set to magnification mode.


Does anyone know if this can be fixed, if this IS the problem? If so, I'll set it all up again and try taking a photo in close up.

Note that I've tried:

-the VGA to SCART RGB cable
-a VGA to SCART RGB board by the same seller
-endless combinations of settings and of the MiSTer Multisystem's Sync on Green/Purple physical switch
-various MiSTer "cores" and games

And note that I get more than one shade of purple, so if a game has a white border maybe I get very obvious purple, if the border turns yellow or green maybe the whites turn pale purple, and with a red or blue border maybe the whites are actually white! The change is instant and clean, nothing gradual. I've got white or I've got two specific purples.

Could it be some caps need changing, something like that?

[Maarten]

Purple white is not a convergence problem if the whole white detail is purple. Maybe a purity problem, but could be a signal problem.

[PenroseHouse]

Thank you, Maarten.

Yes, it's a very clean and perfect purple, there's no way you'd know it wasn't intentional if you didn't know the Spectrum.

Probably too complicated for an amateur like me to fix? I recently opened and discharged another monitor, and I've recapped a Vic-20 before, but that's about it...

[Maarten]

A very clean and perfect purple is either an RGB gain or bias (adjustment) issue on the monitor (convergence or purity are ruled out at this point) or a problem with the signal from the MiSTer.

[PenroseHouse]

I think we've exhausted possibilities with the MiSTer's settings - I may buy a real Spectrum to test with as I don't have anything else RGB.


-RGB gain or bias (adjustment) issue on the monitor

Could that be screws or similar inside that can be fiddled with? (While risking death because you'd have to have the case off while it's switched on?)

[Maarten]

Usually the RGB gain and bias are indeed adjustable. In a decent monitor they have one trim potentiometer (twiddly screw) each, so 6 in total. Usually on the CRT board or on the main board near the signal cables to the CRT board.

Of course the risk is that if you get the adjustment correct that way, it might just be to compensate for a faulty monitor or faulty signal - you then have to readjust when you repair or apply a correct signal.

In case of a purple image, a quick test would be turning the G gain up a little bit.

If you set the monitor up properly you'll first have to feed a black picture, turn up the brightness such that you can just see an even lighted background. Then set the bias pots so the background is the desired temperature of white. Then turn back the brightness to normal setting and apply a white picture. Then set the gain pots to the same desired temperature of white. The temperature would probably be 6500K but if you can't measure that, adjust it so that it looks good.

[PenroseHouse]

Marteen, this is worth trying even though the purple changes?

For example, if I have the ZX Spectrum core with an entirely light grey screen, there is very obvious purple. If I bring up the MiSTer's menu, which effectively reduces the amount of light grey on the screen (because the menu has lots of dark text etc), the purple turns to PALE purple.

[PenroseHouse]

OK... I've opened it up, it has four VERY plasticy pots, or I assume that's what they are, they have slots in like screws and they're on the side of a small board on the back (not the main board on the bottom).

Two blue labelled B, each labelled with what looks like:
VR803
VR805

Two red labelled R, each labelled with what looks like:
VR801
VR804

Probably doesn't mean anything, just the shape of the plastic screw caps ... but one red is turned at a different angle to the other three which are the same.

---

If they ARE worth twiddling with, I don't quite understand the instructions.

I set the computer so that it has a black screen as the output? Then turn up the brightness on the monitor so I can just about see some white (well, grey)? While this is so, adjust pots so it's a clean white/grey rather than purply?

Then set the computer to show a white screen again, and do the same?

Thanks!

[cmjones01]

I've just looked at this thread again and wonder if there's something more subtle going on. Before twiddling R and B cutoff and drive presets, double check whether the colour rendition on the screen changes when the border colour changes. If it does, you can bet your bottom dollar that what we're seeing is a black level clamping problem.

Summary: at the start of each line, there's supposed to be a short period (not normally visible on the screen) which is black. The monitor uses that as its reference, and, basically, shows all colours on the screen relative to that. So if that period contains red, the whole picture will end up looking cyan. If it contains blue, the picture will look yellow - basically the opposite of whatever is in that "blanking period".

So if the MiSTer is delivering a signal which isn't quite standard, and isn't giving enough blanking time after each horizontal sync pulse, the picture colours will depend on the border colour. This, incidentally, is also the reason why the picture from a ZX81 looks so dark on many TVs: its video output is bright white during the blanking interval so the TV uses that as its black reference. So then the white parts of the screen (which is most of it) look black, and the black parts (the text) look even blacker, and you can only see anything by cranking the brightness right up.

Chris

[PenroseHouse]

Thank you for this, cmjones01

I think you're right - and I'm thinking there's nothing can be done about it without an expert building some special board that changes the signal...?

As for what happens:

The colour DOES change, very cleanly and perfectly, and instantly so far as I can tell, when the border changes.

Manic Miner on the Spectrum, you're on a screen with a red border, the whites are about right - you exit to a level with a yellow border, suddenly the whites turn purple.

Boulder Dash on the C64, which has a black border, whites are perfect - you collect the final diamond and the border flashes white for about a half second - or it's supposed to, but of course it flashes purple, and anything else white on the screen turns purple.

It couldn't be a cleaner effect.

Note:

"So if that period contains red, the whole picture will end up looking cyan." - bear in mind I'm using a MiSTer with the Multisystem which has a toggle switch for sync on purple/green, which I often toggle (and I can't recall which is the default setting), so if what I've described above doesn't quite match what you expect, that could be why.

All in all, I'm thinking fiddling with those pots will at best correct the problem for SOME border colours, but then make it worse for others?

[cmjones01]

Originally Posted by PenroseHouse

All in all, I'm thinking fiddling with those pots will at best correct the problem for SOME border colours, but then make it worse for others?

Exactly right. Twiddling the pots won't fix the problem of the colours changing. I'd leave them alone. It does sound like the symptoms are consistent with what I described. Unfortunately I'm not familiar with the MiSTer and its settings and behaviour so can't offer much advice, other than what it's outputting to the monitor is wrong. Changing the sync-on-green setting won't make any difference. I wonder if there's something about its video timings which is a hangover from the VGA/31.5kHz/scandoubler timing. VGA monitors don't, in my experience, do the black level clamp at the start of the line, so it's possible to get away with having a coloured border straight after the sync pulse. But now I'm just theorising.

My next move would be to get a scope on the R/G/B video signals, and the composite sync signal, and examine them while experimenting with the MiSTer settings.

Hang on a moment - I've just had a thought. Is there anything like a "horizontal shift" or "position" adjustment in the settings? That would probably adjust the video timing relative to the sync pulses and, if adjusted too far "left", would start the video too soon after the sync pulse, leading to the black level clamping problem you're experiencing. If such a setting exists (I'm clutching at straws!) then adjusting it might well help.

Chris

[PenroseHouse]

I can't see any settings suggesting "horz shift" or "position", though I've attached the settings .ini in case anyone fancies taking a look through.

I'll get back onto the MiSTer forum with what you've told me at some point...

Note that I'm using the VGA output with an adapter that actively converts the signal to RGB, so the original output IS VGA.

[PenroseHouse]

Attached...