Indivision ECS problem with CRT&PSU?

[jimbob]

Hi, I'm having an odd problem with my Indivision ECS in combination with a Sony Trinitron HMD-220. Hope someone can offer some thoughts.

First of all I think the Indivision card is fine. I just tried it on a Samsung syncmaster TFT with no problems at all. I think the problem may be partially power related because the symptoms are different with different PSUs.

Using an A500 brick, the monitor will take a while to pick up the video signal, (ie. come out of auto st/by). Then, the amiga display will show for brief amount of time, then back to black. Amiga display will alternate with the black over increasing periods of time until eventually the amiga display is on all the time and apparently working normally.

While the black is still alternating, drawing extra power, like by directory listing on an external floppy drive, will trigger the black screen again. Once the display has stabilised for a while this will not happen.

OK, but when using an ATX PSU, the amiga display doesn't appear at all. Instead of just the black screen though, I get an info box from the monitor saying, "out of scan range".

Oh, Amiga is an A600, with ACA630. I don't think it should matter but the Indivision has an extra socket between itself and the A603. I had a bit of trouble fitting it past the ACA at first, I could probably remove it now but it is a bit of pain extracting the Indivision from the A603 so unless someone thinks this could be a problem I'll leave it.

Any ideas?

[Photon]

Testing the CRT with a PC and measuring any difference in voltages between brick and ATX PSU. Those are the only things I can think of, and you probably did that already.

[jimbob]

Originally Posted by Photon

Testing the CRT with a PC and measuring any difference in voltages between brick and ATX PSU. Those are the only things I can think of, and you probably did that already.

Now I have

Results. Monitor behaves itself perfectly with a PC.

I have removed the extra socket and reseated the indivision.
ACA630 temporarily removed.
Keyboard and external floppy disconnected.

ATX supply is pretty weak. 4.82V, 11.86V, -12.04V

Maybe I did something wrong. I read somewhere you sometimes need to do something with the 3.3V line to ensure the regulation is working properly?
Anyway, with the ACA630 etc disconnected, this now behaves like it previously did with the brick, (black screens alternating, then eventually stabilising).

A500 brick seems good. 5.03V, 12.06V, -12.13V

With the ACA630 removed, this works perfectly. So I replace the ACA630 and it still works perfectly BUT 5V has dropped to 4.98V.

Theory - This particular monitor, (and maybe CRTs in general), is more sensitive than the TFT to some signal from the indivision. This signal is sensitive to the 5V level. I'm kinda expecting that reconnecting the keyboard and floppy will pull it down enough to cause the fault again, but with luck removing that extra socket will help to keep it just the right side of working.

Failing that, what is the best alternative for a replacement power supply? I want it to be tidy and silent like the brick. No fans and wire harnesses trailing about like an ATX.

[FOL]

Have to ask, have you tried the ECS config tool?

[jimbob]

Originally Posted by FOL

Have to ask, have you tried the ECS config tool?

Yep, I had adjusted the user settings to fit the screen nicely but I have set it all back to default since the problem appeared. Only thing I have done now is set the mode to 1.25 for 60Hz. The monitor reports 39KHz, 62,5Hz. This should be OK, on the PC there was no problem from 37.7Khz -> 63.9Khz @ 60Hz. 1.00 setting for 50Hz behaves the same way.

I'm not sure it could be anything in the config, as the trouble only seems to occur on the first cold boot with a weak 5V. To elaborate on my above guesswork, maybe once the indivision has warmed up a bit, the problem caused by a slightly weak 5V goes away. (so long as the 5V isn't too bad, like the ATX with fully loaded amiga). The time it takes for the problem to go away after a cold boot suggests to me something warming up anyway.

[Photon]

But you had it working even with the weak 5V perfectly with the LCD monitor?

[jimbob]

That's right, no trouble at all on a TFT monitor. I just put it all back together and I think it is a bit better. Though still not perfect, the display stabilises a lot quicker, then again it is a much warmer day here today so who can say.

At this point the only thing I can think to do would be to try and recreate the effect using a custom timed resolution in windows but frankly, I can't be bothered. If the problem gets worse, I'll look into a different PSU.

A bit embarassing when I try to show off the blazing boot times and the screen takes 5 minutes to come up.

[Photon]

"That's not right." (c) Galaxy Quest.

You could try to get a second opinion from a borrowed CRT or something maybe. Maybe it's the monitor ID pins that aren't correctly sent from the CRT, and Windoze lets the driver force a resolution while Indivision waits for an ID before enabling output and. Just a wild guess, though.

[jimbob]

Originally Posted by Photon

"That's not right." (c) Galaxy Quest.

You could try to get a second opinion from a borrowed CRT or something maybe. Maybe it's the monitor ID pins that aren't correctly sent from the CRT, and Windoze lets the driver force a resolution while Indivision waits for an ID before enabling output and. Just a wild guess, though.

Interesting, I don't know anything about how this works but from the VGA pin out I gather there is a 5V line for DDC, (display data channel). Might this be relevant, given the suspect 5V in the amiga?

Also, I found the manual for the monitor -

Horizontal sync width should be more than 1.0μs
Horizontal blanking width should be more than 3.0μs
Vertical blanking width should be more than 500μs

and

Even if the frequency is within the proper range, some graphics boards may have a sync pulse that is too narrow for the monitor to sync correctly.

I have no other CRT to test with anymore. I do have a CRO but I don't really wanna go there

[Photon]

Yeah, well, was thinking you could get hold of another one cheap or borrow one.

Looking at these standards, I think Indivision doesn't care about monitor ID nor output DDC or read EDID, but I'm not sure.

Could be the sync width, but I suspect a component in the monitor that's bad and goes good in 5 minutes. This is just guesswork; why not ask your reseller for support?

[jimbob]

Thanks for your thoughts Photon. I'm as certain as I can be at the moment that neither device is really faulty as they both work perfectly in other set ups. Flakey 5V upsetting the sync timing just enough to upset the monitor is my best guess. I will get round to trying out a stiffer PSU at some point.

Meanwhile I have an excellent workaround. Leave much shorter periods of time between power ups of the Amiga

[Photon]

But both 4.82V and 4.98V is perfect. (Have you measured this by the Indivision output, though?) Also, symptoms worsen when you have better 5V, although it totally depends on the 5V Amperage of the ATX PSU.

You may indeed have a bad 5V, with two different PSUs or a creeping short/crack on the mobo that fixes itself with heat, but no CRT in working order is that picky about sync pulses. It's normal TTL levels.

All CRTs require the HSYNC to be "XORED" with a gate chip into VSYNC, too keep the flyback oscillator going. You can measure this (on vsync pin) with a multimeter that has a frequency measurement mode; it could read ~31kHz or vary and that's fine then, but if it settles on a frequency (ie. the vertical frequency), the Indivision doesn't support VGA properly.

I think Indivision does (and other users could confirm if it has this?) If it works with any CRT, it has it. Whereas few LCDs require it.

Bit of a stretch I agree, especially since it doesn't explain the "works when warm" behavior, but it was the one thing that supports "all hardware in working order but difference in sync between LCD and CRT".

Anyway, just some last remaining thoughts from a desperate cornered small mammal. Very strange problem this.

[jimbob]

But both 4.82V and 4.98V is perfect. (Have you measured this by the Indivision output, though?) Also, symptoms worsen when you have better 5V, although it totally depends on the 5V Amperage of the ATX PSU.

No, Symptoms were worse with the ATX which has a lower 5V. I measured at the drive light connector. Not to sure where to probe on the Indivision.

a creeping short/crack on the mobo that fixes itself with heat

I wish you hadn't said that! I was thinking more like component value drift with temperature. I'm just gonna put my fingers in my ears and keep thinking this

All CRTs require the HSYNC to be "XORED" with a gate chip into VSYNC, too keep the flyback oscillator going. You can measure this (on vsync pin) with a multimeter that has a frequency measurement mode; it could read ~31kHz or vary and that's fine then, but if it settles on a frequency (ie. the vertical frequency), the Indivision doesn't support VGA properly.

I could try to get the scope on this but it's a bit tricky to get at the signals, plus I'm not sure I could identify an temporary fault like this, unless it was really obvious like the sync pulse visibly stretching as it warmed up.

I'll just put up with it until a CRT to test against shows up for free around here and I can dig out some other PSUs to try. It's much more fun just to keep the Amiga on a lot than try to fix this. Thanks

[Photon]

"You must do what you feel is right, of course." (c) Obi-Wan



Joking, test CRT sounds like a good idea.