Having issues fixing comp video on +2 grey.

[HexTank]

OK, issue solved. I found a TV with composite input and sure enough, it works (as in no pixels in border), so my theory if not the TEA it would be after it was correct.

I guess TVs have a better filtering circuits in them and capture devices and are very particular, damnit Amstrad.

FML.

Was a great journey, though.

Thanks.

[Ast A. Moore]

damnit Amstrad.

Heh. Amstrad has nothing to do with it. No Spectrum ever produced a hundred-percent–compliant video signal, be it RF or composite. It’s because TVs (old TVs in particular) are quite tolerant of the little discrepancies that they can produce decent picture coming from the Speccy. A great deal of modern TVs will have a hard time doing that.

The RGB output of the 128K machines is fine, though, and the picture is excellent.

[HexTank]

Heh. Amstrad has nothing to do with it. No Spectrum ever produced a hundred-percent–compliant video signal

Citation? As it looks like you're confusing crappy picture with compliance.

[Ast A. Moore]

Citation?

You can take my word for it. The VSync and HSync timings are off (and the VSync is basically faked), the waveforms are slightly glitchy (depending on a particular ULA revision), and the picture is progressive, instead of being interlaced. This is fine for analog circuitry—old TVs can still maintain sync lock—but digital TVs will trip over the Spectrum’s fake VSync pulse.

As it looks like you're confusing crappy picture with compliance.

Oh, there’s confusion about this, all right, but none of it is on my side.

[HexTank]

Wow, don't stroke your ego too hard.

I'd like to think you've not bothered to read the thread and just jumped on board because my Amstrad comment touched a nerve.

I can only say stuff based on facts I see first hand.

I have four other spectrums, two pluses and two rubber keys, only one of these has been recapped, and none of them show symptoms that this Amstrad is showing. The clocks to the TEA chip, on my particular board seem reasonably in spec, and here's the kicker, which had you read the thread , might have have stopped you being obtuse.

I'll group my findings here, to make it easier, in the off chance you might have some musings for experimentation that might locate the issue, as I believe it's fixable.

This isn't 100%, it's determined upon power on.

If I short the luminance delay output and input pins, it fixes without the need to power cycle ( hence this having zero relevance to ULA clocks )

Pressing reset has no effect, but my current theory is some inital capacitor charge which gets the whole show rolling on power up.

The location and the artefact itself when happening is 100% consistent.


Oh, and I was testing on a newfangled digital TV, Logik I think the brand is, a proper dumpster TV, this handles it absolutely spot on.

[RMartins]

Just curious, but does it output the same problem with the capture card and a regular TV, when connected directly ?

I have had lots of problems with capture cards, that do not like the fact that the Spectrum only sends odd PAL frames.
i.e. every single frame is odd, so it does send 50 frames, but all odd. An even frame is never sent.

[HexTank]

You mean at the same time? If so I have no means of coupling the signal to test at the moment.

But I'm using the same rgb to composite cable for both tests, direct in to the capture stick and direct to the TV.

I have a few leads to test out, one being the RAMP signal to the TEA, so any naughtiness on the 12v line could be a factor. If it was always screwed I'd probably leave it and move on, but there's clearly something that can be fixed, and it's fun. but it being inconsistently consistent

[1024MAK]

The reset button only resets the Z80 processor (plus anything connected to the same reset line, like the memory bank register, the AT sound chip and any expansions if fitted), the ULA and the rest of the video circuitry will be unaffected. Only a power cycle will change the ULAs operation by forcing the circuitry to restart.

Point of order: most 1980s eight bit computers do not produce a fully complaint TV standard video signal.

Some capture cards and some LCD / ‘LED’ TVs definitely don’t like the video output from some 1980s computers. As some of these end up as edge cases (nearly working, or only just working), comparing different models of the same make/type sometimes results with some working and some not working. I have various Atari ST/STFM/STE machines and some VGA monitors. With the Atari’s configured to the high res. mode, one Atari will work with one monitor, but another Atari will not work with the exact same monitor.

Unlike CRT displays, LCD and ‘LED’ monitors sample and capture the incoming video signal. Then try to work out what has been received. Convert it to a format suitable for a LCD, then send it to the LCD.

Also for some reason, some LCD TVs are more tolerant of a sub-standard signal via the UHF TV aerial input compared to the composite video input.

By messing with the 2000 chip, it is possible that the resulting ‘blip’ in the received video signal is causing the video capture card to resync to the video signal, thus ‘clearing’ the fault.

Mark

[Ast A. Moore]

Wow, don't stroke your ego too hard.

I read the thread quite thoroughly, and if someone’s ego was hurt, it certainly wasn’t mine. I’ve a pretty thick skin. I simply pointed out that your supposition of Amstrad’s machines being an edge case was incorrect, and clarified that no Spectrums produced a standards-compliant video signal. You asked me for a “citation” (hello, hurt ego) and suggested I didn’t know what I was talking about (“confusing a crappy picture with compliance”). I then explained—in greater detail—what exactly was the culprit of the out-of-spec behavior, and now it seems to me you got your panties in a bunch.

Trust me, you’re not the first person who’s observed this kind of odd behavior of the Spectrum (or many other micros from the 80s).

[HexTank]

Thanks Mark, that's actually some useful information that can be actually taken away,

The last paragraph in particular, it might be a case that it's causing a blip forcing a resync making it work, but it's a blip that has a 100% constant. Shame there's not a collection of edge cases to view.

[1024MAK]

There has definitely been reports of problems when trying to use some capture cards in the past. But most users only had a single ZX Spectrum to test with.

The normal advice with display problems, is to try and test a ZX Spectrum with a CRT TV or with as many different LCD TVs as they can....

Mark