exxos's Atari, Amiga & retro forum

BennehBoy wrote: 28 May 2021 19:37 Congrats & cheers! Whisky o'clock here, been diagnosing an A600 PLCC RGBtoHDMI adapter for the last 2 hours. It's giving everything a blue tinge, including the standard RGB out - anyhow looks like the flip flop might be duff as I'm getting a partial short on the red high bit. It worked when it arrived, replugged it and it's gone bad, bah.

So in short, I have some sympathy with you at the moment.

Have you tested those roms I sent in no3? No rush, just my OCD worrying that I may not have burnt them correctly :D

Bourbon o'clock here 😁 Great minds.. I haven't tested the ROMs as yet I'm afraid due to the need to adapt them slightly due to the pinouts and getting the keyboard rebuilt. I will be testing them as soon as I have the video out sorted on the poorly machine. Ordered some transistors for the video circuit on the basis that they are easy to change. I'll check what the good machine is doing for a baseline.

So, I have some display options on the horizon.

SCART cable is ordered, as is an adaptor that will allow me to connect an aerial lead to the UHF modulator, plus I have ordered the transistors that are in the composite circuit.

I think the next steps are to probe the good machine video circuit as I know that works just fine, then I can compare to the No2 machine, which I now know boots, but has no video output.

Video schematic section is sitting on my printer now awaiting my attention over the weekend :)

Hopefully the replacement power connectors will come today too and I can tidy all that up too.

I like progress.

The best video option is RGB to a CRT monitor. Second best is RGBtoHDMI. Third best is a SCART lead to a TV.

(3) has problems with a lot of TVs though because the BBC outputs nearly PAL timings... so many modern TV chipsets will drop sync. Some people report that swapping the main 16MHz crystal for a new one fixes this but I haven't tried it. I just use Samsung TVs which seem to fare better.

cmorley wrote: 29 May 2021 12:26 The best video option is RGB to a CRT monitor. Second best is RGBtoHDMI. Third best is a SCART lead to a TV.

(3) has problems with a lot of TVs though because the BBC outputs nearly PAL timings... so many modern TV chipsets will drop sync. Some people report that swapping the main 16MHz crystal for a new one fixes this but I haven't tried it. I just use Samsung TVs which seem to fare better.

Good job my TV with SCART is Samsung I guess :)

I managed to test my known good video ULA in the machine with no video today, No2, but still nothing (the machine IS booting, I get a BooBip and responds to blind typed commands, caps lock, shift lock etc), and the signal coming out of the RGB lines looks wrong compared to the working machine, No3. Similarly I tried the video ULA from the machine with no video in the good machine and it wasn't happy, but you could at least see the command prompt on boot but with blue bars and flickering. Enough at least to tell if there is any output to the TV, which there isn't on machine No2.

On the No2 machine, there is a 15Khz signal at the composite port, but the RGB lines out of the video ULA seem to have a square wave, not at all like the same on the working machine, whereby you see the pulse for the on screen text and it looks more 'analog'.

I have checked the continuity of all the RGB lines from point to point and they are fine. The next step really is to figure out why there is a square wave coming out of the RGB lines of the video ULA, because I see this even with the known good ULA installed, so it is something to do with what that's receiving.

Of interest, on boot, so in mode 7, I see the same square wave on the RGB out of the SA5050, so that suggests some connection between them. Blind typing MODE 1 causes the RGB output to stop from the SA5050, which I presume is correct, as not used in MODE 1. At this point I still see the square wave on the RGB outputs of the video ULA.

My current theory is that something feeding into the video ULA (Possible the SA5050 too) is bad, resulting in the square waves on the RGB output, which I also see at the composite socket. I need to find what is wrong with the inputs for the video ULA in order to have it generate the correct signal again. At least I have the working machine for reference now.

The next chip to test is IC2, the 6845, as this does the clock for the video processor and some other control lines. It also generates the synchronisation signals for the video.

The RGB output from the teletext chip goes directly into to the video ULA and just has a set of pull-ups on the lines.

stephen_usher wrote: 30 May 2021 08:14 The next chip to test is IC2, the 6845, as this does the clock for the video processor and some other control lines. It also generates the synchronisation signals for the video.

The RGB output from the teletext chip goes directly into to the video ULA and just has a set of pull-ups on the lines.

My thoughts exactly :)

I'll look at IC36 & IC41 whilst I'm at it.

Ok, so I promised some pictures, and here they are.

Here are the exploded Rifas:


And compared to new caps:


The freshly replaced spade connectors, with the nasty old ones in a pile in the middle:


And fitted in the machine:


And last, here is the good composite signal from the working machine on the Fisher Price oscilloscope:


And you can see the analog nature of the signal, the spikes representing the on screen text that's visible at boot time. I see this at the video chip RGB output pins too.

And here is the bad signal, looking more like a square wave:


And this is both at the composite out and RGB pins of the video chip, so the bad signal seems to originate at the video IC, so that is where I investigate next.

@rubber_jonnie is there an inverter in the rgb line? Both signals look 15khz but the second one may need inverting to match the other?

PhilC wrote: 30 May 2021 10:43 @rubber_jonnie is there an inverter in the rgb line? Both signals look 15khz but the second one may need inverting to match the other?

Not according to the schematic for the composite circuit.

Plus if I compare the good RGB out at the video IC to the bad, the bad is just this square wave, but the good is the analog signal. Since it seems bad at the source, then it doesn't matter if inverted or not.

I'm deliberately using the composite port because it's the simplest circuit, and because it works on the good machine for easy comparison.

Without the correct signals from IC2 the output from the video ULA will be nonsense. Check there first.