Time to go back in time again - ZX80 replica build

[rubber_jonnie]

Yeah it's very frustrating.

[rubber_jonnie]

Having checked IC's 11,12,13 and 20 and finding them all good, looking at the data path from the RAM, and the fact that the normal/inverse video links are there too (I know the RAM is good as it's known working from another machine) it would seem that IC9 would be a prime candidate to be the issue here as if that is not outputting, then I can't see that the video signal output that is mixed with the output from IC 9 is going to produce a signal that contains any data.

Also, if I flip the video from C-A to B-A, I see no difference in the image at all which is suggestive of a problem in the IC 9 area.

I'll check that next but if anyone has any thoughts I'd be interested (And grateful) to hear them.

[rubber_jonnie]

I have scoped all of the address/data lines on the Z80, ROMs etc and it seems to be doing what I would expect a normal machine to do. CPU has the expected 3.2Mhz clock.

Looking further it seems the shift/load pin (Pin 1) is remaining high on IC9. I'd expect that to be flipping between high and low in order to read and store the memory contents from RAM and then output them on pin 7 to be mixed with the video signal at IC 20.

Pin 4&5 at IC20 have zero signal, meaning there is no video signal coming from pin 6 and what I'm seeing on screen with the white background is simply the SYNC signal and by checking at IC 19 pins 5 & 6 I can verify that part is working.

Tracing things back there seems to be a clock coming from IC 18 pin 9. It's fine there and it's fine up to the left hand side of C11, to the right of that at IC11 & IC16 pins 2 and 9 respectively, the clock signal is very odd and very low. C11 has the correct 47pF cap.

This means that that the shift/load signal from IC 16 is not happening and likely the whole video section of IC's 11, 12, 13 and 20 is not working. I need to figure out why that clock is not getting through and is seemingly quenched after C11 as that is why I just get a white screen.

It doesn't explain why I see that odd shadowy screen and can just about see on screen activity when I pick up from the spot for C3 (Not installed) nearest the modulator.

Basically the lack of clock to the video section means all I get is sync and a nice white PAL screen.

[rubber_jonnie]

OK, so some progress made.

Having looked at the shift/load, and clock signals of IC9 plus the clocks over at C9 and C11 it became clear they didn't look right, yellow being the input to the cap, blue being the output:

The attachment C9_Input_Output.jpg is no longer available

C9_Input_Output.jpg (238.43 KiB) Viewed 101 times

So after a bit of a chat with somebody over at VintageRadio.net I wondered what would happen if I lifted the pull downs R2 and R25 since they influence the clocks as they are pulldowns:

Screenshot 2025-10-05 at 18.53.16.png (142.9 KiB) Viewed 101 times

Lifting both made no difference, neither did lifting R2, however with just R25 lifted I get this:

Output!!!.jpg (234.63 KiB) Viewed 101 times

This is good progress. It still isn't right as the text is garbled when pressing a key, but I'm now wondering if the 470R pulldowns are just too strong. I suspect back in the day it was fine, but with newer IC's it is too much.

[PhilC]

Great to see you've made good progress with this

[rubber_jonnie]

Great to see you've made good progress with this

Thanks

[HigashiJun]

You're getting closer.

[rubber_jonnie]

You're getting closer.

Indeed.

After discussions across several sites I spent some time quite late last night trying different values of pulldown resistor for R25 to try and make the clock a bit nicer. Here are the before and after shots, yellow is pin 2 of IC 11 (3.2Mhz), blue is IC9 pin 2 (6.5Mhz):

IMG20251005233612.jpg (217.52 KiB) Viewed 70 times

With the clocks like this there is no display.

By changing the pulldown to 2K from 470R the clock at IC11 now looks like this (Yellow again):

IMG20251005233718.jpg (204.01 KiB) Viewed 70 times

I then get a workable display:

IMG20251006000504.jpg (195.55 KiB) Viewed 70 times

IMG20251006000529.jpg (280.12 KiB) Viewed 70 times

IMG20251005234557.jpg (234.69 KiB) Viewed 70 times

IMG20251005234309.jpg (165.22 KiB) Viewed 70 times

Of course it was a rather unscientific test, and things are not exactly quite right because the K cursor is missing it's black background.

Interestingly, the graphics characters are unaffected and work just fine. More tweaking required, I could do with some pots so I can tweak to find the perfect resistance.

I may also need to tweak R2 which is the pulldown for the clock for IC9.

I have at least been able to prove what I'd always suspected, that the machine does run.

Since IC11 is so important as it feeds the parts of IC12 that generate video, not to mention IC16 from which the shift/load signal for IC9 is generated, clearly the clock signal for that must meet certain parameters or it isn't going to work.

I expect when the ZX80 was first around there were considerable differences in the ICs compared to now and the original design worked just fine.

[HigashiJun]

[rubber_jonnie]

Not out of the woods yet, but very much closer.