Sinclair ZX80 / ZX81 / Z88 Forums

An Issue 3 I got from eBay was showing a K prompt via RF on my TV, though I did have to crank the brightness. Now I just get a white screen. I haven't done anything to this one yet, no mods or anything, so whatever has gone wrong had probably occurred just through handling the motherboard.

I think I read somewhere that you only get the K if the RAM self test is ok, so is a plain white screen usually indicative of a RAM fault? I have a 32K SRAM ready to perform the internal 16K upgrade, so should I push ahead with that, or are there a ton of other issues that could cause a white screen? What should I check? I have a multimeter.

A white screen only means that ULA is working.
Everything else may be at fault.. Luckily there is not much else at ZX81 .. and the ULA is difficult to replace..
If it was working well and suddenly it stopped I would check for any loose IC or miss contact.
If you have a multimeter first thing to check are voltages at vcc and gnd leads on every IC.. Check MAK1024 signature for a link to all the IC pinouts..
If you have a known working rampack then it will rule out the internal ram as it overrides it completely..
IA wouldn't start an internal ram modification without a good diagnosis.
Next step would be to check the Z80 CPU.. best way is to replace it with known working one.. if that's not possible then you can do some measurements on it but an oscilloscope or a logical probe are better tools for that. A rom failure is very unlikely but it can happen..

Pretty much what bola_dor says!

The ZX81 is both simple (not many chips) and blooming complex all at the same time.

In order to get a K almost everything needs to be working. This is because the Z80, the RAM, the ROM and the ULA together are all vital to producing a screen display with any pixel or text information. So if the Z80 can’t run the ROM code, or there is a problem with the RAM, then it will either crash or get stuck in an infinite loop. Then all you get is a white screen.

Which (if any) chips are in sockets? If any are in sockets, have you tried firmly pushing them to make sure they are fully seated?

I assume you have a digital multimeter?

Mark

Unfortunately, or fortunately (depending on how you look at it) this Issue 3 has all of its chips soldered to the board - so loose connections aren't likely and all the solder blobs look sound and remain nice and shiny.

I did find the problem though luckily. I remember now that at one point the other night I had gone to fetch a different screwdriver from my understairs cupboard and I had the bare ZX81 board in one hand. I bumped it against the door frame very lightly by accident as I steadied myself leaning across all the clutter in there. I thought nothing of it and couldn't even remember which one of the two zeddies it was, but I just remembered that event and took a closer look at the region that got bumped (which I did recall as I have quite a photographic memory). It was the set of resistors above the CPU and they all have quite long legs on this board. Well a few of them were very close and there must have been a short. I touched those it while it was on and connected to the TV (as this all occured to me) and bam - the K appeared . Furthermore it wasn't helped by the fact that I need to prod the trimmer in the original modulator a little to get any kind of RF picture to register on my LCD TV.

So now it works again, I'll composite mod this guy right away using the second mod I bought but without the 555 IC inserted.

I've circled in red where the resistor legs were shorted:

That’s part of the oscillator circuitry. That may have disrupted the 3.25MHz clock to the Z80. A Z80 without the correct clock signal will do nothing.

Mark

Just as Mark said. That's the clock generator.. the second lemon yellow component is not a resistor but a capacitor.. I have had troubles with the resonator of one of my boards before..
Nice you could solve that..

Works perfectly over composite. The yellow banding on the screen there isn't visible in real life so it may be a camera CCD to screen frequency mismatch artefact.
This time around I've made the wires as short as possible, and I routed the ground wire directly to the back corner of the can with a very short stub: Mark - am I correct in assuming the 10K trimmers have no purpose once the 555 is removed from the composite mod circuit?

patters wrote: ↑Sat Mar 27, 2021 3:56 pm Mark - am I correct in assuming the 10K trimmers have no purpose once the 555 is removed from the composite mod circuit?

Yes.

One is to adjust the time period that the NE555 times for (the duration of the pulse), the other adjusts how hard the transistor for the ‘back porch’ output is driven.

Both do absolutely nothing without a chip fitted.

Mark

Is it at all likely that I could have fried something in the ULA (the tape input) when C6 had shorted against R4, or are these things pretty resilient (compared to say RAM chips)? This was the same machine which currently won't load from Maxduino nor from WESPI hard wired in.

I don’t know for certain, as the long term effects of a ULA being powered up, but with the oscillator stopped is not really known. That’s not to say that we are expecting this situation to cause problems, and there are no known documents that advise against this.

The ear / tape input circuitry in the ULA is relatively simple, and I would not expect this circuitry to need a signal from the oscillator or clock circuitry.

So in all likelihood, I think it very unlikely that stopping the oscillator would cause damage to the ear / tape input circuitry in the ULA, if the ULA can still supply a clock signal to the Z80A and can produce a video signal.

Mark