Flaky ZX81

[Lardo Boffin]

I have just tried it with two working Memotech 16K RAM packs and both produce the same effect as nothing being plugged in - grey screen, white screen or K (after the usual 16K delay).

So I’m guessing its not the RAM?

I’ve used the same voltage regs on at least 20 zeddies without issue before. I bought 4 of them this time round and the first two went into zeddies fine (one was an issue three).

I think I will put the old voltage reg back in and see if that helps any just to be sure.

Not sure I’m brave enough to try and do a reset!

[Moggy]

Just a faint stab in the dark Lardo.

I see this an issue three board which means either yourself or someone else has socketed this as the chips are not socketed on this board.
As you are aware I'm sure, some of the ULA pins are soldered top side and if this isn't soldered well it can be a bit intermittent regardless of whether a multi meter shows continuity between board and said pin.

You may recall when doing some experimenting with my ZX80 I thought it was totally borked (blank screen) until Mark insisted I check the CPU socket connections. This I did but the multimeter showed all ok even so Mark was still insistent about the socket.

To cut to the chase some of the soldered top side CPU socket pins looked a bit dodgy probably because of my penchant for swapping CPU's, one re-soldered socket later and my little 80 sprung back into life.

The intermittence does suggest something of this nature as the active components either work or they don't I would have thought.

[1024MAK]

Yes, I leaning towards one (or more) of the ULA pins being problematic. Either with the chip leg to socket connection or with the socket pin to PCB solder joint. Or possibly one of the PCB tracks has a hairline crack.

Mark

[Moggy]

I totally concur with Marks suggestion. In my case everything pointed to the socket pins being ok and if I had not heeded his advice I would have thrown a perfectly good ZX80 in the skip!

On a slightly different tack I have to say that is a nice clean looking board(sorry the whisky is kicking in)!

[PokeMon]

Not sure I’m brave enough to try and do a reset!

Looking back at your pic, the reset capacitor was replaced. Normally it is an bigger axial, no radial type. Are you sure there is enough distance between the board tracks and the capacitor legs as they are quite long. Just an idea. Maybe worth to isolate under the legs.
When putting it into the case, the capacitor could be pressed on the board.

[Lardo Boffin]

Thanks all!

I have just tried it again - with the ZXPand plugged in it switches on fine every time. With a memopak in it either does the grey, white screen and occasional K.

The switch on issue happened before I changed the 1uF cap. I had run out of axials (some on order and will change it) but made sure there is plenty of clearance.

I have socketed all of the chips so hopefully there are no issues there! The problem occurred before I socketed the chips. I socketed them so that I could chip swap as I assumed one of them had died. However the chips seem fine. A known good set are in there at the mo.

The only work I did before noticing the issue was to replace the regulator. I removed the original one intact so will refit it when I get chance to eliminate that from my enquiries.

[Lardo Boffin]

I refitted the original regulator and now it switches on first time every time with nothing attached!

Will try it later on with something attached to see if makes any difference.

[PokeMon]

That`s why I advised to do a manual reset after power-on.
I guessed that there could be a timing mismatch.
Solution would be to choose a larger capacitor than 1uF as reset cap.

[Moggy]

I find that Memotech's RAM modules have a tendency to not instantly fail like Sinclair's offerings, but tend to throw up various strange behaviours before giving up the ghost. Also I recall Mark saying that they do need at least 9-11 volts to work properly due to the way the chips get the 12 volts of the mixed voltages they require. They don't seem keen to work with regulated power packs set at say 7.5 or 8 volts used as an aid to cool running.

I do know one thing though that when they do blow they nearly always take the ULA with them especially Andy Rea's device which whilst splendid in operation is a bit more fragile electronically than the original in my hard learned experience and is something to bare in mind if you suspect a dodgy Memotech RAM module.

[Lardo Boffin]

That`s why I advised to do a manual reset after power-on.
I guessed that there could be a timing mismatch.
Solution would be to choose a larger capacitor than 1uF as reset cap.

Thanks. I think I have some 10uF axial caps somewhere. Would that do the trick?

The thing thats confusing me (the most) is why it worked fine with the ZXPand+ plugged in?