Sparkly pixels

[Lardo Boffin]

I saw this one based on a link from another thread on here:-

IMG_5472.JPG

Yup that's my reg of choice.

Thanks. Is it U.K. spec? Not really following the comment about it being Australian...

[1024MAK]

Prices on these DC/DC converters/regulators often change.

Here are some part numbers and links:-
Recom R-785.0-1.0
https://www.rapidonline.com/recom-r-785 ... er-84-2573
http://uk.farnell.com/recom-power/r-785 ... dp/1793143

Traco Power TSR 1-2450
https://www.rapidonline.com/tracopower- ... 1w-51-7072
http://uk.farnell.com/tracopower/tsr-1- ... dp/1696320

But also check the prices of other suppliers

Mark

[Lardo Boffin]

Well, I've sent this zeddy off to Mutant Caterpillar Games to be looked at. As soon as I have any feedback I will post what the issue was!

[Fwapp]

Looking forward to hearing what they say.

Ian's a very clever chap so it's in good hands.

[Lardo Boffin]

That's the Australian version, you want to be ordering a UK-spec component

B

Ok. Head in hands moment - I just got the aussie gag. On my iPhone the image is the correct way up not upside down!

Only just saw it on an actual computer for the first time....

[Lardo Boffin]

So, back to the sparkly pixels...

Well the good news is that it's been fixed!

The bad news is that the cause is only suspected rather than known with certainty. So much for closure!

Long story short all the chips in the sparkly (well, now ex-sparkly) zeddy work fine individually - no faults found. When combined together with a 'spand however sparklyness ensues. In the end the working zeddy ended up in this state by chip swapping until a working combination of chips was found. This initially was just the CPU replaced after some swaps.

I decided to have a ZXVid fitted while it was away being sorted and once that was done the sparklyness returned. You just can't keep a good sparkle down it would seem.
After this the CPU was put back and the ROM changed and the sparkles went away again. The assumption was that there was still interference going on but it was hidden by the RF output whereas the ZXVid made it visible again.

So what caused it? Answers on a postcard please...

I know in the world of mechanical engineering components in a system that are slightly out of tolerance can have a number of effects ranging from nothing, through excessive wear, right the way to spectacular failure. I assume the same applies to electrical stuff with timing and other interactions etc. This kind of complex interaction issue is suspected here (exacerbated by the modern 'spand) by the experts at MCG.

So there you go!

Lardo

[Moggy]

If it is some kind of interaction with the spand then it begs the question why yours, solitary out of the at least a hundred sold is there this "interaction"?

I assume they checked the resonator was chiming at the correct frequency after all it is the source of all timings?

[1024MAK]

In a computer, the timing between when the CPU (or video system) asks the memory for data (address bus and control lines) and when the memory is able to drive the data onto the data bus for the CPU to read, is everything.

Some Sinclair ROM chips can be a bit slow. Because of the way the video system in the ZX81 (and ZX80) works, the timing is more critical than people think.

Add in a modern (fast responding CMOS SRAM) chip, and in addition, you may get reflected pulses on the data bus lines. With different chips, there may have been a slightly different impedance on the data bus. This would then change the resulting reflected pulses...

Which of these (if any) caused the issue is not clear

Have fun

Mark

[Lardo Boffin]

If it is some kind of interaction with the spand then it begs the question why yours, solitary out of the at least a hundred sold is there this "interaction"?

I guess I'm just lucky!!!

Seriously though if Sir Morris has sold a couple of hundred 'spands and each owner has tried them out of 10 or so zeddys then that is a at most a couple of thousand zeddys run with spands. According to Wikipedia (which we all know is 100% accurate) they made 1.5 million of them. That is a tiny percent of zeddys that have been connected. It is always possible that other weird and wonderful issues (that are nothing to do with the spand) are waiting to be found!

I have personally tried to tune in to an old CRT TV something in the region of 15 zeddys over the last couple of years. A couple have produced near perfect pictures while the majority range from poor to unusable. That is ignoring the dark output of the early issue ULAs. Clearly there is some degree of variability in these computers else they would all produce a nice picture.

I actually had a slightly similar problem to this with a BBC Micro - in mode 7 (teletext mode) after a while the lower case 'e' letter on the screen flashed slightly and lost the bottom part of the top half of the letter (if that makes sense). All other characters were unaffected. I rectified this by replacing the teletext chip. If the solution to my sparkles was just to replace the ROM that would have made more sense to me as the character patterns in low res are taken from the ROM.

I wish I had thought to buy a ZXblast before having this fixed to see if that had the same problem but only just thought of that!

All I can say is hmmm!

[siggi]

Is the +5V power OK? Or is there a lot of "noise" or peaks or voltage drops? Maybe you have to replace old capacitors (if not already done) ...

Regards
Siggi