Issue 4B Spectrum with dot crawl in black and white

[Lardo Boffin]

I have an issue 4B Spectrum which has what can only be described as minor dot crawl in black and white - if after switching on I fill the screen with text the text is far from stable. You can’t really see it in a photo but on the tv it looks a little like dot crawl. Which you normally only get to a noticable level when colour is present.
I have tried it with a Spectra SCART and the picture is not perfectly stable with that either which it normally is.
I have replaced all capacitors and fitted a new regulator. I also replaced the Z80 with a new one as the M1 line was missing and my DIVMMC did not work.
Any suggestions? Are there any ‘service fixes’ required on these that I should do?

Thanks!

Motherboard:-

[1024MAK]

What display are you using? Keep in mind that as long as the video source supplies a colour burst signal after the line sync signal, a TV or monitor using a composite or RF input will still treat the image as a colour image, the white being the effect of the red, green and blue video channels all being on to give the white. This is not the same as a monochrome TV/monitor or a colour display receiving a monochrome video signal with no colour burst signal present.

Has this ZX Spectrum been modified to output a composite video signal? If yes, which version of the mod. was done?

The Spectra (at least the first versions that Paul sold) gives a slightly wobbly picture on one of my ZX Spectrum+ machines, which otherwise worked fine. I did not investigate any further. I forget the details now, as this was some time ago.

I see that you used a DC/DC converter rather than a 7805. Was this problem present before you fitted the DC/DC converter? Which type of DC/DC converter are you using?

If you have a spare ULA (from another machine, either the same ULA version, or a later version), try fitting that.

By “all the capacitors” I presume you just mean all the electrolytic capacitors?

Have any of the 4116 DRAM chips ever been changed? Look at the soldering to see if you can see any signs of any rework, like brown flux residue.

Is the crystal and surrounding components in good condition? If you have a suitable test gear, test the clock output from the ULA on pin 32.

Mark

[Lardo Boffin]

Thanks as ever Mark.

What display are you using? Keep in mind that as long as the video source supplies a colour burst signal after the line sync signal, a TV or monitor using a composite or RF input will still treat the image as a colour image, the white being the effect of the red, green and blue video channels all being on to give the white. This is not the same as a monochrome TV/monitor or a colour display receiving a monochrome video signal with no colour burst signal present.

A full hd LG LCD telly. It is quite a picky one but I have another 4B that doesn’t do this. The black text is rock solid. I need to try other tellies to see if it is as bad.

Has this ZX Spectrum been modified to output a composite video signal? If yes, which version of the mod. was done?

Just the simple desolder the wires and direct connection. No extra transistors etc.

The Spectra (at least the first versions that Paul sold) gives a slightly wobbly picture on one of my ZX Spectrum+ machines, which otherwise worked fine. I did not investigate any further. I forget the details now, as this was some time ago.

All of my Spectrums apart from this one produce a very stable picture using this.

I see that you used a DC/DC converter rather than a 7805. Was this problem present before you fitted the DC/DC converter? Which type of DC/DC converter are you using?

I use the Murata convertor from here.

https://www.retroleum.co.uk/spectrum-repairs-mods

I have fitted at least 5 of these in Spectrums and probably 10 in ZX81s without issue so far.

If you have a spare ULA (from another machine, either the same ULA version, or a later version), try fitting that.

I will have a look. I have a nebula ULA replacement chip.

By “all the capacitors” I presume you just mean all the electrolytic capacitors?

Yes - just those!

Have any of the 4116 DRAM chips ever been changed? Look at the soldering to see if you can see any signs of any rework, like brown flux residue.

Will have a look.

Is the crystal and surrounding components in good condition? If you have a suitable test gear, test the clock output from the ULA on pin 32.

I only have a logic probe so thats probably not going to help!

[1024MAK]

One thing that can help, if the problem is noise on the +12V supply to the 4116 DRAM chips, is to fit larger value ceramic decoupling capacitors in place of the 22nF types in positions C5, C6, C7 and C8. Or maybe just C5 and C7. The capacitors that I recommend for the best performance is 1uF multilayer ceramic types (5mm lead spacing should fit most boards). Multilayer ceramic types have a much, much better performance compared to small value electrolytic capacitors of the same value. They are available from many different suppliers.

Mark

[Lardo Boffin]

Thanks. I had seen that mentioned as a potential cure for jailbars here:- https://spectrumforeveryone.com/technic ... -spectrum/
although they suggest 150 to 220 nF. I will get some 1uF ordered and see how it goes.

[McKlaud]

220nF would be sufficient. This is not an issue with the capcistance only. In this application the main factor is ability to fast charging and discharging of these caps. So, the ESR would play the main role.

[Lardo Boffin]

Thanks. So the next obvious question would be what ESR rating should I be looking for?

[McKlaud]

Any ceramic caps (single or multilayer) are good for this application. Don’t use tantal or electrolytic ones.

[1024MAK]

220nF would be sufficient. This is not an issue with the capcistance only. In this application the main factor is ability to fast charging and discharging of these caps. So, the ESR would play the main role.

Thanks. I had seen that mentioned as a potential cure for jailbars here:- https://spectrumforeveryone.com/technic ... -spectrum/
although they suggest 150 to 220 nF. I will get some 1uF ordered and see how it goes.

Um, the idea of changing capacitors C5 to C8 as far as I know was an idea first suggested my me over on World of Spectrum forums many years ago. I originally suggested renewing C1 to C8, as the problem that was trying to be resolved was jail bars caused by inadequate decoupling of the supplies to the 4116 DRAM chips.

I suggested multilayered ceramic 1uF because (1) this is the highest value ceramic that is commonly available (with a low ESR) and I knew that the 12V supply to these chips was not very well decoupled. The DRAM does not care too much about this, but the same 12V supply is used by the video circuits, so it is best to nobble any ”switching” noise as close to the source as possible.
And (2), fitting 1uF types did have the intended effect, with users reporting a much better picture.

Since then, it has been discovered by others that there is no need to change C1 to C4. And some users are apparently happy with only changing two of out of C5 to C8. I have not tried this myself.
I keep a large number of 1uF in stock anyway. The next value down that I stock in large numbers is 100nF.

I was not aware of the spectrumforeveryone.com article. I wonder when that was first published.

Mark

[1024MAK]

Thanks. So the next obvious question would be what ESR rating should I be looking for?

They have to be ceramic multilayer type. No other type is suitable.

Ceramic multilayer type are low ESR by design.

I have recently seen some axial type at a reasonable price on eBay, Link. They are also available from the bigger electronic parts suppliers.

Mark