Spectrum +2 (grey) jailbars.

[Lynxman]

Is there a known way to eliminate the jailbars on a +2 with RGB out? It's really only visible on the white loading screen.

I have recapped the whole machine and it didn't eliminate them. I tried increasing the size of the electrolytic cap next to the RAM from 22 uF to 47 uF which I think helped slightly.

It's not super bad but I'd like to get it perfect.


There's also slight ringing but I think taht comes from not using coaxial between the port and SCART.

[Ast A. Moore]

I have virtually eliminated them on my Toastrack. There’s a good chance you can get similar results on your +2.

[R-Tape]

Heck - not a hardware person, but I have these on my +2 too. I actually thought it was a "feature".

It's been a while, but presumably the next time I send a broken Speccy to Mutant Caterpillar, and ask for this to be remedied as an aside, they'll understand all this.

What a difference a clean 5V rail makes!
Replaced C28 (22 µF, near the ULA) with 47 µF, and C8 and C7 (22 nF ceramic axial jobbies next to two RAM ICs) with 1 µF multilayer

[Ast A. Moore]

Heck - not a hardware person, but I have these on my +2 too. I actually thought it was a "feature".

Heh. No, it’s just that the RGB output on 128K Spectrums was never meant for analog TVs, so they didn’t bother to clean up the 5V rail. A digital monitor would (for the most part) ignore the small intensity fluctuations and display a much cleaner picture without any modifications done to the Speccy.

It's been a while, but presumably the next time I send a broken Speccy to Mutant Caterpillar, and ask for this to be remedied as an aside, they'll understand all this.

Ouch. For the kind of money they charge, you could probably recap half the Spectrums on the planet.

[Jamhamster]

I joined just to report that I have come up with a fix that gets it near enough perfect. This problem has been annoying me since before Rick Astley promised not to give anything up!

I got the fix together with more luck than skill and help from Brendan Alford. Swap C52 from a 22uf capacitor to a 100uf Capacitor.
Add 8 1uf ceramic capacitors in positions C12 - C19 (underside of the board is best). I have a guide and the results on my channel here:

https://www.youtube.com/watch?v=prFIgkQS_A0

[1024MAK]

You guys really must specify which board you are talking about, as there are different issue boards.

For example, on the Z70700 ISS.1 (manufactured in week 41 in 1986) that I have in front of me, C52 is the decoupling capacitor for the +5V to the ULA internal interface circuitry, but will also decouple the +5V supply for the nearby glue logic and DRAM chips. But being an electrolytic, it’s rubbish at high frequency signals (meaning, there is not a lot of point in increasing it’s value as larger value electrolytics are made for low frequency capacity rather than high frequency use). There may be some benefit if you use a ‘low ESR’ type in this position.

You may want to renew C63 as well, as this also decouples the main +5V supply.

C56 to C59 and C72 to C79 are the high frequency decoupling capacitors for the DRAM chips. They are all 100nF.

C53 (100nF) is the high frequency decoupling capacitor for the main ULA supply.

So I would suggest the following:
Renew C52 with the same value and connect a 470nF or 1uf ceramic capacitor across the legs of it.
Replace half (every other) of C56 to C59 and C72 to C79 with either 470nF or 1uf ceramic capacitors.
Fit a 470nF or 1uf ceramic in place of C53.

This may be different on different board versions...

Mark