48K issue 3 - colours all wrong!

hellraiseruk
Posts: 18
Joined: Tue Oct 12, 2021 5:58 pm

48K issue 3 - colours all wrong!

Post by hellraiseruk »

Hi,

Recently got hold of a rubber key Speccy with a keyboard fault - replaced the membrane, then noticed when testing that I had no colour. Thought nothing of it at the time, and assumed it would most likely fix itself after a re-cap. Replaced the usual caps, along with c5-c8 for the "jailbars" issue, composite modded, replaced the 7805 with one of those little switching dc power things, works like a charm but still no colour.

So, swapped out the 4.433MHz crystal, same issue, has to be the LM1889 - replaced this, and I now get some colour, but nothing like what I should get - magenta border is a dirty green for example, whilst green border is a weird light blue - also this colour drops out and goes back to monochrome with an interference pattern instead of colour.

Sniffing around the pins of the LM1889 I get the following voltages:-

1 - 9.21v
2 - 4.77v
3 - 4.88v
4 - 4.81v
5 - 0v
6-9 - 0v
10 - 4.29v
11 - 4.25v
12 - 11.39v
13 - 4v
14 - 11.6v
15 - 11.49v
16 - 11.6v
17 - 8.1v
18 - 9.22v

From what I gather most of those voltages seem about right with the exception of pin 17 - why is it so high? Anyone got any ideas as to what may be causing this?

I ought to add I've tested this on a number of LCD TVs and also a scart-HDMI adapter, and all manifest the exact same symptoms.

I've replaced TR1 and TR2 as I thought this may be the cause of the issue, still no change. Also replaced TR8 and TR9, again no change.

I've attached 2 JPEGs of what I'm seeing, anyone got any ideas?

Today I've done some further testing and pins 15 and 16 of the ULA show 2.85V and 3.8V respectively when showing the copyright notice. Is this a sign the ULA is the issue? I'm running out of bits to change :( I've also switched to a 7V switchmode PSU I had lying around, this gives me slightly lower voltages on the LM1889, but the overall symptoms are identical.

Cheers

JD
Attachments
this is the weird colour dot crawl I get on the copyright notice
this is the weird colour dot crawl I get on the copyright notice
This is what I get when colour comes in
This is what I get when colour comes in
This is when colour drops out.
This is when colour drops out.
User avatar
1024MAK
Posts: 5101
Joined: Mon Sep 26, 2011 10:56 am
Location: Looking forward to summer in Somerset, UK...

Re: 48K issue 3 - colours all wrong!

Post by 1024MAK »

IC14 (LM1889) pins 14, 15 and 16 should all be the same voltage (about 12V), as they are all connected together.

I recommend a minimum of 8V supply voltage input. 7V is a bit low for the simple DC/DC converter (based on TR4 & TR5) to cope with in trying to boost it up to 12V for the 12V rail.

Is the ULA in a socket? If yes, are pins 15 and 16 making good contact?

Have you tried putting the original crystal back in?

Do you have a spare ULA to swap in?

TR1, 2, 8 and 9 very rarely fail, because they are not worked very hard. Although TR2 can fail if the composite video output is shorted to 0V/GND. But you had a video signal, which proves that TR1 and TR2 were fine.

Mark
ZX81 Variations
ZX81 Chip Pin-outs
ZX81 Video Transistor Buffer Amp

:!: Standby alert :!:
There are four lights!
Step up to red alert. Sir, are you absolutely sure? It does mean changing the bulb :!:
Looking forward to summer later in the year.
hellraiseruk
Posts: 18
Joined: Tue Oct 12, 2021 5:58 pm

Re: 48K issue 3 - colours all wrong!

Post by hellraiseruk »

Hi,

Yes, ULA is socketted, I've reseated it with no difference, don't have a spare to test with unfortunately.

I've also tried with the original sinclair power supply, and get the same issue. Does this seem to be pointing to the ula? I've also tried swapping back to the old crystal, with no change in symptoms.

Thanks for your help 👍

Jd
User avatar
Tiger
Posts: 88
Joined: Tue Sep 01, 2020 10:37 am

Re: 48K issue 3 - colours all wrong!

Post by Tiger »

I remember having a similar fault and there was C69 broken. Indeed, the cap itself had a fine cut between case and wire! I discovered it as I touched this area with my fingertip - the color changed and so I could localize the damaged part (I also swapped the ULA and LM 1889 before with no result). I think TR 8/9 could also produce such a fault. But it is also possible that C72/C38 R44/45 was damaged by replacing the crystal X2, or some solder mistake in this area...
hellraiseruk
Posts: 18
Joined: Tue Oct 12, 2021 5:58 pm

Re: 48K issue 3 - colours all wrong!

Post by hellraiseruk »

Hi,

Checked around that capacitor and no sign of problem, checked R44/45 which are measuring 1k and 2k, c72/c38 all look fine, and double checked around the crystal for issues, but no sign of anything untoward. The issue was apparent before I replaced the crystal and has not changed in the meantime, so it looks like the fault lies elsewhere.

I may have to order a vLA82 or similar and see if that helps, if not at least I can sell it on :)

I've also just done the DC-DC mod as it's an issue 3 board, this has made no difference either but I didn't expect it to, just didn't want the board to blow up once I'd found the fault which is what usually happens :)

Any more help would be appreciated, and thanks to all those who have helped so far!

JD
hellraiseruk
Posts: 18
Joined: Tue Oct 12, 2021 5:58 pm

Re: 48K issue 3 - colours all wrong!

Post by hellraiseruk »

Hi,

Well, as is always the case noone has stocks of either nebula or vla82 now I'm in a position to order one! So, whilst I wait for stocks, anyone got any other ideas?

Cheers

Jd
User avatar
1024MAK
Posts: 5101
Joined: Mon Sep 26, 2011 10:56 am
Location: Looking forward to summer in Somerset, UK...

Re: 48K issue 3 - colours all wrong!

Post by 1024MAK »

Maybe, but I’ll be busy with real life until Thursday or Friday…

Mark
ZX81 Variations
ZX81 Chip Pin-outs
ZX81 Video Transistor Buffer Amp

:!: Standby alert :!:
There are four lights!
Step up to red alert. Sir, are you absolutely sure? It does mean changing the bulb :!:
Looking forward to summer later in the year.
User avatar
Tiger
Posts: 88
Joined: Tue Sep 01, 2020 10:37 am

Re: 48K issue 3 - colours all wrong!

Post by Tiger »

Charlie Ingley from New Zealand offers them:
https://www.vretrodesign.com/products/v ... eplacement
hellraiseruk
Posts: 18
Joined: Tue Oct 12, 2021 5:58 pm

Re: 48K issue 3 - colours all wrong!

Post by hellraiseruk »

1024MAK wrote: Mon Oct 18, 2021 8:57 pm Maybe, but I’ll be busy with real life until Thursday or Friday…

Mark
That's fine, I'm in no rush, just want to solve the mystery ...

Cheers

Jd
hellraiseruk
Posts: 18
Joined: Tue Oct 12, 2021 5:58 pm

Re: 48K issue 3 - colours all wrong!

Post by hellraiseruk »

OK, decided to have another look today - I've been over the PCB with a magnifying glass, and no sign of any issues, so I thought I'd check voltages on pins 15-17 on the ULA. Border and paper set to each colour, so I have a full screen the same colour, this is what I get:-

1 Blue 15 - 2.452v 16 -3.95v 17 - 2.447v
2 Red 15 - 3.028v 16 - 3.95v 17 - 2.342v
3 Magenta 15 - 2.451v 16 - 3.95v 17 - 2.234v
4 Green 15 - 3.46v 16 - 3.95v 17 - 2.123v
5 Cyan 15 - 2.854v 16 - 3.93v 17 - 2.017v
6 Yellow 15 - 3.46v 16 - 3.94v 17 - 2.041v
7 White 15 - 2.847v 16 - 3.94v 17 - 1.814v
0 Black 15 - 2.856v 16 - 3.95v 17 - 2.565v

So I'm confused - surely pin 16 is supposed to vary and not be the same constantly? Does this point to a ULA issue?

JD
Post Reply