Issue 3 Spectrum 48K dodgy output

Lardo Boffin
Posts: 2160
Joined: Sat Nov 26, 2016 2:42 am

Issue 3 Spectrum 48K dodgy output

Post by Lardo Boffin »

I got an issue 3 Spectrum a while back which I finally got round to doing something with - in this case I have got as far as composite modding it (as it makes it easier to test while doing all the other stuff).

The picture alternates between light green, light blue or black and white with stripes at an angle when just plugged in doing nothing:

42D03EAA-44ED-4D0C-B020-B5080DE7E2B2.jpeg
FBEAE212-E2DB-4222-B534-62DFB7D986C7.jpeg
24A817AA-5B33-4353-B123-B18FF81039CC.jpeg
E566EE60-D327-4FD7-8E9D-C841A5F9FFD9.jpeg

I have not done anything else with it yet - my plan was to replace the voltage regulator and electrolytic caps.

Any ideas what could cause this?

Many thanks

Lardo
ZX80
ZX81 iss 1 (bugged ROM, kludge fix, normal, rebuilt)
TS 1000 iss 3, ZXPand AY and +, ZX8-CCB, ZX-KDLX & ChromaSCART
Tatung 81 + Wespi
TS 1500 & 2000
Spectrum 16k (iss 1 s/n 862)
Spectrum 48ks plus a DIVMMC future and SPECTRA
User avatar
Tiger
Posts: 88
Joined: Tue Sep 01, 2020 10:37 am

Re: Issue 3 Spectrum 48K dodgy output

Post by Tiger »

I think it is more a problem with "scaling up" the video signal on modern TV's. Also the video lead can be the reason for "switching color" (turn the plug a little in the jack - this will show the effect). Most you'll get poor signal quality on 4K-(or 8K) TV's. Most "normal" HD-TV's are ok. Sometimes it helps to take an other chinch cable - I found that some cheap cable work better than "High-End" products. It is also possible that the plug of the modulator is corroded or has no proper connection. Best is to clean it and bend it a little bit...

Stripes can also occur when the modulator isn't switched off. The + Volt connection at the modulator must cut.
By the way it seems there is a long wire used for video mod - that's also a problem for "dodgy signals".
hellraiseruk
Posts: 18
Joined: Tue Oct 12, 2021 5:58 pm

Re: Issue 3 Spectrum 48K dodgy output

Post by hellraiseruk »

Might be worth checking pins 15-17 on the ula, as well as the LM1889. I'd recap it first as dodgy caps can cause problems...

There's a few posts in here about colour issues, have a read 😀

JD
Lardo Boffin
Posts: 2160
Joined: Sat Nov 26, 2016 2:42 am

Re: Issue 3 Spectrum 48K dodgy output

Post by Lardo Boffin »

Cheers both.

I have shortened the wire to the modulator, cleaned the socket etc. but no difference.

BB09839B-5449-4887-ACA7-BD39D6B6B49C.jpeg
The TVs I used to test with work fine with my other composite modded spectrum and many other 8bits so are not too picky.

All of them produce a different but poor result - the CRT TV has a black and white output, the non-HD LCD TV also has black and white but with diagonal stripes, the HD ready LCD TV has alternative green / blue screen and the large HD TV in the living room has a steady blue screen. Clearly each is interpreting the poor signal differently. I also tried a cheapo comp cable.

Time to do some reading. :geek:
ZX80
ZX81 iss 1 (bugged ROM, kludge fix, normal, rebuilt)
TS 1000 iss 3, ZXPand AY and +, ZX8-CCB, ZX-KDLX & ChromaSCART
Tatung 81 + Wespi
TS 1500 & 2000
Spectrum 16k (iss 1 s/n 862)
Spectrum 48ks plus a DIVMMC future and SPECTRA
User avatar
Tiger
Posts: 88
Joined: Tue Sep 01, 2020 10:37 am

Re: Issue 3 Spectrum 48K dodgy output

Post by Tiger »

As I can see you take the video signal directly from emitter TR 2 ( connector 15A at the expansion bus) without an capacitor between. I for myself prefer the solution with an 100µF cap. Also I made the mod with an transistor, but this produces sometimes poor pictures an some TV's. In your case I would replace TR 1 and TR 2 (ZTX 313) with BC549. Recapping (in this case specially C25) will do the rest. I think you'll get this "swapping" signal quality because C25 is in a very poor condition and all the other will not be very much better...
Lardo Boffin
Posts: 2160
Joined: Sat Nov 26, 2016 2:42 am

Re: Issue 3 Spectrum 48K dodgy output

Post by Lardo Boffin »

I have never tried the capacitor approach - never needed to before.

Regardless, fully recapped and new regulator but no difference:

D39C8975-2F4F-4750-A752-FA49892D1D6B.jpeg

Transistors next I guess, once I have ordered some. Based on this http://blog.retroleum.co.uk/electronics ... ideo-mods/ I assume they have to be fitted opposite to the silk screen?
ZX80
ZX81 iss 1 (bugged ROM, kludge fix, normal, rebuilt)
TS 1000 iss 3, ZXPand AY and +, ZX8-CCB, ZX-KDLX & ChromaSCART
Tatung 81 + Wespi
TS 1500 & 2000
Spectrum 16k (iss 1 s/n 862)
Spectrum 48ks plus a DIVMMC future and SPECTRA
User avatar
Tiger
Posts: 88
Joined: Tue Sep 01, 2020 10:37 am

Re: Issue 3 Spectrum 48K dodgy output

Post by Tiger »

"I have never tried the capacitor approach - never needed to before." - yes, you'll have no problem with most TV's. But it is a security option to prevent interferences/spikes between the power supplies.
What happens if you disconnect the cap soldered on pin 3 of the LM1889? Do you get a better picture then?

Yes the BC549C have to be fitted opposite!
If TR1/TR2 will give no better results, check C30 ( there must be +3V on pin 13 ULA) and X2 ...
Lardo Boffin
Posts: 2160
Joined: Sat Nov 26, 2016 2:42 am

Re: Issue 3 Spectrum 48K dodgy output

Post by Lardo Boffin »

Its now a capacitor comp mod and I found a couple of Bc549s and fitted them. I also temporarily disconnected the cap connected to pin 3 of the LM1889.

No changes - low res TVs are still black and white and HD alternate between blue and green.

AA1D1EF6-570F-42FA-A3B2-B86857BB8D79.jpeg
Where exactly is C30?
ZX80
ZX81 iss 1 (bugged ROM, kludge fix, normal, rebuilt)
TS 1000 iss 3, ZXPand AY and +, ZX8-CCB, ZX-KDLX & ChromaSCART
Tatung 81 + Wespi
TS 1500 & 2000
Spectrum 16k (iss 1 s/n 862)
Spectrum 48ks plus a DIVMMC future and SPECTRA
hellraiseruk
Posts: 18
Joined: Tue Oct 12, 2021 5:58 pm

Re: Issue 3 Spectrum 48K dodgy output

Post by hellraiseruk »

Set border and paper colours the same, then work your way through the colours measuring the voltage on pins 15-17 on the ula and post the results - it could well be the ula that is faulty. Alternatively replace the lm1889 but I'd do the ula check first as it's quick and costs nothing! Also check tr8 and tr9 as they are used for colour as well.

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

Re: Issue 3 Spectrum 48K dodgy output

Post by Tiger »

I think there had must be a difference result by disconnecting C 76 (47nF) from pin 3 of the LM1889. If there is no change it or the LM1889 could be damaged. But first do the tests wich hellraiseruk suggested and X2 (4.433619MHz) is still suspicious. I think you'll find the solution soon.
Post Reply