Issue 3 Spectrum 48K dodgy output

[1024MAK]

Check crystal X2 and the nearby components for dry joints, solder splashes, track damage or any physical damage.

Further to what hellraiseruk said, ULA pins 15 and 16 are the colour difference outputs. As he says, set the border and paper to the same colour, then record the voltage on each of these. Do this for all eight colours. If one pin always has a constant voltage, the the ULA is suspect.

Do the same with pins 2 and 4 on IC14 (LM1889). The details of the voltages are in the service manual. This chip only does the colour encoding. Test the supply voltage to it on pins 14, 15 and 16 (they should all be the same voltage).

The output is on pin 13. Test the voltage here.
Capacitor C65 (22µF on later boards, but I’m not sure what value is on your board) is the coupling capacitor that feeds the colour signal to join the monochrome signal at the base of transistor TR2. So check it’s connections (did you renew this one?). And check / test the PCB tracks between pin 13 of IC14 and it. And from the other leg of C65 to the base of TR2.

If you want, you can carefully lift one leg of C65. Then the machine should produce a rock solid and consistent monochrome image only.

If you have suitable test gear, you may want to measure the frequency that the oscillator in IC14 is running at. See the service manual for further details.

If you don’t have any suitable test gear, try replacing crystal X2.

Note that failure of TR8 or TR9 is rare. In terms of semiconductor failures, the normal causes are IC14 or the ULA.

Mark

[Lardo Boffin]

Thanks all!

Working my way through the list as and when time permits.

I lifted the leg of C65 (has been replaced) and get a rock solid picture. I checked the components around tbe modulator and nothing stands out. The X2 soldering wasn’t brilliant but reflowing it made no difference.

Measuring voltages is next. Where is the best (safest) reference point on these for measuring (presumably the 0v on each chip?)? Mostly only do ZX81s…

[1024MAK]

Suitable 0V/GND points: If it still has a 7805, the heatsink. Otherwise, the main case (not the lid) of the modulator or it’s phono output socket.
Or any negative lead of any of the electrolytic capacitors where the negative lead is definitely connected to the 0V/GND rail.

Mark

[Lardo Boffin]

Thanks Mark.

On IC 14 -

13 : 4.15V
14, 15 and 16 : 11.59V

ULA

Colour 0
Pins 15/16 :
2.26 / 2.02
Colour 1
1.67 / 2.03
Colour 2
2.42 / 2.06
Colour 3
1.86 / 2.05
Colour 4
2.62 / 2.06
Colour 5
2.05 / 2.07
Colour 6
2.84 / 2.03
Colour 7
2.26 / 2.02

Not looking hugely good for the ULA then as pin 16 is pretty consistent and colour 0 and 7 seem to match on pin 15.

[hellraiseruk]

Doesn't look great, mine failed high with pin 16 at just under 4v, but I'd definitely expect it to vary. Seems these ULAs are fragile things!

I got hold of a Nebula from Retroleum and that worked great, plus you no longer have to worry about replacing the socket as they now have fine pitch pins. Still takes a bit of force to get them in though! Might be worth seeing if he has any in stock, but first I'd let someone else reply and confirm our suspicions....

JD

[Lardo Boffin]

Well, its not the ULA. I borrowed a ULA from a known good spectrum and put it in. No difference in output. The good ULA still worked back in its original computer (thankfully).

Maybe time to replace X2 and then IC14.

If I was to carefully transplant the X2 from another Spectrum would it just work, as in be compatible?

[hellraiseruk]

Yep, its just a crystal - same as transplanting any other component. I'm surprised it's not the ula with there being no difference in voltage on pin 16, that makes little sense.

I guess sniffing around ic14 makes sense, but I'm still confused by pin 16 on the ula...

JD

[Lardo Boffin]

but I'm still confused by pin 16 on the ula...

JD

Me too! I will try the ULA from the dodgy speccy in the good one to see if it works there - should have done that first time around.

[1024MAK]

Me too! I will try the ULA from the dodgy speccy in the good one to see if it works there - should have done that first time around.

Maybe we’ll make a fault finder out of you yet!

If you have a spare working machine, it’s alway worth thinking about moving the suspect chip by swapping it into the known good machine, then if the fault moves with the chip…

Mark

[Lardo Boffin]

I took the ULA out of the dodgy spectrum and put it in a good one and that continued to work fine so the ULA is presumably fine.

The question is then what could be causing the ULA colour voltages to be wrong when a good ULA is present in the board?

I guess from a brute force perspective I could change X2 and IC14 and hope that sorts it.