ZX +128 12v rail issues

[bola_dor]

Hi, I'm refurbishing a Toastrack for a friend, blue yellow ghosting fix sound leveling etc. I found two wide horizontal dark bands in the display ,

I though it could be 100Hz ripple and I found a 2v ripple at the "9v" input, using an original PSU giving 11.7v with machine on. It felt strange as that PSU doesn't cause any issues on other Spectrums..
I replaced C131, C45 and C25,, nothing changed.
I tried a switching 9v PSU and saw strange colours in the boot attempt and a hang.. again tat PSU works ok on other machines..
Then I switched to a lab PSU and it took 15v to work properly.. at 0.90A. Banding disappeared but I was sure something was wrong besides the PSU filtering..
5v rail is OK as the linear regulator won't let the ripple trough.
Then I changed C44 thinking in the charge pump DC DC and now it can boot with 10.5v, that's better but not enough.. drop down the input voltage to 10v and colors will get strange and system hang up.. TEA2000 is not being feed well...
I then changed C80.. that was dead, I can only test lower values capacitors but nothing changed.. still needs 10.5v.. I am not sure if to progressing to a full recap may improve something.. or if using a bigger capacitor at the C44 position is a good idea? May TR5 be failing? Its not ease to replace it for me.. and how is the voltage being regulated? I spected a 12v zener somewhere but couldn't find it.

And finally. I lifted C126 to separate audio. That improved something but audio is still messing with composite video output....

[1024MAK]

2V ripple on the 9V nominal output of the original PSU is not good. Best to renew it’s internal electrolytic capacitors.

You don’t say what voltage you get on the +12V (IC33 pin 14), -5V (across zener diode D19) or the -12V on IC33 pin 1.

If you have found faulty or suspect electrolytic capacitors on the board, it would be wise to also replace C50 (22µF) on the +9V (nominal) rail, C34 (22µF) on the +5V rail.

The other important electrolytic capacitors in the DC/DC converter circuit are C74 (4.7µF), C111 (47µF) and C112 (47µF).

The DC/DC converter does not use a zener to provide a reference voltage for the +12V rail. Instead it uses the +5V rail as a reference voltage. The function of C74 is to smooth and filter the feedback voltage from the +12V rail to TR5, which is the ‘error’ amplifier. By error, it is meant that it compares a reference voltage (in this design, the +5V rail is used as the reference voltage) on its emitter with the feedback voltage (via R59 and R58 which are a voltage divider) on its base. If the voltage on the base is more than about 0.6V lower than the voltage on the emitter (about 5V), it turns on more, and hence current is increased to TR4 base circuit, which in turn should cause TR4 to pump more current through the coil, hence raising the voltage on the +12V rail. Do keep in mind that TR4 is an oscillator so the current at its terminals is a switching waveform, not DC.

The current pulses caused by TR4 oscillating create a magnetic field in the coil. Which in turn further drives the base of TR4. The high voltage pulses caused in the coil when TR4 switches off, are what enables the output voltage of the DC/DC converter to be higher than it’s input.

Capacitor C80 has a hard life, as all the power for the +12V output of the DC/DC converter goes through it. It’s provided to stop DC current from the +9V (nominal) rail from leaking through the coil of the DC/DC converter if it is not working correctly.

D15 and D17 rectify the ‘AC’ pluses, and C44 smoothes (filters) the +12V.

As the voltage on the +12V rail increases, eventually there will be an equilibrium, as the feedback voltage on TR5 base increases to a point where TR5 is delivering just the right amount of current to TR4, which in turn holds the current through the coil at a level to keep the +12V steady.

Is TR4 a ZTX650? If it’s been changed to a different type, this may cause poor performance.

Otherwise, generally speaking, it either works well, or doesn’t work at all.

TR5 is not normally a problem, unless TR4 has died. A ZTX213 would normally have been fitted originally. But any PNP signal transistor with a reasonable gain should be okay here (obviously the lead out may be different).

If the voltage on the +12V output is not being correctly regulated, and the +5V is okay on the emitter of TR5, then it could be that TR5 is faulty. But also test R58 and R59 (you may have to lift on lead to get an accurate reading), both should be 1% types.

Mark

[bola_dor]

2V ripple on the 9V nominal output of the original PSU is not good. Best to renew it’s internal electrolytic capacitors.

I did it.. ripple is still there, it has more a serrated shape than a Sinusoidal one, I tested with another toastie and ripple is not present, same PSU. So I guess it comes from inside the machine,

You don’t say what voltage you get on the +12V (IC33 pin 14), -5V (across zener diode D19) or the -12V on IC33 pin 1.

IC33 pin14 is 10.4v (same as psu input)
IC33 pin 1 has 0.8v
-0.6v across D19, I changed it, still -0.6v

If you have found faulty or suspect electrolytic capacitors on the board, it would be wise to also replace C50 (22µF) on the +9V (nominal) rail, C34 (22µF) on the +5V rail.

Will do

The other important electrolytic capacitors in the DC/DC converter circuit are C74 (4.7µF), C111 (47µF) and C112 (47µF).

Will do too

The DC/DC converter does not use a zener to provide a reference voltage for the +12V rail. Instead it uses the +5V rail as a reference voltage. The function of C74 is to smooth and filter the feedback voltage from the +12V rail to TR5, which is the ‘error’ amplifier. By error, it is meant that it compares a reference voltage (in this design, the +5V rail is used as the reference voltage) on its emitter with the feedback voltage (via R59 and R58 which are a voltage divider) on its base. If the voltage on the base is more than about 0.6V lower than the voltage on the emitter (about 5V), it turns on more, and hence current is increased to TR4 base circuit, which in turn should cause TR4 to pump more current through the coil, hence raising the voltage on the +12V rail. Do keep in mind that TR4 is an oscillator so the current at its terminals is a switching waveform, not DC.

The current pulses caused by TR4 oscillating create a magnetic field in the coil. Which in turn further drives the base of TR4. The high voltage pulses caused in the coil when TR4 switches off, are what enables the output voltage of the DC/DC converter to be higher than it’s input.

Capacitor C80 has a hard life, as all the power for the +12V output of the DC/DC converter goes through it. It’s provided to stop DC current from the +9V (nominal) rail from leaking through the coil of the DC/DC converter if it is not working correctly.

D15 and D17 rectify the ‘AC’ pluses, and C44 smoothes (filters) the +12V.

As the voltage on the +12V rail increases, eventually there will be an equilibrium, as the feedback voltage on TR5 base increases to a point where TR5 is delivering just the right amount of current to TR4, which in turn holds the current through the coil at a level to keep the +12V steady.

Thank you, what a tutorial!!

Is TR4 a ZTX650? If it’s been changed to a different type, this may cause poor performance.

Otherwise, generally speaking, it either works well, or doesn’t work at all.

TR5 is not normally a problem, unless TR4 has died. A ZTX213 would normally have been fitted originally. But any PNP signal transistor with a reasonable gain should be okay here (obviously the lead out may be different).

If the voltage on the +12V output is not being correctly regulated, and the +5V is okay on the emitter of TR5, then it could be that TR5 is faulty. But also test R58 and R59 (you may have to lift on lead to get an accurate reading), both should be 1% types.

Both TR4 and TR5 are the original ones, TR5 emiter has 5.2v smooth and steady.
R58 and R59 are both into the specs..
So should I change Tr5? I've used BC557 on 48K boards
C45 has 9.4v and the same kind of ripple of about 1v amplitude.. I guess that's my display banding...

Mark

Thanks a lot..

[Paul]

If your PSU ripple remains after updating the electrolyte Capacitors then change the diodes. There might be one at the edge. Or just replace them all as I did. Not seen often but our Sinclair's are worth the effort.

[bola_dor]

If your PSU ripple remains after updating the electrolyte Capacitors then change the diodes. There might be one at the edge. Or just replace them all as I did. Not seen often but our Sinclair's are worth the effort.

Ripple is not there without a charge, nor with another similar computer. And it's at 100Hz I think it's post rectification.

This is DC coupled
But I'll check diodes

[Paul]

As with rectification the diodes mirror the lower half of your amplitude this doubles the frequency to 100hz if you start at 50hz (60/120).
Anyway, if the PSU loaded with different loads does not show this behaviour it's definitely okay.
Without load I wouldn't expect to see any ripple.

[1024MAK]

Can you compare the current drawn on the faulty unit compared to the working okay unit?

If the faulty unit is drawing a lot more current than normal, this would increase the 100Hz ripple.

The +12V voltage being the same as the +9V input indicates a BIG problem, as this is not supposed to happen with this version of the circuit.

The -5V and the -12V being well outside the required voltages prove that the TR4/TR5/coil circuitry is not operating correctly.

Is there a short circuit between the +9V input and the +12V rail?

Mark

[bola_dor]

Can you compare the current drawn on the faulty unit compared to the working okay unit?

both are getting 0.90A (tested both at 11v )

If the faulty unit is drawing a lot more current than normal, this would increase the 100Hz ripple.

The +12V voltage being the same as the +9V input indicates a BIG problem, as this is not supposed to happen with this version of the circuit.

The -5V and the -12V being well outside the required voltages prove that the TR4/TR5/coil circuitry is not operating correctly.

Is there a short circuit between the +9V input and the +12V rail?

I mesured 1ohm between the 9v entrance pin and pin 14 at ic33 and 12 ohm from the pin to C45 (+)lead. (R62+L6?). 1.2ohm to C80(+) and several Kohm to its(-) so the leaking is before C80.
And I found it .. someone was here before

IC33 pin 14 is fed up directly from the PSU by that red wire...
If I desolder it I don't know what would I find..
I

Mark

Again, thanks a lot

[bola_dor]

Well I replaced both TR4 and TR5 with BC635 and BC557 (I double checked pinouts and had to cross Tr4 leads) I've used successfully this combination on 48K issue 6A boards.. lifted the bypass wire. And lifted r62 just in case...
Didn't work at all, 12v is about 0.6v...
I am starting to think about a modern DC DC booster.
I have a XL6009E1 arround.. how much current will it drain along the TEA2000, if I feed it from the 5v rail? It has a common 7805 linear regulator. Or is it advisable to use the PSU input itself ro feed the DC DC?. Thing is that if someone uses a psu giving more than 12v it may destoy the DC DC circuit. Perhaps I should use a 11v zener in its entrance.
I know this is an unorthodox solution and that serial ports nor keypad will work again.. but its on my friend expectations...

[1024MAK]

I can’t remember how power hungry the TEA2000 is. But it’s not going to be too much for a boost converter run off the +5V rail.

The alternative arrangement would be to link the +9V and +12V rails again (and remove any of the existing components that would not like this) and use a 12V regulated PSU. The downside being (1) the 7805 and heatsink would get hotter and (2) it will only work correctly with a 12V regulated PSU.

Mark