Apple Lisa Datapower PSU with bad +33v

[pentium]

For reference, the PSU schematic is here and here.

When I got my Apple Lisa originally the system seemed okay aside from a bad analog board. While I was troubleshooting that I had C7 explode. I ended up replacing C1, C3-C7 which were all originally paper with polyesters. C3 and C4 were replaced with 600v rated caps but were otherwise safe substitutes.

After that I noticed I wasn't getting any high voltage at all. I checked and +5 and +12 and they were in spec. I then checked the +33v line and it was reading +70v. Oh dear. So that massive voltage boost blew up a regulator on the analog board and that is why I have no high voltage.
Anyways, I checked components R23, CR21, CR22, C22 and C25 out of circuit to see which one was bad. Diodes were good, capacitors didn't seem shorted and were giving their capacitance rating and the resistor was in spec.
Best I can think is the worst and that would be the 33v winding has shorted out but would that one exploding cap really cause damage this far down the line? Why only to +33v?
The PSU can only easily be tested under load. Without the system card cage installed there isn't anything to tell the PSU to turn on.

[budm]

Re: Apple Lisa Datapower PSU with bad +33v

C1, C3-C7 are AC line filter to reduce the noise from power line getting into the power supply and also reduce the noise generated by the switching power supply from injecting the noise back into the power line. The unit will still operate if all those caps are removed. The caps probably not safety caps X/Y type, if they are the cap would have fail open instead of shorting out and blew up, you should should use X/Y type such as Panasonic ECQUL series. They will not cause the 33V to go up to 70 volt.
But if you look at the circuits, the 33v, 5v, +/-12v are generated by the same transformer, so if the 33V went up, then the 5v and 12 would have gone up also since they are driven by the same transformer drive circuits. C25 for the 1000uF50v for the 33V would have blown up if 70v was applied. If the winding shorted out, it would have taken out the transistors that drive the transformer.
So did C25 also blow up?

[kaboom]

Re: Apple Lisa Datapower PSU with bad +33v

C1, C3-C7 are AC line filter to reduce the noise from power line getting into the power supply and also reduce the noise generated by the switching power supply from injecting the noise back into the power line. The unit will still operate if all those caps are removed.

Correct.

C25 for the 1000uF50v for the 33V would have blown up if 70v was applied.

Unless the cap was dried out beyond sanity.

Pentium, if you have an x100 probe, can you scope the 33V output? I'm also curious as to what +/- 5/12 look like on the scope.

If C25 is completely open, depending on your meter, there is a chance that the high frequency pulsating DC is giving you a false reading.

[pentium]

Re: Apple Lisa Datapower PSU with bad +33v

I unfortunately only have X1 and X10 probes. I can't be 100% on C25 but I guess tomorrow I could run to the shop and get a replacement cap.

[budm]

Re: Apple Lisa Datapower PSU with bad +33v

His 10x PROBE should be just fine, but I would just replace C25 any way, caps in this power supply are REALLY OLD, it should be recapped anyway.

[kaboom]

Re: Apple Lisa Datapower PSU with bad +33v

Whoops.

Must've had a moment when I suggested the x100.


A 10x probe is fine.

[pentium]

Re: Apple Lisa Datapower PSU with bad +33v

Replaced the 50v cap and the voltage dropped down to 68v.
You sure I should be directly poking a this with my scope? I use it a lot for TTL and CMOS stuff but 70v is far beyond what I ever used it with.

[budm]

Re: Apple Lisa Datapower PSU with bad +33v

Can we see some pictures? I realy like to see what you are seeing, unless you may have bad ground some where to give you this strange reading without the new cap blowing up.

[Th3_uN1Qu3]

Re: Apple Lisa Datapower PSU with bad +33v

You sure I should be directly poking a this with my scope?

If you tell us which scope you have. An analog will be fine, if it's one of those USB thingies, i wouldn't be too sure. Read the manual.

[pentium]

Re: Apple Lisa Datapower PSU with bad +33v

It's a Tektronix 335. It's a fancy little thing that I can even plug into the car.

As for photos, here's the general area of the circuit. I marked the components.

[kaboom]

Re: Apple Lisa Datapower PSU with bad +33v

I wonder just how much current is actually behind this "70V."

Put a 10K 1w resistor across that output and check it again.

If that new cap is taking 68 volts, it should warm up a little, at least while it tries to form up to that voltage.

Looks like that TO-3 device on the black heatsink, behind the 3 inductors, is the +5 rectifier...

[pentium]

Re: Apple Lisa Datapower PSU with bad +33v

Does it matter if I use a 5w 10K I have handy instead?

[Th3_uN1Qu3]

Re: Apple Lisa Datapower PSU with bad +33v

That's even better.

[aurel6814]

Re: Apple Lisa Datapower PSU with bad +33v

Hi, I'm new here and just have a question.
I'm also trying to repair a Lisa power supply.
How did you manage to test the power supply outside the computer?
Do you apply a tension on the X (on/off) contactor?
Thanks! Aurélien

[pentium]

Re: Apple Lisa Datapower PSU with bad +33v

Put the 10K resistor in as a dummy load. Voltage now sits at 60v.

I'm also not testing the voltages out of the system. It's a lot of taking it out, tearing it apart to replace or inspect something, putting it back together and reinstalling it in the lisa before poking my DMM at the connector for the analog board. I'm sure it would be easier if I knew what the Soft Power circuit expected for an "ON" signal and it would sure as hell be easier then lugging this $#*!ing lisa on my bench.

[budm]

Re: Apple Lisa Datapower PSU with bad +33v

The ON/OFF control signal should be >VR (2.52V, which feeds PIN5 of the ICZ3B-LM393 to keep the output High, when you apply 3VDC to pin6 (ON/OFF) of ICZ3B, the output will swing low and turn on the power supply.
I am still not convinced that the output is 60V (you may be getting voltage leakage reading), I would use 500 Ohms resistor as a load, at 70V with 5000 Ohms, it will be drawing about 14mA, about 1 Watt.

[budm]

Re: Apple Lisa Datapower PSU with bad +33v

I just read the note on the diagrams, it shows it is reversed engineered, so you must verify the connections to make sure they match your board.

[pentium]

Re: Apple Lisa Datapower PSU with bad +33v

Double checked the schematic with the actual board and there are no deviations.

I don't have a 5K 1W resistor on hand or anything close. For me to try that it will take me a few days.

[budm]

Re: Apple Lisa Datapower PSU with bad +33v

At least now you can turn it on manually using that ON/OFF pin.

[pentium]

Re: Apple Lisa Datapower PSU with bad +33v

Added one 5K 1W resistor and the voltage remains at 58v. I also have replaced the supposedly blown 7824 but prior to installation I compared the two and they both gave identical readings. I will not however reinsert the analog card in the event that it just goes and blows up the replacement as well.