Troubleshooting a IIsi Sony APS-06 Power supply - blown capacitor caused by bad diodes in series with C105

[joshc]

This post is a recreation of a previous troubleshooting thread (lost in the recent forum crash) where I had some problems identifying the cause of a blown capacitor in a recapped Sony APS-06 from a Macintosh IIsi.

The exact PSU in question:



I used the correct value capacitor and it was placed in the correct orientation, this is after removal, C105:



C105 is in series with D105 ...

I removed D105, and low and behold, bad diode:



Hiding underneath a heatsink, there's another diode which is in the same series - D106. This was also bad.



After some digging around and help from other members, we managed to find the schematics for this PSU, which are included with the IIsi schematics from Bomarc:

Macintosh 68k Schematics : Bomarc Services : Free Download, Borrow, and Streaming : Internet Archive

Macintosh 68k Schematics========================Bomarc Services reverse engineered several Apple Macintosh computers in the1990s and sold the schematics to...

archive.org

This was the relevant bit of the schematics:


Unhelpfuly, 22UA did not really identify exactly what kind of diode these were - I couldn't find anything online, and it seemed like these were old parts no longer made with that specific part number.

But it was really @techknight who was able to make the most sense of what was going on, I've pulled what he said out of the Google cache from the lost thread:

ok so heres the deal. I looked at the schematic (never looked at a IIsi power supply in this depth before)

it appears the UA23s are your main DC Rectifier diodes, and TH101 is your surge suppression resistor. Make sure this resistor is not open. it may need replaced.

The recommended design for that STR switcher regulator is an RM1C which is EOL, so following those characteristics, You can probably replace those two UA23s with 1N4007s or something along those lines.

At his suggestion, I replaced the two bad diodes with 1N4007s or similar - what I actually found to be closest to the spec needed was:

Vishay 400V 1A, Diode, 2-Pin DO-204AL 1N4004-E3/54

so I ordered some of those, put them in, and low and behold - it worked!

I hope this thread proves useful for someone else diagnosing/troubleshooting this particular PSU. When I had the problem, I couldn't find anyone online who had the same issue. Most problems with the IIsi PSU relate to either just the radial capacitors, the SMD caps on the soft power daughter board, or a problem with TH101.

 

[belzrebuth]

Can I test this power supply without attaching it to the motherboard?
I tried to turn it on by shorting trickle voltage to ground but it didn't turn on..
I hope I didn't manage to break something.

 

[joshc]

I don't know the answer for definite for this one, but like a lot of computer power supplies, it probably needs some load to start.

 

[cheesestraws]

I tried to turn it on by shorting trickle voltage to ground but it didn't turn on..

Also, shorting trickle to ground won't turn it on anyway . The logic for most Mac PSUs is the opposite to ATX; for ATX power supplies you need to ground the soft power, but for most Mac power supplies (including, I think the IIsi), you need to take the soft power pin high, to the standby/trickle 5 V rail. You might want to try that.

 

[re4mat]

Can I test this power supply without attaching it to the motherboard?

Yes, but you need to give it a load. And it's been awhile, but I think to turn it on in that scenario you have to short pin 9 to pin 10. (Hopefully I'm remembering that right.)

I tried to turn it on by shorting trickle voltage to ground but it didn't turn on..
I hope I didn't manage to break something.

That sounds normal. According to the AGMFH, shorting a power out to ground will cause the PSU to shut off.

 

[belzrebuth]

I do have an electronic load so I could connect that to the 5V rail which I assume would be the reference for the rest of the voltages produced.
So 9 to 10 it is.
I will test and report back.

 

[captjon320]

I hope this thread proves useful for someone else diagnosing/troubleshooting this particular PSU. When I had the problem, I couldn't find anyone online who had the same issue. Most problems with the IIsi PSU relate to either just the radial capacitors, the SMD caps on the soft power daughter board, or a problem with TH101.

I've been following this thread along with others on the IIsi PSU. I have four IIsi PSUs that I've been working on over the last month. Three are Sony models and one is a MagneTek. I have recapped them all (along with the daughter boards) and have one Sony working and the MagneTek working.

The two Sony PSU that are not working are both DOA when plugged in. I understand it has to be a dead component on the PSU, but don't know where to start testing. I would rather test than throw components at them blindly. Both fuses are in tact in both PSUs. Should I start with TH101? This is the first thread that I've come across referencing this component. Any help or suggestions to move forward are appreciated.

 

[joshc]

Should I start with TH101?

Do you have a component tester? You should be able to start testing for bad resistors / diodes. There's also a thermistor on these that can be bad, that's TH101.

 

[captjon320]

Do you have a component tester?

I actually have one coming tomorrow, so I am planning on learning how to use it quickly! I will start with resistors/diodes and move on from there.

 

[dougg3]

Adding more info to this thread about the IIsi power supply and the unidentified 22UA diode:

My IIsi power supply was dead. Of course I recapped and heavily cleaned it before even trying to use it, but that didn't fix it. The symptom was the 5V standby rail was fine, but as soon as I pressed the power button or power key, it would just make a faint ticking sound every couple of seconds like it was failing to start.

I ended up tearing it apart and finding a shorted diode, also a 22UA: D103. With enough research/asking around, I figured out that this diode is used in what is called an "RCD snubber" circuit for protecting the MOSFET (Q101) from voltage spikes. The recommendation I found for a diode used in this circuit was to make sure the reverse voltage rating is at least as high as the MOSFET is rated for, be able to handle 1 or 2 amps, and also make sure it has a very low reverse recovery time.

The 2SK1402A has a reverse breakdown voltage of 650V. One person recommended a Vishay MPG06K (800V), but it doesn't have fast recovery time. There's a variant of it called the RMPG06K that has a faster recovery time. I suspect it probably would have worked fine.

Instead, I located another RCD snubber circuit in a random other power supply board I had sitting around, and stole the diode from it: a BYV26EGP. It's rated for 1000V and has a very fast recovery time. It's bigger than the original but it was still easy to fit. Bottom line: replacing it fixed the problem and the power supply works! (It doesn't ever shut off, but that's a different problem for another day/thread).

BTW, I found a mistake in the Bomarc schematics relating to this part of the circuit and confirmed it on the actual board. They mistakenly showed the snubber going to the "primary common circuit" but it doesn't. It goes to the same voltage that goes to pin 4 of the transformer. This makes way more sense because it matches all the example RCD snubber circuits you can find online too. Here's the original:



And here's the corrected version showing where R103 and C103 actually go:

 

[dougg3]

The 2SK1402A has a reverse breakdown voltage of 650V

I don't know what I was thinking when I wrote this, but I was trying to say the 2SK1402A is rated for a maximum of 650V so that's the spec I was shooting for on the diode. My bad on the wording there.