IIsi Sony PSU trickle current startup

[RickNel]

Been having intermittent startup problems with my IIsi. That thing when the relay snaps off as soon as it comes on.

When it behaves, the whole system starts up with no symptoms. So I don't think there is any problem with the main circuits in the PSU or on the logic board. It all looks clean and logic board has just been recapped with tantalums, the 5v and 12v rail filters with low-ESR electrolytics.

I put an oscilloscope on the PSU outputs, with the PSU removed from chassis. The trickle current that is supposed to power the startup is on pin 10. On AC power-up, this pin takes about a second to rise from 0 to just under +5volts, the relay snaps on, then off, and the trickle returns to 0. End of story till next AC on.

The result is the same if I bridge pin 10 to ground, which some threads suggest should force the PSU to start up. I'm not so sure about that, because the case and keyboard power interrupt switches are both "normal on - instant off", which suggests that the power-on circuit is triggered by a ramp up rather than a steady on like the ATX "power good". Please correct me on this if you know more.

The PSU pins that are supposed to be GROUND are carrying about +.3volts as long as AC is on. Is that normal or does it suggest a possible diode leak?

Inside the PSU is a small PCB with one chip and a bunch of other components, including two electrolytic caps that look like the 220uf types I have just replaced on the logic board. Looks like that PCB may be what ramps up the trickle current. I'm suspicious of those caps.

Is there a schematic around for the Sony PSU? This is a 240vac model AP56 , part number 699-0567.

Any other thoughts on cause/cure?

Rick

 

[uniserver]

Hey sir looks like this some how posted twice, might want to delete the other one before too much time passes.

One of the members found a capacitor in the Macintosh IIsi PSU, that seems to be a pattern failure. As I remember it was only one.

I will dig around and see if in can find that thread.

 

[Trash80toHP_Mini]

The double post appears to have disappeared.

 

[RickNel]

Thanks for the idea (and the heads-up about the double post - I've asked admin to delete).

If you mean this thread, I've looked hard but no evidence of a burst cap like that one. I think the issue is more likely in the power-up logic.

Rick

 

[uniserver]

Have you checked that cap's ESR? I tell you what I can't tell you how many times the big 470uf or 220uf (with the SE/30/IIsi/IIcx/IIci) axial caps looked great, and in-fact they were bad and causing annoying issues.

With the LC psu's that I fix, some of the pattern failure caps have not even started leaking, but they are still bad.

 

[RickNel]

So, you're telling me I need to take the PSU out of its case and check ESR on the major caps?

Well, I did ask for advice and suggestions.

For accurate ESR reading wouldn't I have to desolder the caps? Or can I get away with just putting a multimeter across them and watching the resistance climb and fall?

BTW my Sony PSU is similar, but not identical to the one pictured in the other post. Some of the caps in that one are placed differently, and the ramp-up circuit seems to use discrete transistors rather than an IC, so not sure I can use the pictures for exact reference.

In testing since my original post, I have had the unit boot up a few times with two different logic boards. Sometimes it will cold boot properly using the reset button - ie. not immediately on AC power. Other times it boots immediately on AC power. It will restart properly from MacOs software instruction, but does not recognise the keyboard or cabinet back button hardware resets. To me, that still points to the the soft power system and mostly likely that part of it in the PSU, unless both of my logic boards have the same fault in them. It could relate to the availability of current in particular caps at particular times in the boot sequence, so I will try to check the PSU caps. The Sony PSU seems to have about 15 of different values

Rick

 

[uniserver]

You could see if maybe a air can upside down( COOL ) or Hair dryer (hot) makes a change weather or not it functions.

 

[RickNel]

Well I got deeper into this. Without a proper cap tester, I tried continuity testing across the terminals of all caps in the PSU. No shorts, most showing a ramp up to charge, discharge, and residual resistance.

Biblits earlier thread about a cap fix relates to a Plessey PSU I think, but the failed component was a high-voltage cap, so I decided de-solder the only large electrolytic on the primary side of my Sony PSU - 4.7uf x 400volt. It still seemed to test OK off the board, so I soldered it back in. There was also a possible dry joint on one leg of a smaller-value tantalum next to it, so I reflowed that.

Test result was a big bang, fluid squirt and impressive flow of grey smoke - while the 11si continued to boot (until I turned it off). So that HV cap was probably in the process of failing and the heat from the de-solder/solder probably finished it off. That can sits about 1mm from a large HV MOSFET and its heatsink, so has probably been cooking slowly over the life of the unit.

Now thinking I should probably also replace the other caps on the primary side - 2 x 220uf @ 250v, plus a couple of smaller filters. I'd rather not swap out the 9-10 electrolytics on the secondary low-voltage side if this is not necessary - not just from laziness, but not wanting to put more heat stress on the board than necessary.

Anyone done a major recap on a PSU like this?

Rick

 

[RickNel]

Well I might as well keep chatting to my self on this.

I got a full set of electrolytics for my Sony PSU - just to be sure.

Note this is a different Sony PSU to the Plessey one in Bilbit's thread - maybe because this is the "international" 250v model? Or different factory?

There are only 3 high-voltage electrolytic caps in this PSU -

2x 220uf/250v at C101,C102

1x 4.7uf/400v at C105

C105 was the one that blew, and it connects to the power relay, so I replaced it first and then tested. Bootup OK, so PSU now functional with no problem except still no onboard or keyboard soft power function.

Next thing I'll try will be the small electros on the PCB daughterboard that houses what I think is the counter chip for the PSU. I'll leave that till I get a better desoldering iron that is on its way. PSU traces sink a lot of head and make desoldering a pain. I won't do the full re-cap unless I need to, but I have the caps ready just in case.

Rick

 

[RickNel]

Since the previous post, I found that changing the HV caps HAD NOT after all eliminated the startup problems. So, as foreshadowed, I next changed the caps on the small daughterboard.

The board is about the size of a large postage stamp and stands vertically off the main PCB, connected by 10 through-hole soldered pins. To tackle this, I waited till I had received a better desoldering iron I had ordered - one with the heated hollow tip that goes over the pin to be desoldered, with a hand-compressed piston that is released for a one-shot solder suck. It can be used one-handed, which is good when needing to hold something steady. This iron worked perfectly, clearing each pin and hole of solder on first shot.

The two SMD electrolytic caps are the same type as the ones we generally replace on the logic boards. I snipped them off, then thought about taking a photograph. Sorry the image is blurred - I didn't notice until too late after I had soldered it back into place.

The chip is a dedicated switch-mode power controller, managing the Pulse-Width Modulation and other timing functions for the switching circuits that determine voltages delivered by the PSU under various conditions, including the startup trickle current. The board also has a tiny trimpot, which I didn't touch. Much of the surface of this board was blackened and a bit sticky - I could not clean it fully with alcohol, acetone, or detergent, and did not want to scrape off the insulating lacquer, so left if looking fairly smutty once it was clean enough to check carefully for any broken or burnt traces. I found none.

Looking from the component side, pins down, the cap on the left is 100uf x 6.3volt, negative side down. The cap on the right is 33uf x 25volt, negative to right. Unlike the main PCB, the sub-board carries no markings to identify locations or show the polarity of caps, so it is important to NOTE the polarity of the caps before removing them.

I cleaned up the pads with a bit of flux, then used spots of solder paste on the new components, holding them in place with tweezers while flowing with a sharp pointed iron.

With these two caps replaced, the PSU behaved perfectly - that is to say it did not start immediately on power-up, and it responded correctly to the soft-power switches on the back of the IIsi and on the keyboard.

I have a full set of replacements for all remaining electrolytics, but will hold them in reserve until there is failure.

For reference, the full list of eletrolytic caps on the main PCB is:

C101,C102 220uf x 250v (primary)

C105 4.7uf x 400v (replaced) (primary)

C202,C203 2200uf x 6.3v (secondary)

C211,212 1200uf x 16v (secondary)

C302 56uf x 35v (primary)

C401,C402 100uf x 25v (secondary)

Rick

 

[Trash80toHP_Mini]

Keep up the chatting, this is an interesting exercise! Let us know of any changes, several folks have had problems with IIsi PSUs lately. Mine released the magic smoke some time ago, but it's going to get an ATX transplant at some point. :approve:

 

[uniserver]

i think this belongs in the caps replacement thread, however i do not moderate that, I will be personally bookmarking this useful thread. Thanks for the good work.

 

[Trash80toHP_Mini]

Thread linked: Capacitor Replacement Thread . . .

I haven't updated in two months, please PM me regarding any other new caps threads, uni.

 

[tecneeq]

I have a semi-failing IIsi PSU as well and monitor the thread with my steel blue hawkeyes¹!

¹) One lazy, one milky.

 

[RickNel]

Trash80toHP_Mini » 26 Feb 2013, 01:05 Mine released the magic smoke some time ago, but it's going to get an ATX transplant at some point.

I'm curious to know which, if any, caps in the Sony PSU have blown smoke other than the C105 4.5uf x400v.

I have a hypothesis that overheating may be a problem because the air inlet vent above the PSU heatsink is very easy to cover up in normal desktop use..

Rick

 

[RickNel]

I've now completed the replacement of all electrolytics in my IIsi PSU.

Hopefully, this will have corrected the lingering system problem I have had, where the keyboard would start to randomly repeat keys after the IIsi had been connected to power for several hours. Wierdly, this fault would start whether or not the IIsi was powered up all that time. It would self-correct when the power was fully disconnected for a few hours. So the issue, I am guessing, was in a buildup of ESR or loss of capacitance in caps feeding the 5v rail, eventually producing some ripple that disrupted the ADB hub.

As I changed all the caps I found considerable sticky evidence of electrolyte leakage under the caps, especially at the 1200uf and 2200uf secondaries. No corrosion, though.

When I removed the cakes of hotmelt glue that had been slopped around at original assembly, I found a couple of corrections to the table of cap values posted earlier. Here is the corrected table. It does not cover the several tantalum and ceramic caps that I have not seen necessary to replace (so far!)

C101,C102 220uf x 250v

C105 4.7uf x 400v

C106 100uf x 10v

C202,C203 2200uf x 6.3v

C211,212 1200uf x 16v

C302 56uf x 35v

C401 100uf x 25v

C402 47uf x 25v

After seeing how many problems an unreliable PSU can cause, I strongly recommend PSU re-capping as much as logic-board recapping. Don't wait for the smoke.

Rick

 

[CC_333]

Excellent work!!!

I have a IIsi PSU which seems to have died, and I noticed a bunch of corrosion on that same vertical board (and a few other places :-/ ), so that could be causing my supply's "deadness".

With your wonderful list as a reference, I will see about recapping it. I have little to lose at this point.

c

 

[uniserver]

oh yeah , fix your psu CC_333 and post here and let us know how it worked out.

 

[Trash80toHP_Mini]

Trash80toHP_Mini » 26 Feb 2013, 01:05 Mine released the magic smoke some time ago, but it's going to get an ATX transplant at some point.

I'm curious to know which, if any, caps in the Sony PSU have blown smoke other than the C105 4.5uf x400v.

I have a hypothesis that overheating may be a problem because the air inlet vent above the PSU heatsink is very easy to cover up in normal desktop use..

If I ever find my other IIsi case again (it looks like a well stocked vintage Mac store exploded in my apartment, ATM) I'll have a look see at the PSU's guts.

I'm sure you're right about the overheating issue, any tilt/swivel base monitor would partially, if not completely, block those top vents. The 12" 'zaToppers and the Portrait have feets high enough to allow for proper airflow to the PSU. Those were the only adequate resolutions supported by the IIsi's VampireVideo setup anyway, but few would opt for the expensive grayscale Portrait to get an adequate pixelcount . . .

:O THX, after looking all over the joint . . . thrice . . . I remembered having swapped in the IIsi yesterday, when I took the DuoDock out from under the Portrait! :lol:

. . . my IIsi PSU appears to have suffered from what appears to be major cap leakage (brown discoloration on the PCB) centered on the tall cap tank farm in the corner.

I'm prepping a couple of other packages to deliver into the care of the USPS, so PM me you address and I'll sent it along for the research project. I don't need it back, I'm halfway through doing an ATX conversion inside the sheet metal enclosure.

Just checked out the wiring harness, which I need to desolder for the ATX hack . . .

. . . leakage seems to have made it all the way down into the Mobo connector along a few wires that had been in contact with the PCB! 8-o

I'll post some piccies when I get a chance, there also appear to be burn spots on the bottom of the PCB, likely from overheating if that's what they really are.