Suspect Bad PS Apple AcBel API 1PC36

[buz]

Re: Suspect Bad PS Apple AcBel API 1PC36

Ok - think I can decipher this .... you don't assume too much .... I just don't know enough. I'll try and get this tonight, hopefully post before calling it a day.

Regarding the 'zero' readings ... the 4 pins are loose, they wiggle easily. It seems there is tension on the leads springing them to the edge. So when I take the readings I try and center the lead, pushing it slightly to center it in the hole.

[buz]

Re: Suspect Bad PS Apple AcBel API 1PC36

Think I got it ....

2 & 3 = 244
2 & 1 = 246
5 & 6 = 242
5 & 4 = 240

Reversed were all infinite.

[Toasty]

Re: Suspect Bad PS Apple AcBel API 1PC36

Very good. That means it is not the diode packs. dang. More digging on my end....

You can solder them all back in place.

[Toasty]

Re: Suspect Bad PS Apple AcBel API 1PC36

Again - Warning
You are going to measure the "HOT" side of the PSU power.


Please check the voltage at the points indicated in the picture.
1 - with the unit -just- plugged in
2 - with it jumpered -only- at the mobo plug
3 - with it jumpered at the mobo plug AND the "bypass" on the opto

Toast

[buz]

Re: Suspect Bad PS Apple AcBel API 1PC36

Again - Warning .....
AND the "bypass" on the opto

Toast

I understand except for this - the bypass on the opto - is that the wire I soldered to the M8 pin and then to ground on the mobo plug?

[Toasty]

Re: Suspect Bad PS Apple AcBel API 1PC36

Yes

[buz]

Re: Suspect Bad PS Apple AcBel API 1PC36

Well, I said I understood and now am not so sure.

1. 'Just' Plugged in = 387
2. (no power) Jumpered Only = started around 300+ then kept trickling down
3. (no power) Jumped and ByPassed = same as 2 but started at around 200 then trickled.

[Toasty]

Re: Suspect Bad PS Apple AcBel API 1PC36

Somethings wrong. Go over your setup. You should have something on the 3rd one near the first reading, as you are "forcing" the supply on.

Remember that 390v area voltage stays high for at least 20 minutes.

[buz]

Re: Suspect Bad PS Apple AcBel API 1PC36

Ok - my question at this point - do I keep it powered on (plugged in) while I do steps 2 and 3? Thinking now I am supposed to

[buz]

Re: Suspect Bad PS Apple AcBel API 1PC36

Plugged in for all 3 tests - all 3 tests = 387

[Toasty]

Re: Suspect Bad PS Apple AcBel API 1PC36

Ok. You did solder the diode packs back in, correct?

Force it on again and measure right off the board, from the bundles of wires, each of the voltages again. Negative is the black bundle in the middle of them.

3.3 (orange)
5v (red)
12v (yellow)

[buz]

Re: Suspect Bad PS Apple AcBel API 1PC36

Yes, the diode packs were soldered back (post 83).

Jumpered green to ground on plug, and used the resistors from M8 to ground on plug (post 60)

Orange 3.33
Red 4.14
Yellow 8.96

This is from on the board where the harness comes 'from'.

[buz]

Re: Suspect Bad PS Apple AcBel API 1PC36

Hey Toast - first want to thank you for all the effort you have put in to helping me try to repair my PSU. I know my 'novice' abilities have been a challenge for you.

Have we run out of ideas or perhaps come to a point where my ability isn't enough to proceed? It isn't a must that this gets repaired. It certainly has been an interesting exercise for me so thanks for that too. There's some nostalgia with this 'old' computer and had hoped to keep it going. I know there are other avenues to do that although at this time finances aren't going to permit them. (i.e. get another psu or computer)

Thanks again.

[Toasty]

Re: Suspect Bad PS Apple AcBel API 1PC36

I am just trying to figure why this is being dragged down on the 5v & 12v lines. Once the power is generated for the primary side of the transformer and is created on the secondaries, it is simply rectified thru the diode packs then filtered with the output coils and caps. There is no regulation or any other devices it passes thru. Especially with the 12v as it comes from the diode packs, hits the large cap and is output. There is no other component to filtering the 12v line.

Since this is happening with the supply un-loaded, and an overheating problem was encountered, I am puzzled that nothing indicates damage to the unit, yet there obviously was.

I have a suspicion that the transformer may have been compromised (internally shorted winding) prior to your recap. Once you got it working, it then failed after being loaded for the short time it was working. Yet, the voltages look perfect on the input and it doesn't seem that the driver MOSFET was damaged either.

I am very leery of having you work on the primary side. It is a safety concern of mine with anyone as I don't want to see anyone injured or to have a problem with the unit (fire) if indeed we did get it working.

At this point, I would like to have this unit on my bench to test. I feel that with 2 units side-by-side I should be able to find the problem. If you are good with that idea, then we can arrange to have the unit sent here for me to look at. If you want to save on postage, there would be no need to send the case, just the board. I feel that you could reinstall it properly in the case. But if you would feel more comfortable having me do it, then please do. Feel free to PM me.

Toast

[buz]

Re: Suspect Bad PS Apple AcBel API 1PC36

PM'ing

[Pyr0Beast]

Re: Suspect Bad PS Apple AcBel API 1PC36

Leaky shottky diode (it shorts when temp gets high)
Bad cap.
Bad optocoupler with compensation and feedback circuit.

Paint stripper fan and dust-off spray would probably be the next thing to try with.

[buz]

Re: Suspect Bad PS Apple AcBel API 1PC36

Victory!!!

The PSU is doing it's job .... starting/running the computer. 97 posts later and an end run. Your generosity Toast is hard to quantify. Thank you.

I had a 'heart stopper' though. Had it fired up, connected to the Internet, watching the Temp Monitor, listening - all seemed toooo good. Of course it wasn't under any load but still considering these famously noisy MDDs this one is ultra quiet now. So quiet I had to pop the case open to make sure the fans were running. Had the case door about half opened and *BLINK* it turned off. Can't list the expletives I had for the computer/situation. Unplugged it and went for a short drive. Came back, plugged it in and it fired right up and as quiet as before.

The carpenter guy here says 'ah, gotta be a short somewhere' the wannabe electronics guy says 'hmmm'.

Bottom line at this point, I'll appreciate it as long as it runs and I just won't open it up while it's on (something I do all the time). Still not 'sure' the fans are running, CPU is only at 55ºC atm, gradual rise since boot up. Leveled off at 58.6ºC Guess the fans are running

Thank you again Toast.

[Pyr0Beast]

Re: Suspect Bad PS Apple AcBel API 1PC36

What was the fault ?

[Toasty]

Re: Suspect Bad PS Apple AcBel API 1PC36

Check all the connections to drives, leds, motherboard, etc. for a tight fit. The unit is -very- sensitive to power interruptions. Senses an imbalance in the load. Actually did it to me when I moved a ground wire while under load. Picked up a bad connection and -blip- off. Quick off/on restart and was fine.

@Pyrobeast - Another small cap had failed. Read 82Ω where it should have been <0.8Ω (800mΩ). Everything else about the cap was correct. Value, leakage.

[buz]

Re: Suspect Bad PS Apple AcBel API 1PC36

Ok, thanks (some more) - I'll check those connections, fairly certain things are tight but you never know. There does seem to be something else at play but I don't know enough to figure it out myself. I'll copy and paste some more detail from an Apple Discussion thread ....

https://discussions.apple.com/thread...rt=15&tstart=0

Guess I'm not done with this thread. In the mean time here, the PSU failed and it's repaired now. I am still wrestling with the MDD. The good news is that it starts now and has a reasonable start up time. I didn't want to stress it right away so was only running Temperature Monitor and Safari - and the assorted MDD mods Chud, and NapEnable. As it idled along I was composing a reply to another thread and out of the blue the fans ramped up the temps began rising. I was concerned maybe all the fans weren't working so opened the case and *Blink* it shut down.

Didn't know what to think so let it set a while and it booted right back up, but with the fans ramped. My first effort was to try and boot to Disk Warrior and had various unexpected behaviors from spitting the disk, to shutting down. Finally got booted to DW and just as it was ready to build a new directory it shut down again. Gave up on DW, I was able to get a graph from DW that showed >1% so figured it was ok. Then went to ASD and ran it, fortunately I was watching most of the way through and was passing all the tests, but right at the end the machine shut down again. Haven't restarted it yet so don't know if I got a report or not.

These shut downs aren't complete, the Power button still glows (pulsing) and the red LED on the board is still on. The display goes off, the fans quit, the drive quits but the power stays on. Pressing the Power in this state will power it back up.

The machine is 'stock' and am using only 1 ram chip.

Back to the beginning, when it started to ramp up initially, it had idled along for 20 or 30 mins just as quiet and peaceful as could be - then wham!! it took off and with no apparent extra load. At one point I looked at Activity Monitor and saw nothing unexpected.

Any ideas for me??