Suspect Bad PS Apple AcBel API 1PC36

[buz]

Re: Suspect Bad PS Apple AcBel API 1PC36

Ok ... tell me how to safely jump off a cliff and I'll probably do it ...

This may take me a bit (fitting it in around some other 'duties') but think I understand the 90%.

Soldering to Pin2 would be next to the square solder pad. (?)

Found a piece of 24ga wire should be good (better than the 16ga next choice)

Need to ask (because I've wondered) - I did the cap soldering with a chisel tip and wonder if I should have used a needle tip. The chisel tip seems to melt the board solder 'almost' before it touches (fast at any rate)?

Measuring the 3.3/5/12 would be as done when checking voltages initially (using a pinout diagram and instructions)?

[buz]

Re: Suspect Bad PS Apple AcBel API 1PC36

Reporting:

5v test(s) ... pins 1 and 3 ... 4.7v and 4.1v

12v test(s) ... pins 10 and 24 ... both 8.8v

3.3v test ... pin 6 ... 3.3v (pins 17, 18 and 19 also tested 3.3v)



EDIT: Daughter board 5th pin tested ~2.4v occasional drop to ~1.5v

[buz]

Re: Suspect Bad PS Apple AcBel API 1PC36

oops .... 5th pin results above was jumpered

unjumpered 5th pin tested 4.4v

[Toasty]

Re: Suspect Bad PS Apple AcBel API 1PC36

>>Soldering to Pin2 would be next to the square solder pad. (?)<<

Pin 1 on *most* IC's and such devices is *usually* indicated by a square solder pad.
It is also *usually* indicated by a dot or dimple on the device itself.
(see pic 1)

Pin 2 of M8 is the round pad next to Pin 1 (the square pad on M8). You just tested it yesterday. (see pic 2)

The 2.4v on DB pin 5 has me concerned as it may not be low enough to fully turn on the opto. The 1.5v is the IC trying to start up and check the voltages.

I need you to solder the 2 resistors in that package together in parallel. (see pic 3)
This makes the resistance 1/2 of 1kΩ or 500Ω. That should provide a low enough resistance path to enable the opto fully.

Rerun the prior test and repost the voltages.

Toast

[buz]

Re: Suspect Bad PS Apple AcBel API 1PC36

Pin 5 jumpered ... 1.0v - 1.6v

unjumpered .... 5.56v

[Toasty]

Re: Suspect Bad PS Apple AcBel API 1PC36

Other voltages 3/5/12...?

[buz]

Re: Suspect Bad PS Apple AcBel API 1PC36

Pin 1 - 4.75
Pin 3 - 4.1
Pin 24 - 8.9
Pin 10 - 8.9
Pin 6 - 3.3

with doubled resistor

[Toasty]

Re: Suspect Bad PS Apple AcBel API 1PC36

Good. Those voltages are a clue that there is something wrong with the 5v/12v line and that's why the supply is not running. Something is not coming up strong enough to run that section. It could be on the input side or one of the diode packs has failed.

Let me get some other things together for you to test.

Good work!

Toast

[buz]

Re: Suspect Bad PS Apple AcBel API 1PC36

Thank you a million for hanging in here w me - hate to say it but it's kinda fun on this end.

[buz]

Re: Suspect Bad PS Apple AcBel API 1PC36

I figured I should be as sure as possible about my soldering to either eliminate it as a current problem or call it 'good enough'. I know it needs improving. The photos are of what I think is the worst of it.

Hate to be chasing other troubleshooting unnecessarily.

[Toasty]

Re: Suspect Bad PS Apple AcBel API 1PC36

Set DMM to Diode Check function.

PSU unplugged and no jumpers.

Do the following and hold each for 10 seconds before you take a reading.

With DMM black lead on ground / black / negative at mobo plug:
5v / red
12v / yellow

-then reverse-

With DMM red lead on ground / black / negative at mobo plug:
5v / red
12v / yellow

Report readings.

Toast

[buz]

Re: Suspect Bad PS Apple AcBel API 1PC36

Black to pin 12
Red to Pins 10 / 21 / 22 / 24
Reading = 189
Reversed = 155

Black to Pin 12
Red to Pins 3 / 4 / 15
Reading = 23
Reversed = 23

[Toasty]

Re: Suspect Bad PS Apple AcBel API 1PC36

Please Stop using pin numbers. I have to go look this up -every time-.

Use them the way I gave them, please 12v/yellow - 5v/red

AND, you can do it right on the board without using the plug.

Saying that, the 189 looks suspicious as I get .250

Your meter may vary, but isn't there a decimal point? Or a zero in front of the 23?

Ok. Time for pics to mark next locations.

Toast

[buz]

Re: Suspect Bad PS Apple AcBel API 1PC36

Ok - sorry - had no idea it was a problem and understand.

I retested at the board and received the same results - no decimals but there is a zero in front of the 23 so - 023.

My novice is showing.

[Toasty]

Re: Suspect Bad PS Apple AcBel API 1PC36

Did you ever get the solder-wick? The next test will require unsoldering some leads so the items can be tested "out of circuit", but not off the board.

I suspect one of the 12v dual-Schottky diode packs to be leaky.
This PSU does not have a pi filter on the 12v line, so I'm doubting it's the 12v cap (2200uF/16v) you replaced.

[buz]

Re: Suspect Bad PS Apple AcBel API 1PC36

Yes - got the wick, .10" blue label. Ready when you are.

[Toasty]

Re: Suspect Bad PS Apple AcBel API 1PC36

Here we go. See the pics as to what these devices are and where to unsolder two legs of each. Wiggle the leads around slightly to make sure you don't have contact to the foil traces. You can use the continuity test on the meter to be sure.

Then test them as individual diodes (4).
They should all give the same readings.
The odd ball will likely be the bad one.

Toast

[buz]

Re: Suspect Bad PS Apple AcBel API 1PC36

Took me a bit to figure out the testing but think I got it. The picture shows the 12vShottky and hoping my diagraming isn't too hard to follow.

I number the connections 1, 2, 3 across the 'top' and 4, 5, 6 across the 'bottom' as viewed from the pictures

Tested 1 to 2 and 4 to 5, both read 159

Tested 1 to 3 and 4 to 6, both read 0

[Toasty]

Re: Suspect Bad PS Apple AcBel API 1PC36

Seems like the 0 readings indicate you still have contact with the foil.

Should have 8 readings. 4 diodes X 2 readings for each.

Start with black on center pins (cathode) and red (anode) on outer. Then reverse the leads to red on center and black on outer.

The desoldered pins should be "loose" in the hole and you should just be touching the pin(s) with the probe on those. Push them a bit to see if they wiggle and the reading changes.

1 - 2
2 - 3

4 - 5
5 - 6

then:

1 - 2
2 - 3

4 - 5
5 - 6


Does that make sense? Sometimes I assume too much.

Toast

[Toasty]

Re: Suspect Bad PS Apple AcBel API 1PC36

Actually, unsolder the remaining two pins. If everything is good, the first 4 readings should match and the second 4 should read infinite (open).

0 = zero ohms, as if the meter leads are touched together

Infinite would be as if the leads were not touched together. The meter should not register anything.