I mac G5 PSU, ALS - odd voltages...

[trex66]

Re: I mac G5 PSU, ALS - odd voltages...

I can tell you pin 15 (on/off) sits at 3.3v (off) and drops to 0 when the supply comes on.

My pin 15 was at 5V before this now it's 3.3V.

I want to jump it to see what will happen, but I am afraid something might pop if I do.

Can't help thinking this change has something to do with the Q1 transistor I replaced. I thought I did a good job, but it's possible that I overheated it or something. It's so small I don't know how you guys do it.

[Toasty]

Re: I mac G5 PSU, ALS - odd voltages...

Use a ~50Ω resistor as the jumper.

[trex66]

Re: I mac G5 PSU, ALS - odd voltages...

jumped it and there was no change, all the voltages stay on.

[godonr]

Re: I mac G5 PSU, ALS - odd voltages...

Check your soldering again.

[trex66]

Re: I mac G5 PSU, ALS - odd voltages...

looked over all the solder points and they looked good. I replaced the Q1 transistor just to be sure it was OK. And now I'm back to where I started. When off I'm getting 45-47V on the 3.3V pins. When jumped I get 3.3V on the 3.3V pins, the 12V pins are 12.6V and the 5V pins are 5.3V. Also the 24V pin stays at 24V when on or off.

I don't see the voltages rising like the did before, maybe that's fixed.

[trex66]

Re: I mac G5 PSU, ALS - odd voltages...

I was reading some other post and checked D10. It had continuity both ways, so I pulled it out and tested again, same result.

The code on this diode is T2D 40. Toasty posted a pdf for T2D diodes, just not sure which one (UF4001-UF4007)applies to my Diode(T2D 40).

[godonr]

Re: I mac G5 PSU, ALS - odd voltages...

Wow... in the same but different world now. The AcBels I have are AP13PC96s... close but not the same as yours. Mine only has one 3845B PWM and fewer caps. Does your board use a LM339 Quad Compartor? Your voltages seem right when on and something's telling it it's ON all the time. Are you seeing 47v @ pins 1,2 and 12 unjumpered?

[trex66]

Re: I mac G5 PSU, ALS - odd voltages...

Yes, 1,2, and 12 pins are at 47V when unjumpered. But the 5V pins (not trickle) are 0V before jumpered so I don't think it's on all the time.

I'm sorry godonr, I have no idea what a LM339 Quad Compartor is? If you could describe what one is I will look.

I'm an electrician by trade, I thought this would come easy to me but it hasn't.

A month or so ago I had a monitor quit on me and there were some guys on here that helped me through the troubleshooting and got it back up. I was a really great experience.

This Imac has sat in the basement for a while because it will just shut down whenever it wants to. So no one wants to use it.

I bought another Imac that had a known good PSU to get this one up, but it turned out to be bad as well. It's a Celetronix and from what I hear, not worth the time to fix.

I thought I had this PSU almost fixed when the voltages came up normal and stable (but on all the time). After I replaced the Q1 again, I am back to square one. Now I found a shorted Diode that is impossible to trace. I'm starting to get a little frustrated with these PSU's.

[Toasty]

Re: I mac G5 PSU, ALS - odd voltages...

It is very odd that D10 reads as shorted as it is across the primary of T2, the 5/24sb transformer. If you are shorting -across- the primary, there is no way that the 5/24sb voltages could be present.

Are you using an analog or digital multimeter? When you test the diode, you get the same reading both ways, as if you touched the test leads together?

I've pulled one from a good unit and it shows open both ways on both types of meters on all scales. It does not forward-bias until it's over 80v. Reverse is better than 500v.

I'm very puzzled as to how to classify this diode...?

Looking at the TOP datasheet, this could be an ultra-fast such as the UF numbers, or an avalanche or transient voltage suppressor (TVS). None of the circuits are a match to this PSU's circuit. There is always another device or 3 between its cathode and the other side of the transformer primary. The anode is always tied to the Drain on the TOP and the primary of the transformer.

Toast

[trex66]

Re: I mac G5 PSU, ALS - odd voltages...

I have a few DMM, the most expensive one is a fluke 113. I retested the diode with the fluke and now it is testing open both ways. I tested with two other DMM I have and same result.

I'm not crazy, yesterday it tested shorted both ways?

I had just had the PSU powered up before I tested it, maybe was locked closed?

I could always put it back in the PSU and see if it will test shorted again.

I was looking for that diode and ran across this post from jimwarholic site concerning a similar diode - T2D 47:

I can`t find info for this diode, but their purpose is to rectify the winding output AC voltage and provide DC supply to the PSU controller IC. To confirm that, trace the cathode (the terminal in the band side) and check if that ends in the IC controller (one of the two SMDs ICs on the back of the board, as I see in your pictures).

If so, you can safely use an 1N4934 or other FAST RECOVERY RECTIFIER DIODE, of 1A and 100V.

[Toasty]

Re: I mac G5 PSU, ALS - odd voltages...

His assumption is incorrect. This diode is not rectifying the output AC. Rather it is snubbing or clamping the output (drain) of the TOP, driving the transformer (T2). The secondary on the primary side is driving the diode D18 (just behind the TOP in front of T2) which is creating the PWM and Optoisolator power.

>>maybe was locked closed?<<
They don't do that.

>>I could always put it back in the PSU and see if it will test shorted again.<<
Shorts don't come and go...

[trex66]

Re: I mac G5 PSU, ALS - odd voltages...

I went ahead and put D10 back in and jumpered the PSU, then unplugged it. And it tests shorted again, both ways. I don't know what to make of it.

I started to look for short on the 3.3V pin. The only thing I can find is that it has some continuity to the ground pins 141Ω. The 5V pin has 57Ω and the 12V pin 155Ω. The other pins have no continuity to ground. These are the pins that are turned on when the PSU is jumpered. On the 24V area there is a M17 that runs to C67 and the 24V pin. I am reading some continuity between the middle leg (ground) and the leg that runs to the 24V pin which means I see continuity between the 24V pin and the 3.3V pin. I pulled the C67 which spans these two legs, and still see continuity. Does this mean the M17 is bad?

As I mentioned before the 3.3V reads 47V with little current, when I measure the 5V and 12V pins they show a higher voltage but quickly drop until they reach 0.1V. So I think all three pins are affected by this, I don't know why the 3.3V pin shows a high Voltage?

[Toasty]

Re: I mac G5 PSU, ALS - odd voltages...

>>I went ahead and put D10 back in and jumpered the PSU, then unplugged it. And it tests shorted again, both ways. I don't know what to make of it.<<

Again, out of circuit? Now it shows shorted?

WTH is a M17???

>>The other pins have no continuity to ground.<<

What other pins? You mean the ground pins...? Continuity could be through a capacitor that you are charging when you apply the ohmmeter.

Do you have a 1/2 watt or better resistor in 47-75Ω lying about? If so, I want you to put it on the connector from any of the 3.3v lines to a ground pin. Then measure the 3.3v and see what you get for readings.

Toast

[trex66]

Re: I mac G5 PSU, ALS - odd voltages...

>>I went ahead and put D10 back in and jumpered the PSU, then unplugged it. And it tests shorted again, both ways. I don't know what to make of it.<<

Again, out of circuit? Now it shows shorted?

I didn't take it back out to test it again, it tested OL in and out of circuit before power and 0.00 in circuit after power - I don't know why it's testing open and shorted. But I'm getting the 5/24V so I thought I should look somewhere else for the problem.

WTH is a M17???

I know, but that's how it labeled on my board, pic below

>>The other pins have no continuity to ground.<<

What other pins? You mean the ground pins...? Continuity could be through a capacitor that you are charging when you apply the ohmmeter.

I mean the other points on the board where the wires for power go to the connector (24V, 5V, 5Vsb, ON/Off, PG).

Do you have a 1/2 watt or better resistor in 47-75Ω lying about? If so, I want you to put it on the connector from any of the 3.3v lines to a ground pin. Then measure the 3.3v and see what you get for readings.

Toast

1/2Watt 66Ω = 14.6mV on the 3.3 pins

[Toasty]

Re: I mac G5 PSU, ALS - odd voltages...

Again, you can not measure D10 with it in circuit. It is -across- the T2 primary and will -always- read shorted in both directions. The only exceptions would be and open primary in T2 or a broken connection between the two components.

M17 is a KA78M24. A 24 volt, 1/2 amp, positive voltage regulator. It's what "makes" the 24 volts.

[trex66]

Re: I mac G5 PSU, ALS - odd voltages...

Toast, you never said it couldn't be measured in circuit. I powered up the PSU and removed D10 and it tests shorted - both ways. This is exactly how it tested before, tomorrow it will probably test open.

So is this a bad diode, should I replace it and if so where do I get one? Or is this acting normal and I should be looking somewhere else for the problem? If this is bad, how could it leak voltage to the 3.3V pins?

[Toasty]

Re: I mac G5 PSU, ALS - odd voltages...

Seem to be stuck in a loop here.

IF the diode were bad, you'd have NO 5/24vsb nor would the "Power On" circuit work.

Where this diode is located is on the "HOT" side of the power supply. Meaning it is connected to the mains power. Your problem is on the "COLD" or isolated side.

Check C59 & D21. All are near the big cap in front of M17. On the output leg from T2 that feeds D21, is the -only- place I could find anything near 47v when the supply was off -or- on, and that was when the meter (analog) was set to AC. That side of the secondary is the feed for the 24vsb.

If you look at the underside of T2, with it towards you, the 4 pins to the left are output. The center 2 are tied together and go to ground. The pin closest to you, and the edge of the board, is the 5vsb supply which goes to D20 and then is filtered by C40, a coil, and C45. The pin at the top of those 4 from T2 feeds the 24vsb as discussed above.

Yet, none of this is connected to the 3.3v line.

Toast

[trex66]

Re: I mac G5 PSU, ALS - odd voltages...

I pulled and tested C59 and D21 and they are oK.

I been tracing the 24V and the 3.3V circuits for hours and I just can't see any component they share. They don't even come that close to each other. I'm still trying to figure out how the 24/5V circuit could produce 47V?

[trex66]

Re: I mac G5 PSU, ALS - odd voltages...

I think I found the problem. C45 has badly burned pads, I think it's shorted into the 3.3V trace. And the two coils on the 3.3V line are increasing the voltage to 47V.

If this is where the 47V is coming from then do I need to fix it or will it function OK?

[godonr]

Re: I mac G5 PSU, ALS - odd voltages...

That would make sense... sorta. I can tell it was a partial recap from the green Sanyos. Was C45 one of them? Did this fix your problem? It's always something simple... after you find it, right!