Another sick iMac G5 PSU

[epolinda]

Hi, everybody!

This is my first post in this forum, so I'd like to start saluting all of you guys who provide this great help to us who suffer from the capacitor plague.

I happen to have a iMac G5 17" ALS that was dead when I bought it. I had to replace the mobo, once the recapping was not enough to bring it back to life.

It's PSU was also damaged, but I managed to make it work by changing the caps and one of the transistors close to the PWM IC (3845).

Some weeks ago, a power surge killed the PSU again. When I tested it I could only find the 24V and 5V tensions that I now know (by reading this forum) come from the section thats driven by the TOP245 IC.

When i was testing the transistors close to the PWM, those voltages have also stoped working and I started to search for some problem close to the TOP245. Finally, I found an open resistor connected from the primary voltage (in my case, nearly 300V) to the transformer thats driven by the TOP245. As I didn't have at hand a resistor with exactly the same value (5,1 ohms) I replaced it with another one with a higher value (10 ohms) just for testing purposes. When I connected the mains, both the resistor and the TOP245 exploded leaving behind just ashes and debris...

I could not find a TOP245 to replace the blown one in my city (Brasília, DF, Brazil), so I tried to find it on the net. Even though I could not find it in any of the sellers I know in Brazil, I managed to find the TOP247 that the datasheet points as equivalent and capable to drive higher power.

I replaced the remains of the TOP245 with the brand new TOP247 and put also a new resistor where the old one once was. When i connected the mains this time (very carefully and from a long distance!) nothing exploded, but the voltages that should be coming out from the PSU are reduced. Where there should be 24V, I found 10V and where it was supposed to be 5V, it was 2V.

This is the point where I'm stuck right now. I'd really like to hear some thoughts about this situation. Are TOP245 and TOP 247 really equivalent? Don't you find strange that the voltage was reduced to a integer value in both cases? Is it a coincidence?

Any ideas at all will be very, very, very welcome!

Thanks again

Emerson

[Toasty]

Re: Another sick iMac G5 PSU

247 is 2x the current rating of 245. Otherwise equivalent.

Apple P/N please. Mains voltage? Pics of unit also.

Sounds like the flyback diode is failed or the 5vsb transformer is shorted internally.

[epolinda]

Re: Another sick iMac G5 PSU

Hi, Toasty!

Thanks for the answer.

The Apple part number is 661-3627. Mains voltage is 220V.

I'm not at home right now, but when I'm back there latter, I'll take some pictures to post here.

Is the flyback diode the one that is connected parallel to the transformer primary? If that's the case, I've tried to test it and it appeared open to me. But, reading another thread in this forum I saw a post from you saying that it could not be measured with a ohmmeter, and that it would test allways open. So I left it there believing it should be good.

To make things worse, that diode has a marking that can not be found anywhere on the net. Do you know of any equivalents that I could try to use to replace it?

Thanks again,

Emerson

[seanc]

Re: Another sick iMac G5 PSU

I just found a 267 on eBay for mine (should arrive tomorrow), had a quick look for a 245 for you and found one from the same seller. It's from the UK so shipping may take a while for you.
He has lots of bits for iMac and Apple PSUs.

http://cgi.ebay.co.uk/1-x-TOP245-TOP...item335981cfb0

[epolinda]

Re: Another sick iMac G5 PSU

Toasty,

I've taken some pictures of the PSU. Sorry they are not very good, once I've used a cell phone for that.

I hope the problem is the diode. I wouldn't know how to replace a transformer like that.

I had no time to replace the diode last night but I'll try to do that tonight. Do you believe I could use UF4007?

Seanc, thanks so much for the info. I'll try a little more with the 247 and if it doesn't work, I'll go for the guy you pointed.

Best regards,

Emerson

[Drexter]

Re: Another sick iMac G5 PSU

I'm not an expert, but I would recomend you check first what caused the failure, you might keep burning the TOP245 or damaging componets till you fix first what caused it.

[epolinda]

Re: Another sick iMac G5 PSU

I've got some news.

Measuring the voltages on TOP247, I found out that the tension on the pin connected to the optocoupler SFH615 was always the same (5V), no matter what the PSU output was (yes, it was changing each time I conneted it to mains).

I decided to take it off the board and test it. Not to my surprise, the photo-tansistor side of the optocoupler was shorted, keeping 5V at one of TOP247 pins (I believe it's the regulation one).

I'll try to replace it tonight and see what happens.

The results will be posted here.

[epolinda]

Re: Another sick iMac G5 PSU

Well, new developments.

I've replaced the optocoupler and turned the PSU on. The tensions are still wrong and now the PSU makes a ticking noise. It ticks something like once a second. Another problem is that now TOP247 gets hot. I haven't measured the temperature, but it is still possible to touch it and keep the finger on it.

Checking other components close to the TOP247, I found resistor R70 open and the capacitor close to it (C41) seems to have some leaking. I've removed the resistor and confirmed it is open, but I didn't remove the capacitor because it seems to be a little damaged and I was afraid it would not survive the operation, but I believe I'll take it out tonight and test it off circuit.

I'd really appreciate that anybody could tell me the values of those components, once the cap has no writing and the resistor was impossiblo to read.

I could not make a positive identification of the maker of my PSU. All I could find was a writing on the PCB saying "API4PC47". It also says "Rev. B" somehere.

Could anybody please help me?

[Toasty]

Re: Another sick iMac G5 PSU

Your board layout is -similar- to the 661-3350. There is a spot for R70 but it's not used on these. (It is next to C39 chip cap between the transformer legs, correct?) The number is super tiny and you need a jeweler's loupe or similar magnifier with a 10x power.

C41 is a can cap or chip?

Check the transistors around the 3845 PWM chip. The last fix was Q1.
See: https://www.badcaps.net/forum/showthread.php?t=8665

AcBel is manufacturer.

[epolinda]

Re: Another sick iMac G5 PSU

Toasty,

Thank you so much for your help. It is specially very useful to know who was the maker of the board.

You are right. R70 is placed between C39 termminals and is connected to it. C41 is a SMD cap. I'm attaching a little schemmatic (very basic and quite incomplete) for this section of the PSU. I hope it serves to make easier for me to explain what I'm doing. Maybe it's also useful for someone else.

Correct me if I'm wrong, but it seems to me this section is independente of the others and it should work despite of the other tensions.

Let me tell you what I've done untill now. I've replaced R70 with another smd resistor. It was completely impossible to find a 6R8 SMD resistor, so I tried to use a 4R7 one. That value of 6R8 was taken from the TOP242-TOP250 datasheet. It appears there in some sample circuits in a configuration very similar to the one I found on the PSU.

Replacing the resistor did not bring the section back to life, but it changed (as a matter of fact almost disapeared with) the ticking noise.

Measuring the tensions, I found 5.3V on pin 1 of the TOP247. Once the tension on pin 4 of the SFH615 varies from 2 to 4 V, it seems to me it will never be correctly polarized and so the optocoupler will remain inoperative. Then I thought that the SCR (MCR100-6) could be damaged. Took it off, tested it and it really seemed to be defective. I've already bought a new one and I'm waiting to have the time to change it. Until then, I'm going to read a little more of the datasheet and try to understand more of TOP's workings.

For my despair, I'm starting to believe that the transformer could really have some shorts. Do you think it's possible to test it? How could that be done? If it's really damaged, do you there's something I could do to replace it? Should I just dump this PSU?

Once again, thanks for your interest and help.

[Wizard]

Re: Another sick iMac G5 PSU

Draw out the schematic bit more please. I'm not seeing complete circuit and it is drawn like both hot side and cold side cross connected unless dashed line is sectioned thru the transformer and seperate the 2 windings with this dashed line and marked as such hot side and hot side. Typically most SMPS schematics had all the hot side windings on one side of transformer and other side of transformer as cold for clarity.

Cheers, Wizard

[epolinda]

Re: Another sick iMac G5 PSU

Wizard,

I've made a mistake when uploading the schematic and it was cut on the right side. Now I'm sending it complete. I've also drawn a thin blue line separating hot and cold side. Hot side is left, cold is right.

Last night, after my previous post, I could not sleep and decided to try to change the MCR100-6 that I had bought. No change. I finally decided to replace also TOP247 with another one. No change.

The PSU still makes a noise (something like a very brief "cheep!", with nearly 3 seconds of interval. It is coming out of the transformer shown on the schematic. In a desperate action, I shorted the secondary of that transformer and the noise stoped.

I'm more and more convinced that the problem is the transformer, but I won't give up without proving that.

But at this moment, I confess I'm out of ideas.


Emerson

[Toasty]

Re: Another sick iMac G5 PSU

Hot side(mains) and cold side(output) are -completely- electrically isolated. The only "connection" between them is the optoisolator and the transformer. Your schematic is not correct in that respect. It is very difficult to trace these particular PSU's because of the many thru-connections from the solder side to the top and back again.

Have you tested the transistors surrounding the 3845 PWM?
Was the original R70 actually marked 6R8?
Did you replace the opto with an identical unit?

The ticking or chirping sound is the unit trying to start and the PWM protection kicking in and shutting it down.

[epolinda]

Re: Another sick iMac G5 PSU

Toasty,

I'll try to revise the connections and draw everything from the beginning.

I did not test all of the transistors, because I thought that section of the circuit was dedicated only to the other voltages (+5V, +12V and +3.3V). But if you really believe I should do that, then I will.

The markings in R70 were unreadable, even with strong lenses. I took that value from the datasheet. I thought it could be correct, once the caps surrounding it had the same values on the board and on the sample circuit of the datasheet. I could not find a SMD resistor with that value, so I'm trying to use a 4R7 one. Do you believe that value would be too critical?

Finally, I could not find an identical opto unit. I'm using a TLP621 wich has the same pinout and similar characteristics. I'll check a little further for the correct opto.

Thanks for the ideas.

Emerson

[Toasty]

Re: Another sick iMac G5 PSU

See stj's response in this thread:

https://www.badcaps.net/forum/showthread.php?p=95789

[koektrommel]

Re: Another sick iMac G5 PSU

And a bump upwards!

I have a PSU with a Q1 blown also. I've found a (zener)diode with weird readings;
It's the same as the diode directly under the circled position ; diode ZD6
See an earlier posted picture from another user: https://www.badcaps.net/forum/attach...chmentid=13706

It's readings out of circuit, using a DMM in diode-measurement: 045 and 045. Both directions the same? It's allready replaced once. On another identical PSU which didnt show ant resoldering marks, there's another diode that has only one big yellow stripe on it. That one was working well and was the same type PSU.

Can someone please tell me what type this diode at position ZD6 is?