iMac G5 PPC Power Supply

[CarlDay]

Hi

I have a iMac g5 psu to repair. I have changed caps but I noticed that Q1 has a hole in it and ZD6, D2 and R7, all have heat damage and I would like to change them.

Does anyone know what value they are?

The psu board is marked AP14PC47

If anyone can help that would be great.

Thanks

Carl

[Toasty]

Re: iMac G5 PPC Power Supply

Greetings!

Search is your friend:

https://www.badcaps.net/forum/showthread.php?t=13439

Are ZD6 or D2 shorted?

https://www.badcaps.net/forum/showthread.php?p=213069

D2 is an ultrafast diode of LL914/LL4148 type.

R7 is marked 102 = 1k - You figure out the metric

Photos please.

With Q1 destruction, damage to the 3845B & 4863B is also likely. Check 2 - 10Ω resistors nearby for damage (pock marks).

Toast

[CarlDay]

Re: iMac G5 PPC Power Supply

Hi

Thanks for the reply.

I have tested diodes and they are not shorted i can get a reading both ways is that good.



Cheers for help.

Carl

[Toasty]

Re: iMac G5 PPC Power Supply

>> i can get a reading both ways is that good. <<

BOTH ways? How do you mean?

Do they read as diodes should, or are the measurements equal or nearly equal? If they're equal, they're bad.

Be very careful doing any soldering in that area. The crispiness of the board means the traces are FRAGILE, and will lift off very easily.

Toast

EDIT: R7 should read <=1kΩ. If it's higher than marked value, it's bad.

.

[CarlDay]

Re: iMac G5 PPC Power Supply

Hi

I have checked again this morning (not half asleep).

D6- I get 1.1v both ways (black probe onto cathode and anode) - I think that's bad.
ZD6 - 0.5v + 1.3v i cannot see the cathode marking it too burnt.
R7 - reads just under 1k

Is it common for zd6, d6 and q1 to fail?

" With Q1 destruction, damage to the 3845B & 4863B is also likely. Check 2 - 10Ω resistors nearby for damage (pock marks)"

I have checked 2 resistors r24 & r27 are the the right ones. They look fine.

3845B & 4863B visually look ok should I change them as well or not.

Thanks again for all your help

Carl

[Toasty]

Re: iMac G5 PPC Power Supply

>>D6- I get 1.1v both ways (black probe onto cathode and anode) - I think that's bad.<<

Do you mean D2?

Remove Q1 first. If it's shorted ( obviously! ), D2 will read shorted. Once Q1 is out, then measure D2 again.

Cathode of ZD6 is on the lower end towards D2 whose cathode is up. The two have their cathodes tied together. D6's cathode is down.

I wasn't specific enough. A photo of the 4863G chip area please (lower right). The 2 resistors should be to the left of that chip and marked 100 = 10Ω. VERY tiny resistors. I have them as R23 & R116. Measure them.

Toast

[CarlDay]

Re: iMac G5 PPC Power Supply

Hi

Here is the photo of r23 and the one next to it r23 looks shot.



What should I order to fix board?

Is it possible to repair or not. I have looked at the cost of parts so far and they are only a few pence will i find a lot need replaceing.

Carl

[Toasty]

Re: iMac G5 PPC Power Supply

Shot's too close. I need to see the larger area and the chip nearby please.

[CarlDay]

Re: iMac G5 PPC Power Supply

[Toasty]

Re: iMac G5 PPC Power Supply

Thanks for the photo! This is a -bit- different, but essentially the same as the 4PC46.

The -46 uses 2 - 10Ω in series, whereas this one uses 1 - 20Ω.

If that is blown (obvious), then -most likely- the 4863G is done also.

Let me backtrack and explain a bit here. The ZD6 zener "clamps" the output (emitter) from Q1 at ~16.7vdc. Since Q1 cratered, it passed B+ (25v or more) on its collector to the emitter, then on to the rest of the the circuit and ZD6 cooked. The 3845B PWM chip can handle up to 30 volts (internal 34v zener), but the 4863G is -max- rated at 22 volts, nominal of 20 volts (internal 20v zener).

All of this comes about from a bad filter cap (C37 - 100uF/50v) on the hot side supply circuit that is powered from the TOP244Y.

This is the reason I insist on changing the small caps in these power supplies. I see so many from other techs (or who claim to be) with this cap not replaced. If it fails, B+ can rise to 40v or better. Next time the switch is thrown, Q1 is sitting there with B+ of 40v and gives the circuit a helluva shot. ZD6 clamps it and Q1 blows (I've seen them vaporized ), passing all that voltage into the 3845B which survives sometimes, but the 4863G torches the resistor and usually not in time to save it.

I would start with Q1 and R23 and make sure those 4 small caps in the Hot side are replaced. A resistance check on the input to the 4863G -may- uncover whether or not it's shorted. No guarantees. Before you power it, make sure that B+ on Q1's collector is right around 24v, no more.

Or, go on eBay and buy a nice Delta unit, recap it and be happy.

Toast

[Toasty]

Re: iMac G5 PPC Power Supply

I've attached your photos here. It's best to use the Advanced editor and use the Manage Attachments feature. That way, future members will have access to the photos in the event the remote server should delete them.

Also, here are the 3845B & 4863G datasheets.

Toast

[CarlDay]

Re: iMac G5 PPC Power Supply

HI

Thanks for the reply. I have a few more questions.

Have you got a part number for the zener diode?
I cannot get the 4863G anywhere is there a replacement?
What is the package size for the resistor is it 0402?

You have been a great help.

Thanks

Carl

[Toasty]

Re: iMac G5 PPC Power Supply

You are in the UK, yes? If you'd update your profile, it would be appreciated.

Regarding the resistor size - go by the chart in Wikipedia:

http://en.wikipedia.org/wiki/Surface-mount_technology

a 0402 metric is 0.41 x 0.20 mm

That's -really- tiny. Actually, it's the smallest. I think these are bigger than that.

More likely it's a 1005 or 1.0 x 0.51 mm

Do you have a bench power supply with variable voltage & current limiting? Might be able to test the 4863 before you pull it. At least put an ohmmeter meter across Vcc to Ground to see if it's shorted - Pins 8 to 6 both ways with the meter. Perhaps you got lucky and the resistor and internal Zener saved the chip.

These are -glued- to the board, so you have to get the legs free FIRST, or you'll damage the traces. Some solder wick first then a pin or other type pick to lift each leg while you apply the iron. Got your magnifiers on???

The 4863 shows in Mouser, Digikey, and Arrow.

If the 4863 proves to be bad, and you are unable to acquire it reasonably, drop me a PM with your address. I have them and can send a few in an envelope by 1st Class Mail. You can PayPal me the money.

Toast