another G5/661-3350 query

[CWB]

hello to all ,
i just joined up and did a quick search for some information relating to the G5 20" swmps .
i noticed in one thread that a schematic was asked for , this is about where i am at ... or a parts list .

in addition to the half-fistful of bad caps that i replaced , D10 has turned itself into a piece of wire (no magic smoke) .
i believe this diode is in one of the smaller supplies (standby/always on perhaps) . it is in association with the TOP244Y driver IC .

the oem number is T2D46 .
i have tried to cross reference this critter to no avail .
without a schematic i cannot readily ascertain if it is a zener or a fast rectifier .
does anybody have any information on this device ?

thanks in advance .

[godonr]

Re: another G5/661-3350 query

I'm looking at D10 on a 661-3350 and it's marked T2D47. Does that help ?

[CWB]

Re: another G5/661-3350 query

no cross for that number ...
it seems as if that number series is about like tracking down a set of snake lips .

do you have any idea if it is a fast rectifier or a zener ?
out of circuit measurements would have to be taken and a zener test performed in order to determine what the part actually is .

that is the problem ...
if i had a good supply in front of me i could figure out what the part is .

[godonr]

Re: another G5/661-3350 query

Maybe a clue here - http://find-chip.com/bauteil-suche/T2D47.html. I would guess fast rectifier. The boards I've worked have zener's marked ZD.

[CWB]

Re: another G5/661-3350 query

the link doesn't work (as of 0650 CST) .

i did find google results for "Bauteilsuche" ... i will explore that a bit further .

yeah , i am used to the zd delineation on boards .
but with this thing ... uuuggghh !
i am leaning towards a fast rectifier as well .

[Toasty]

Re: another G5/661-3350 query

Ultrafast - <75ns - clamping diode

Check Q1 and the 3845 IC also.

What's the PSU's board number? API4PC46 ?

Please have an admin correct the title to: another Apple iMac G5 PSU 661-3350 query.

Please don't encrypt titles. I had not seen this thread until it just happened to come up in "new posts" this morning.

[CWB]

Re: another G5/661-3350 query

the number on the case is API3PC96 .
Q1 checks ok .

nothing on the 3845 ic is dead shorted , the TOP244Y ic checks in a similar manner .
all the same , i am going to replace these two items ... cheap insurance .

just out of curiosity , any idea/thoughts as to why the numbers don't show up for the rectifier ?
the numbers on the device don't seem like "in house" markings .

[Toasty]

Re: another Apple iMac G5 PSU 661-3350 query

>>i am going to replace these two items ... cheap insurance .<<

Unnecessary and a good chance to damage the board.

Somewhere I found a cross to these over a year ago. Determine the output of the TOP and go higher by 20% or more.

"piece of wire" I assume you mean shorted? Or did it burn and that's the remains?

Toast

.

[CWB]

Re: another G5/661-3350 query

the diode went into the shorted mode ...
i don't believe the unit suffered from catastrophic failure , based on the lack of physical evidence (nothing burnt , grenaded or charred) and cursory checks .

absolutely , a board can be damaged by messing with it to no end .
as an old BE friend of mine said about a new young engineer (after the transmitter he had been futzing with for almost the whole day took a dive) :
"yessir , he done engineered it right off the air" .
this supply goes in the comp of a friend that lives in another town and hence the reason for "cheap insurance" .

thanks for the data sheet pdf .

[godonr]

Re: another G5/661-3350 query

FYI -- cheap insurance would be to replace that Q1 regardless. It is the most frequent cause of failure (in addition to caps) for this supply. Just a thought.

[Toasty]

Re: another G5/661-3350 query

Is this diode measured out of circuit? I don't think so.

Take a -real close look- at the circuit it's in.

Now, let's slow down and go back to basics:

-What was/is wrong with the PSU?
-What caps did you use and which ones have been replaced?
-Did/do you have 5vsb and 20-24v when it is not jumpered on?

Toast

[CWB]

Re: another G5/661-3350 query

a little further history ...
i recapped this supply about 8 months ago . the MB (or mid-plane board in applese) was recapped as well .
the original problem was intermittent shutdown , eventually becoming "no start" .
the physical evidence was the typical swelled and leaking caps .
sans the two raw DC input filters , all the electrolytics in the supply were replaced with nichicon and panasonic caps of the FM and FC (etc) line ; the good stuff .
many of the MB caps were swelled/leaking (as denoted elsewhere in these and other forums) . these too were replaced with "the good stuff" .
the supply came back to life and the voltages/ripple looked good . the comp performed normally , running 24/7 for 5 days , before returning it to my friend .

my friend informed me that the machine had been working ok until about three weeks ago ... it would run for several hours and then shut down .
the problem became worse (more frequent , shorter run times) , it finally became "permanent" .
my response was : "get the machine to me" .
(asides : i asked him why he did not call me with this information/condition much sooner .
reply : "... and i didn't want to bother you" .
not a good answer in my book .)

i pulled the supply and performed the cursory checks (the replacement caps are ok) , eventually i found the shorted D10 .
methinks : "aha ... that would explain the no-start condition" .
i consider this to be fortunate as opposed to the "flame out" mode .

the diode measurement is out of circuit . the low value resistor in series with that diode is ok (the resistor is encrusted with that white "toothpaste" adhesive) .
because of this condition (and other technical considerations) i have not applied ac line power .
taking things one step at a time saves a lot of headaches .

@ godonr :
thanks for the heads-up on Q1 ... do you happen to know the replacement for it ?

about the seeming lack of information and schematic availability ...
i have worked on "proprietary" equipment in the past , the kind where you had to sign a non-disclosure agreement .
i can understand/appreciate a company wanting to protect their investments .
in this case though , i really have a tough time considering a SMPS of this type to be made out of (and thought of as being made of) gold plated unicorn dung .

ps ... a message requesting a title change was sent . i imagine admin is pretty busy ... i can appreciate that .

[godonr]

Re: another G5/661-3350 query

Q1 is a MMBT2222A. Mouser carries it with a 350mW power dissip @ 25 for a $.05.

[Toasty]

Re: another G5/661-3350 query

The resistor is not in series with the diode. The diode is across the T1 primary and that's why I said I didn't think it had been pulled. I've not had 1 shorted to date, but then there's always an oddball out there in this business.

[CWB]

Re: another G5/661-3350 query

toasty ,
you are right about the resistor , closer examination shows this .
/me builds pot of stronger coffee and wipes glasses .

trouble shooting this supply would be a lot easier with a real schematic .
with D10 shorted and the fact that it is across the primary of T1 ... i would hazard that *somewhere* there is a fuseable resistor in series with the B+ to this supply . i base this on the lack of the line fuse being ok and no "magic smoke" has been loosed .
i am going to do some more examination/checks , replace D10 and see if the supply comes back to life .
i'll post my findings .

@ godonr :
thanks for the cross number on Q1 .
i have an order going into mouser and i'll order up a few .

ps ... i talked to my friend and brought him up to speed on the happenings .
he informed me that he has located a couple more G5s that are dead in the water ... something about co-workers (graphics design outfit) personal comps .
i dunno ... at this point in time i really have trepidation about opening up another of these supplies .

[Scenic]

Re: another G5/661-3350 query

look at it this way.. if you figured out what's wrong with this one, the other ones might be quick fixes.. and easy money

well.. sort of easy..

[CWB]

Re: another G5/661-3350 query

yeppers ...
no more difficult than nailing jelly to a tree .

it isn't about the money ... i don't believe that one can put a price on sanity .

[CWB]

Re: another G5/661-3350 query

after getting "un-busy" and getting on the downhill side of the creeping crud that is going around i did some more checking on the supply ...

i replaced the diode that is across the transformer and double checked the drain and source leads of the driver ic ... so far so good .
i then powered up the supply and "sniffed" the transformer with a scope .
the supply is trying to fire up ... it pulses for a short interval , stops and pulses again . it does this indefinitely .
it almost acts like there is a helluva load on the secondary .
this whole repair would be easier with a schematic ... one would think that *somewhere* there is a schematic available .
as soon as the other parts i ordered get here (i put them on another order) i will pursue this matter further .

in the meantime , in order to get the comp back to my friend , i am going to order up a (supposedly) brand new supply ...
i noticed that a couple of outfits are selling these supplies as "brand new in the box" . of course there is a wide range in the pricing between them .

[godonr]

Re: another G5/661-3350 query

in the world of apples and oranges... the API3PC96 (661-3350) is the simpler (as in uses only one PWM among other things) of the two (API3PC46) supplies that pop up here. The 661-3350 can shut down because of issues with the m/b as well. I've seen failing HDs cause this problem in iMacs. Make sure you don't have any other issues outside of the power supply. They can shorten the life of these supplies.

[CWB]

Re: another G5/661-3350 query

MB problems ... eeeyeah ...
/start rant
originally my friend had got a quote from a "mac shop" (i'll bet all that work there are left handed , but i shall digress) . the price he was quoted was for a "new" MB ... i verified this by asking him if they had cracked it open and took a look around .
he replied that they did not open it , the tech took a look at it externally and then made his (in my opinion) blind pronouncement .
i asked my friend if they mentioned replacement of any other parts ... he replied "no" .
for the price he was quoted i could build a windows or ubuntu box that would run circles around his G5 ! i believe in fair profit but it was ridiculous .

i opened the machine ... yep , bad caps on the MB .
i opened the swmps ... my suspicions were confirmed ... plumb nasty looking .
in theory ; they would have "replaced" (i have my doubts on this one) the MB (for a high dollar price) and the machine might have fired up long enough to get it home and then fail . the next words would have been "sorry charlie ..." .
from there , i can imagine the pitch to sell him a new overpriced machine .
anyway i look at it i see it as incompetent and questionable in integrity .

i am a BE and have been for many years , along the way i have also worked for many consumer and industrial electronic repair firms/shops , have wall paper from many schools that were required for warranty/certified status , etc ...
i have had to testify in court as to my findings in fraudulent repairs , worked on the "tough dogs" at the request of another shop , etc .
it really p***es me off when i learn of practices/individuals like these that give the industry a bad name .
and when it involves/happens to a friend ... well , you might imagine that it really gets my attention .
/end rant

the checks i performed on the supply after replacing the diode (etc) were done both in and out of the machine .

on a further note ...
as i see it there is marginal cooling and cramped conditions in the supply ... this makes me wonder if a suitable external supply (fed through the old power cord hole) would not be a better/cheaper/easier deal , all things considered .