I'm delighted to see Paul in on this PSU problem. I have a PC Power &
Cooling Silencer, 610W EPS12V. It went belly up about 6 months ago, just
after the 3 year guarantee was up, and I came close to dumping it
several times since but, because it cost far more than the earlier PSUs
I had bought, I hesitated. They were only 250W-350W.

The unit stopped with a loud pop. I was never sure about where it came
from. I had a very good power strip TrippLite and it has been working
fine before and after that pop. So I figured that it most have come from
the PSU or the motherboard. There was no sign of a blown capacitor on
the motherboard, so I guessed that the pop came from the PSU.

Next I took the cover off the PSU and searched for traces of damage but
found none. All capacitors seemed ok and no burn nor melting solder.
The unit had a circuit board about 5.5"x5.5" There were 3 units like a
cart wheel but without the spokes. They were about 1.5" diameter
outside, 5/8" diameter on the inside and thickness about 5/16". A thick
copper wire was wrapped around the wheel passing through it and around
and around again. The ends were then soldered to the circuit board.

Close to one of these copper-wire-wrapped coils was a small very busy
circuit board - about 1.25" x 1.25". It was bent away from the coil and
it seemed to be pushing a capacitor, that was right beside it, - both
leaning over about 30 degrees. These could not be pushed unless the
cover was taken off.

All in all I could not see any damage. Finally, with the cover off, I
decided to plug in the unit and switched it on. There was no fan
movement and no hum. While moving the heavy cables around, the back of
my hand touched on one of the aluminum units in the box. That made me
jump a little, but at least it told me that the unit wasn't totally
dead. There were 2 aluminum units with a strange structure - probably
for thermal dissipation- heat fins?. The lengths were about 4", heights
about 3", and there were 6 fins - 3 top ones about 1 1/4" wide with
about 1/8" between them. Then there were three lower lower ones the same
length - about 4", and of width 3/4" , again with 1/8" space between
them. The section looked like an artificial tree.

Comments greatly appreciated.

Seum wrote:
> I'm delighted to see Paul in on this PSU problem. I have a PC Power &
> Cooling Silencer, 610W EPS12V. It went belly up about 6 months ago, just
> after the 3 year guarantee was up, and I came close to dumping it
> several times since but, because it cost far more than the earlier PSUs
> I had bought, I hesitated. They were only 250W-350W.
>
> The unit stopped with a loud pop. I was never sure about where it came
> from. I had a very good power strip TrippLite and it has been working
> fine before and after that pop. So I figured that it most have come from
> the PSU or the motherboard. There was no sign of a blown capacitor on
> the motherboard, so I guessed that the pop came from the PSU.
>
> Next I took the cover off the PSU and searched for traces of damage but
> found none. All capacitors seemed ok and no burn nor melting solder.
> The unit had a circuit board about 5.5"x5.5" There were 3 units like a
> cart wheel but without the spokes. They were about 1.5" diameter
> outside, 5/8" diameter on the inside and thickness about 5/16". A thick
> copper wire was wrapped around the wheel passing through it and around
> and around again. The ends were then soldered to the circuit board.
>
> Close to one of these copper-wire-wrapped coils was a small very busy
> circuit board - about 1.25" x 1.25". It was bent away from the coil and
> it seemed to be pushing a capacitor, that was right beside it, - both
> leaning over about 30 degrees. These could not be pushed unless the
> cover was taken off.
>
> All in all I could not see any damage. Finally, with the cover off, I
> decided to plug in the unit and switched it on. There was no fan
> movement and no hum. While moving the heavy cables around, the back of
> my hand touched on one of the aluminum units in the box. That made me
> jump a little, but at least it told me that the unit wasn't totally
> dead. There were 2 aluminum units with a strange structure - probably
> for thermal dissipation- heat fins?. The lengths were about 4", heights
> about 3", and there were 6 fins - 3 top ones about 1 1/4" wide with
> about 1/8" between them. Then there were three lower lower ones the same
> length - about 4", and of width 3/4" , again with 1/8" space between
> them. The section looked like an artificial tree.
>
> Comments greatly appreciated.

If you want to learn more about power supplies, Jonnyguru.com has reviews
done with a load box. This is a PCP&C 750W for example. This one is
made by Seasonic. PCPower contracts out, like Antec does. Seasonic
is a primary manufacturer, while PCP&C and Antec are middle-men.

http://www.jonnyguru.com/modules.php...Story4&reid=63

The toroidal magnetics, are to contain the magnetic field. A toroid is
very nice in that respect. You can make "inductors" or "transformers".
An inductor is one coil of wire. A transformer (at least one that isolates)
uses two coils of wire. The ones in the picture look like they're inductors.

There is a "regular" transformer between the two aluminum heatsinks. And
the big primary storage cap is there too. That transformer between the
heatsinks, could be for the single 12V output rail.

Modern 80+ (high efficiency) power supplies are two stage. They work like
this.
12VDC
|
120V --- 300VDC --- switching_conversion ---+--- 12VDC to 5VDC --- 5VDC
transformer/switching and 3.3VDC --- 3.3VDC
transistors, rectifiers conversion
caps. board

In such units, there could be a separate printed circuit board, to do the
second stage of conversion.

There can also be a small circuit board, implementing input filtering.

Or even a PCB near where the cabling is connected. Power supplies with
modular cabling, need to mount the modular connectors somewhere, and
that may involve a piece of PCB material at 90 degrees to the main
PCB at the bottom of the unit.

Notice in the Jonnyguru picture, there is a fuse near the inlet. That
is for safety, in the event there is a short on the primary side.
The muffled pop could be coming from the fuse blowing violently. A
fuse like that would have to be slow blow, because of the abuse the
fuse takes at startup. There is a huge current inrush when an ATX
supply starts up, as the input cap is charged from the line. In the
old days, a negative temperature coefficient device was placed in
series, to reduce the magnitude of the inrush - we had some supplies
at work, where that component was what was causing the power supplies
to fail. That device runs hot (as "hot" is its most-conducting state
and has the least voltage across it). The difference between its cold
state and its hot state, is how it moderates inrush. But the price paid,
is an energy inefficient solution (heat = waste).

I like this diagram, because it's instructive of the basic architecture
of PC supplies. This one is different than the 80+ efficient units,
in that all three major rails, come from the same conversion step.
And that was necessary, because heavy currents could be coming
from any of the major rails, instead of just +12V.

http://www.pavouk.org/hw/en_atxps.html

You can see in that diagram, NTCR1 on the left, is there for inrush limiting.
And you can also see the presence of the safety fuse on the top left as well.

You can check the fuse with a multimeter. But if it is open circuit, then
chances are, there is a short "downstream" of the fuse. Like a blown
bridge rectifier, a bad switching transistor (of the pair of them),
and so on. I'm not sure I could unsolder and test all these items - the
bridge rectifier is relatively easy to test (with the diode range on
my multimeter). The transistor would be more trouble, especially if
it can't be identified. In school, we used a curve tracer to
characterize transistors, but I haven't seen one of those for
eons.

Paul

Seum wrote:
>
> I have a PC Power &
> Cooling Silencer, 610W EPS12V. It went belly up about 6 months ago, just
> after the 3 year guarantee was up, and I came close to dumping it
> several times since but, because it cost far more than the earlier PSUs
> I had bought, I hesitated. They were only 250W-350W.
>
> The unit stopped with a loud pop. I was never sure about where it came
> from. I had a very good power strip TrippLite and it has been working
> fine before and after that pop. So I figured that it most have come from
> the PSU or the motherboard. There was no sign of a blown capacitor on
> the motherboard, so I guessed that the pop came from the PSU.
>
> Next I took the cover off the PSU and searched for traces of damage but
> found none. All capacitors seemed ok and no burn nor melting solder.
> The unit had a circuit board about 5.5"x5.5" There were 3 units like a
> cart wheel but without the spokes. They were about 1.5" diameter
> outside, 5/8" diameter on the inside and thickness about 5/16". A thick
> copper wire was wrapped around the wheel passing through it and around
> and around again. The ends were then soldered to the circuit board.
>
> Close to one of these copper-wire-wrapped coils was a small very busy
> circuit board - about 1.25" x 1.25". It was bent away from the coil and
> it seemed to be pushing a capacitor, that was right beside it, - both
> leaning over about 30 degrees. These could not be pushed unless the
> cover was taken off.
>
> All in all I could not see any damage. Finally, with the cover off, I
> decided to plug in the unit and switched it on. There was no fan
> movement and no hum. While moving the heavy cables around, the back of
> my hand touched on one of the aluminum units in the box. That made me
> jump a little, but at least it told me that the unit wasn't totally
> dead. There were 2 aluminum units with a strange structure - probably
> for thermal dissipation- heat fins?. The lengths were about 4", heights
> about 3", and there were 6 fins - 3 top ones about 1 1/4" wide with
> about 1/8" between them. Then there were three lower lower ones the same
> length - about 4", and of width 3/4" , again with 1/8" space between
> them. The section looked like an artificial tree.

You may want to check HardwareSecrets.com, which explains PSUs and has
lots of detailed photos of their insides, including with parts
removed (they unsolder many of the PSUs they review). Repair
information can be found at www.RepairFAQ.org and the forums of
BadCaps.net (several forums, one just for PSUs).

Don't run the PSU except with its cover in place and fully screwed
down, because it blocks chunks from flying out violently if something
explodes, and it prevents electric shock because there's a lot of
exposed high voltage in the PSU, sometimes including at least one of
the big heatsinks. Besides, there's no need to have the PSU running
or even plugged in to test it. You can check most components with
just a multimeter's ohms and diode check functions -- see the
excellent www.RepairFAQ.org . However you must first discharge any
high voltage remaining in the big capacitors. They almost always
discharge automatically in a few seconds, an exception being if the
PSU fails in the high voltage section. Unfortunately PC PSUs are
usually so cramped that it can be hard to reach the components to
discharge them without first exposing yourself to potential high
voltage.

Usually when you hear a pop, a capacitor has ruptured, at either top
or bottom (rubber plug on bottom), but sometimes it's a fuse, big
transistor, or big diode. Here may be a device that looks like a big
transistor but is actually a chip that handles the +5Vstandby supply.
Fuses are sometimes covered in heatshrink, and transistors and diodes
are on the big heatsinks and hard to remove individually, so it's
usually easier to just unsolder everything attached to the heatsink
and lift out the whole heatsink. Transistors and diodes are usually
electrically insulated from the heatsink with a sheet of silicone
rubber and maybe a nylon shoulder washer for the mounting screw, and
any replacements have to be insulated as needed for their particular
packaging style. Insulation has to be tested with an ohm meter,
especially for high voltage components.

Electrolytic capacitors probably fail more often than any other
components, but they can't be tested very well with just a multimeter,
even if the meter includes a capacitance measurement function.
Capacitors that have ruptured or are bulging are definitely bad, but
so can those that look normal on the outside.

edfair wrote:
> Your cart wheels were inductors with ferrite cores.
>
> The metal stuff was heat sync to remove heat from the switching
> transistors that were mounted to them somewhere, probably near the
> base.
>
> The pop was probably one of the switching transistors. If so, it may,
> or may not have done any outside damage.
>
> Yeah, brushing against a loaded component can get your attention. I
> picked up a Xerox 820 supply with a blown bleeder and partially
> discharged the supply through the palm of my hand. Still hurts to
> remember it.

Many thanks to Paul, L.M.& C, and Edfair. I'll send my first attempt now
and will read your emails today.

This morning I took the cover off the PSU and made a few measurements
(for what they are worth. :-) )

With the cover off I found a small 1.75" square circuit board just
inside the external socket. This little board has 3 connection wires,
one to ground and the other two to the mainboard. I made 3 resistance
measurements on those 3 wires from the prongs of the socket to the case
(the ground) and from the two other connections on the mainboard.

Resistance between the ground pin in socket and the ground pin on the
PSU - 0 Ω.
Resistance between the N(eutral) pin and the connection to the
mainboard: 0 Ω.
Resistance between the L(ive) pin and the connection to the mainboard:
- Infinity Ω.

Is it possible that something blew on that little circuit board? It has
a piece of metal box that passes through the rear wall of the box and
the socket has a plastic piece that is fixed to the rear wall of the box
with 2 screws. I took them out and tried to pull off the plastic piece
in an effort to get the circuit board out of the box to see if there are
any damaged components on it. The I-0 switch is near the socket and is
firmly glued into the little circuit board.

Now back to your writings :-)

Thanks for the help.

Seum wrote:
> edfair wrote:
>> Your cart wheels were inductors with ferrite cores.
>>
>> The metal stuff was heat sync to remove heat from the switching
>> transistors that were mounted to them somewhere, probably near the
>> base.
>>
>> The pop was probably one of the switching transistors. If so, it may,
>> or may not have done any outside damage.
>>
>> Yeah, brushing against a loaded component can get your attention. I
>> picked up a Xerox 820 supply with a blown bleeder and partially
>> discharged the supply through the palm of my hand. Still hurts to
>> remember it.
>
> Many thanks to Paul, L.M.& C, and Edfair. I'll send my first attempt now
> and will read your emails today.
>
> This morning I took the cover off the PSU and made a few measurements
> (for what they are worth. :-) )
>
> With the cover off I found a small 1.75" square circuit board just
> inside the external socket. This little board has 3 connection wires,
> one to ground and the other two to the mainboard. I made 3 resistance
> measurements on those 3 wires from the prongs of the socket to the case
> (the ground) and from the two other connections on the mainboard.
>
> Resistance between the ground pin in socket and the ground pin on the
> PSU - 0 Ω.
> Resistance between the N(eutral) pin and the connection to the
> mainboard: 0 Ω.
> Resistance between the L(ive) pin and the connection to the mainboard:
> - Infinity Ω.
>
> Is it possible that something blew on that little circuit board? It has
> a piece of metal box that passes through the rear wall of the box and
> the socket has a plastic piece that is fixed to the rear wall of the box
> with 2 screws. I took them out and tried to pull off the plastic piece
> in an effort to get the circuit board out of the box to see if there are
> any damaged components on it. The I-0 switch is near the socket and is
> firmly glued into the little circuit board.
>
> Now back to your writings :-)
>
> Thanks for the help.

There is a review article for your PSU here.

http://www.hardwaresecrets.com/print...ply-Review/503

And a picture here.

http://www.hardwaresecrets.com/fullimage.php?image=8441

I think I'm seeing a black, heat-shrink wrapped fuse, on the
main board. The little board with the AC socket on it, could
contain a few filter components.

If we use this as a reference, the little filter board could
contain the stuff up to "C2" and "C3". The bridge
rectifier is more likely to be on the main board, but
I can't see it in the picture. The equivalent of C2
and C3, could be the blue capacitors on the tiny AC socket
board.

http://www.pavouk.org/hw/en_atxps.html

Then the question becomes, what series-connected components
are in the path. The Pavouk diagram has NTCR1, which is the
inrush limiter. I doubt they use something so crude on modern
supplies (mainly because they're a reliability issue).

*******

It's possible that any Active PFC circuit, may have the side
benefit of making it possible to limit inrush current. I've never
looked at a datasheet for an Active PFC, so can't really tell you
whether that is the case or not. An Active PFC circuit didn't
exist at the time the Pavouk site owner traced his power supply.

Hmmm. Well, this isn't going to do it. The Active PFC is a parallel
element.

http://www.hardwaresecrets.com/fullimage.php?image=5584

And according to the text description for that block, the resistor
near the right of the diagram, is the inrush limiter (equivalent
of NTCR1 in the Pavouk diagram). So the Active PFC itself isn't
moderating the current enough, and the circuit is still provided with
a negative temperature coefficient resistor (resistance drops when
it gets hot, after a couple seconds of operation).

http://www.hardwaresecrets.com/artic...Supplies/327/6

*******

If that is a fuse on the main board, then some series connected thing
opened on the socket board. Make sure, when you're doing the ohm
tests, that the switch on the back of the power supply is in the
ON position, as otherwise, the open thing could be the switch itself.

When it comes to "magnetics failures", such as T1 or T5 on the
Pavouk diagram, typically that happens because of cold solder
joints. Some magnetics, consist of enamel wire which has been
stripped chemically. And due to lax quality control, the
device doesn't get soldered into its PCB properly. I've had a
few items like that in my lifetime, that needed to be re-soldered.
But no failures inside a power supply. This was with other electronics.

Good luck and be careful,
Paul

Paul wrote:
> Seum wrote:
>> edfair wrote:
>>> Your cart wheels were inductors with ferrite cores.
>>>
>>> The metal stuff was heat sync to remove heat from the switching
>>> transistors that were mounted to them somewhere, probably near the
>>> base.
>>>
>>> The pop was probably one of the switching transistors. If so, it may,
>>> or may not have done any outside damage.
>>>
>>> Yeah, brushing against a loaded component can get your attention. I
>>> picked up a Xerox 820 supply with a blown bleeder and partially
>>> discharged the supply through the palm of my hand. Still hurts to
>>> remember it.
>>
>> Many thanks to Paul, L.M.& C, and Edfair. I'll send my first attempt
>> now and will read your emails today.
>>
>> This morning I took the cover off the PSU and made a few measurements
>> (for what they are worth. :-) )
>>
>> With the cover off I found a small 1.75" square circuit board just
>> inside the external socket. This little board has 3 connection wires,
>> one to ground and the other two to the mainboard. I made 3 resistance
>> measurements on those 3 wires from the prongs of the socket to the
>> case (the ground) and from the two other connections on the mainboard.
>>
>> Resistance between the ground pin in socket and the ground pin on the
>> PSU - 0 Ω.
>> Resistance between the N(eutral) pin and the connection to the
>> mainboard: 0 Ω.
>> Resistance between the L(ive) pin and the connection to the mainboard:
>> - Infinity Ω.
>>
>> Is it possible that something blew on that little circuit board? It
>> has a piece of metal box that passes through the rear wall of the box
>> and the socket has a plastic piece that is fixed to the rear wall of
>> the box
>> with 2 screws. I took them out and tried to pull off the plastic piece
>> in an effort to get the circuit board out of the box to see if there
>> are any damaged components on it. The I-0 switch is near the socket
>> and is firmly glued into the little circuit board.
>>
>> Now back to your writings :-)
>>
>> Thanks for the help.
>
> There is a review article for your PSU here.
>
> http://www.hardwaresecrets.com/print...ply-Review/503
>
>
> And a picture here.
>
> http://www.hardwaresecrets.com/fullimage.php?image=8441
>
> I think I'm seeing a black, heat-shrink wrapped fuse, on the
> main board. The little board with the AC socket on it, could
> contain a few filter components.
>
> If we use this as a reference, the little filter board could
> contain the stuff up to "C2" and "C3". The bridge
> rectifier is more likely to be on the main board, but
> I can't see it in the picture. The equivalent of C2
> and C3, could be the blue capacitors on the tiny AC socket
> board.
>
> http://www.pavouk.org/hw/en_atxps.html
>
> Then the question becomes, what series-connected components
> are in the path. The Pavouk diagram has NTCR1, which is the
> inrush limiter. I doubt they use something so crude on modern
> supplies (mainly because they're a reliability issue).
>
> *******
>
> It's possible that any Active PFC circuit, may have the side
> benefit of making it possible to limit inrush current. I've never
> looked at a datasheet for an Active PFC, so can't really tell you
> whether that is the case or not. An Active PFC circuit didn't
> exist at the time the Pavouk site owner traced his power supply.
>
> Hmmm. Well, this isn't going to do it. The Active PFC is a parallel
> element.
>
> http://www.hardwaresecrets.com/fullimage.php?image=5584
>
> And according to the text description for that block, the resistor
> near the right of the diagram, is the inrush limiter (equivalent
> of NTCR1 in the Pavouk diagram). So the Active PFC itself isn't
> moderating the current enough, and the circuit is still provided with
> a negative temperature coefficient resistor (resistance drops when
> it gets hot, after a couple seconds of operation).
>
> http://www.hardwaresecrets.com/artic...Supplies/327/6
>
>
> *******
>
> If that is a fuse on the main board, then some series connected thing
> opened on the socket board. Make sure, when you're doing the ohm
> tests, that the switch on the back of the power supply is in the
> ON position, as otherwise, the open thing could be the switch itself.
>
> When it comes to "magnetics failures", such as T1 or T5 on the
> Pavouk diagram, typically that happens because of cold solder
> joints. Some magnetics, consist of enamel wire which has been
> stripped chemically. And due to lax quality control, the
> device doesn't get soldered into its PCB properly. I've had a
> few items like that in my lifetime, that needed to be re-soldered.
> But no failures inside a power supply. This was with other electronics.
>
> Good luck and be careful,
> Paul

Thaaaank you Paul. You are truly a mine of information :-)