[SOLVED] PC power supply short circuit

Hello,

My PC power supply just failed couple days ago and didn't know that until I asked people that my PC just stopped running, I thought it might be the CPU but they told me it's probably the PSU.

So I opened the PSU and checked the rails and found that the 12V rail is short circuit with the ground.

Now how to fix this problem?

Maybe I would learn an easy way to troubleshoot rails short circuits.

Thank you,

Is it an actual short circuit, or just zero output? If it IS a short, then it might be a bad output capacitor or a shorted transformer winding. But I don't really think either of those are very likely. I'm not really sure what the output resistance of a PC supply should be, so you might be making a false assumption here.

I don't know, now after I de-soldered one of the transistors, the rest of the transistors are shorted with the ground.

I'm really now confused about this complicated PCB. There are a lot of SMDs.

The only still problem I think is the short 12V rail. I don't know how to discover the faulty device. There are shorted diodes, SMD resistors, sides of output coils are shorted too, all I'm testing with DMM buzzer.

I get confused, I leave it and check it again.

i should get back to complete the testing after sometime.

Pretty much any coil is going to look like a "short" with a DMM buzzer, that's not a valid test. I think you're going to have to look at things with a scope, there's not not a lot you can tell with just a continuity tester.

Hi,

I agree with Barry.
Your DMM buzzer sends out a DC signal (plus and minus) it checks current flow.
If there is enough current flow, then the bzzer goes ON.
But the threshold is not defined.

With a PC power supply - when flyback type - the 12V output consists at least:
* the secondary winding of the transformer
* the (catch) diode
* the capacitor.

Your testing:
* The secondary winding is very low ohmic. It surely will cause your buzzer to sound
* The diode is low ohmic in one direction, high ohmic in the other direction. It is very likely that with an unconnected diode your buzzer sounds when you connect your DMM in one direction, but it will be quiet in the other direction.
* The capacitor when connected in the correct direction is low ohmic first ... then it becomes charged and thus the current decreases. It is very likely that you hear a short buzz immediately when you connect the DMM, but it will stop buzzing automatically. Connecting the capacitor in the wrong direction will cause the buzzer to sound, but maybe it sounds continously or it will stop.

Now all three components (and maybe others) are connected...this makes it less predictable if the buzzer sounds or not.

Klaus

Of course any coil would measure very low resistance between its ends.

What I mean is each end of the coils is shorted with the ground.

How I measure in this situation, is by connecting one end of the DMM to the ground and the second end to any point of the PCB, I move to pins of transistors and some other devices. But other components; like, resistors (which won't look like a very low ones), diodes. I measure these by connecting the DMM between its ends to see if the diode is shorted or the resistor is burnt.

Because the testing is different.

Here's a screenshot for my PSU PCB.

Note: The red circles are shorted diodes, they certainly are broken diodes.

All this area is a mess everything is shorted.



- - - Updated - - -

KlausST said:

Hi,

I agree with Barry.
Your DMM buzzer sends out a DC signal (plus and minus) it checks current flow.
If there is enough current flow, then the bzzer goes ON.
But the threshold is not defined.

With a PC power supply - when flyback type - the 12V output consists at least:
* the secondary winding of the transformer
* the (catch) diode
* the capacitor.

Your testing:
* The secondary winding is very low ohmic. It surely will cause your buzzer to sound
* The diode is low ohmic in one direction, high ohmic in the other direction. It is very likely that with an unconnected diode your buzzer sounds when you connect your DMM in one direction, but it will be quiet in the other direction.
* The capacitor when connected in the correct direction is low ohmic first ... then it becomes charged and thus the current decreases. It is very likely that you hear a short buzz immediately when you connect the DMM, but it will stop buzzing automatically. Connecting the capacitor in the wrong direction will cause the buzzer to sound, but maybe it sounds continously or it will stop.

Now all three components (and maybe others) are connected...this makes it less predictable if the buzzer sounds or not.

Klaus

Click to expand...

I'm taking into consideration the polarity of the diode and to be sure I tested them again, they are shorted in both sides.

I know when a capacitor is holding a charge and causing the buzzer to be on for a while then when the charge diminish the sound stops if it's a healthy component. And if it's a faulty one then the sound continues.

I'm not testing a winding between its pins or the two ends of a coil, of course they are low resistance and cause the buzzer to be ON, this test is done to check the windings are not broken.

I'm testing each end with the ground. Because I do this kind of testing to know if anything is shorted with ground.


My conclusion here is that my PSU is gone :-?

Replacing these components would be very hard and I have to buy all the faulty parts from China and also I have to buy a heat gun soldering station to replace the SMD.

I have also my work Dell PSUs, I took them from work because they are faulty and they don't want them anymore, so I took 4 of them so I can learn how to troubleshoot modern PSUs but I didn't open any of them. I just opened mine which is Cooler Master V650 +80 Gold PSU. But it's now broken, I would really like to fix it and I'm thinking now how to do that and what parts and equipment I need.

Hi,

Difficult to help without the schematic.

Maybe replace:
* capacitor
* diode

Klaus

What's that big blob of solder in the upper left corner? Actually, there are two. That doesn't look right. Neither does that big blog in the center of the board. Maybe those shorts don't matter, but it sure indicates terrible manufacturing.

How to get the schematic? I tried but couldn't.

Maybe I should contact Cooler Master but I think they won't give me.

- - - Updated - - -

barry said:

What's that big blob of solder in the upper left corner? Actually, there are two. That doesn't look right. Neither does that big blog in the center of the board. Maybe those shorts don't matter, but it sure indicates terrible manufacturing.

Click to expand...

Upper left corner, you mean those lines of solder? I think those are common GND >> I checked now, yes those lines are the common GND for the HIGH voltage side of the SMPS.

The one in the center is also a common GND for the low voltage side of the SMPS.

The high side should be healthy, but I think the problem would be on the 12V rail because the other rails are not shorted. I think if I remove the small resistors and diodes near the 12V rail, then they would be OK and the shorts would be fixed I guess.

I could be wrong but I'm very suspicious that so many things seems shorted, especially when checking with a buzzer.
The big rectifier diodes could well be dual diodes with both anodes/cathodes going to windings on the transformer so they would appear to be shorted in-circuit but are probably OK if removed.

Low voltage side faults are quite rare on PC PSUs with the exception of dried up electrolytic capacitors but they usually fail open circuit not shorted. It is also fairly rare to take feedback from the 12V line to regulate the input stage, usually it is done from the 5V or 3.3V outputs which are more critical. That only leaves the rectifiers and maybe two capacitors (one each side of the choke) as possible culprits.

Brian.

betwixt said:

I could be wrong but I'm very suspicious that so many things seems shorted, especially when checking with a buzzer.
The big rectifier diodes could well be dual diodes with both anodes/cathodes going to windings on the transformer so they would appear to be shorted in-circuit but are probably OK if removed.

Low voltage side faults are quite rare on PC PSUs with the exception of dried up electrolytic capacitors but they usually fail open circuit not shorted. It is also fairly rare to take feedback from the 12V line to regulate the input stage, usually it is done from the 5V or 3.3V outputs which are more critical. That only leaves the rectifiers and maybe two capacitors (one each side of the choke) as possible culprits.

Brian.

Click to expand...

The rectifier bridge is OK and doesn't have any shorts, I have a big assumption that the HIGH voltage side is OK.

The problem is obvious on the output, because 5V and 3.3V are not shorted, 12V is shorted which means there's something faulty in that region.

But all the area in the 12V output is shorted, there are SMDs and everything is shorted so I have to take them out and check which one is burnt. But I'm afraid of how to take out the SMDs because I don't have a heat gun soldering station.

Maybe in our local electronics shops, one told me that there's a shop which wants to sell its equipment as a clearance I don't know I want to check later, but buying one from Aliexpress would cost me like $200+ so I have to think and try to remove them by my 30W soldering iron which I'm sure won't work effectively.

I fix a lot of faulty power supplies, The most common fault is leaking capacitors (look for ones with bulges) The second most common fault is the large rectifiers on the output of the power supply going short circuit. It is normally the 5V rectifier that fails and normally only when the fan stops working.

The rectifier bridge is OK and doesn't have any shorts, I have a big assumption that the HIGH voltage side is OK.

Click to expand...

That's not what I meant. The 12V rectifier is often in a TO-220 package but has TWO diodes inside it. The outside pins go to a center tapped transformer winding and the rectified output comes from the middle pin. If you measure across it you will probably see the resistance of the transformer, not the diodes.

Regardless of that, the 12V output of a PC supply is normally nothing more than the rectifier, a capacitor to ground, a filter choke and another capacitor to ground. The chances of a choke failing are almost zero so it points the accusing finger at the rectifier or the capacitors. I think the chances of any SMD component being faulty are very slim, especially as the power available at the 12V output would probably vaporize them. Most likely, the SMD components are to control the cooling fan speed but they would not cause a short as you describe it.

Note that your buzzer is not an accurate way of testing for shorts as it will 'buzz' whenever enough current flows between the probes and the DMM can almost certainly push enough current through the kind of diodes used in SMPS to make them look like shorts when they are actually perfectly OK.

Brian.

Hi,

What happens if you disconnect/desolder the diode/rectifier? Does the rest operate correctely then?
(edit: deleted schematic discussion, because photo is already there)

Klaus

Hi all,

Thank you all very much for the replies

Here are couple photos I took and described some information I know and some I don't know.

I want to know where are the output diodes, the high frequency transformer, there's another transformer so I don't know what are they!

Hi,

The big yellow taped device should be the power transformer.
I assume what you named "o23N04" are the diodes.
But I'm not sure. Difficult to recognize..

Klaus

Eagle1109, I'm not sure I agree with all your component identification but the critical parts are off the edge of the photograph (I think!).

If you are sure the fault is on the +12V output, show us a top and bottom side photograph of the board showing the edge where the output wires are attached. The high voltage side isn't particularly important, what we need to identify are the capacitors connected to the 12V output wires. The yellow transformers are debatable without further pictures as that kind of PSU will have a similar component for PFC and also for the 'always on' 5V standby supply.

There should be two 12V output points, the positive normally has yellow wires and the negative normally has blue wires. If you can mark where they attach on the underside photograph it would help us.

Brian.

KlausST said:

Hi,

The big yellow taped device should be the power transformer.
I assume what you named "o23N04" are the diodes.
But I'm not sure. Difficult to recognize..

Klaus

Click to expand...

023N04 are power MOSFETs

But the transformers, I'm trying now to understand what I can understand on the PCB and record my description on a video and upload it to YouTube.

Hello,

Finally I uploaded the video

It's difficult to say for sure without seeing the PSU but I think you have some parts identified wrongly.

The MOSFET and 2-pin power diode behind the large black capacitor (as in the image in the post) are for PFC and the yellow inductor under your finger is also part of the PFC circuit.

The Two MOSFETS on the other side of the heat sink are the ones driving the output transformer through the two small yellow wires and I think you will find the mounting pins are also connections to the transformer. The wires do not have to be very thick, remember that power = voltage * current and at that part of the circuit the voltage is very high so the current, and therefore wire thickness is small. The high current parts are after that transformer where the voltages are much lower. The other inductor, nearer the middle of the board is almost certainly a smaller independent supply to keep the "+5V standby" supply running as it has to be present even when the computer is switched off.

The test with the buzzer is not valid and may be leading you to the wrong conclusion about there being a short on the +12V line. Instead of using the buzzer range, use the lowest resistance range on your DMM and if the reading is more than about 10 Ohms you are looking for the fault in the wrong place. Apply the probes and leave them in place for a few seconds then note the DMM reading, then reverse the probes and wait again for a few seconds and tell us both the readings.

I think its far more likely the problem is one of the two capacitors beside the IC (near R14) has gone open circuit preventing the PSU starting up when turned on.

Brian.