Continue to Site

Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

[SOLVED] PC power supply short circuit

Status
Not open for further replies.
What I want to know for now is how each output is handled?

I mean I used to see rectifying diodes after the high frequency transformer, now I don't see anything!

20170817_063943.jpg

- - - Updated - - -

I opened the cooler master webpage for the 650V semi modular and found an interesting information.

20170817_063943.jpg

So I only have one line after the high frequency transformer, because I saw the pins after the transformer are shorted and only one ground, that's my guess for now.

But why two coils?
 
Last edited:

Hi,

If those are "custom DC/DC modules" then isn't it obvious?
Two output voltages --> two DC/ DC converters --> two coils

Klaus
 

Yes two coils for 3.3 and 5 volts.

But going back to the two coils near the high frequency transformer and the two after before the "custom DC/DC modules", because these modules come at the final stage just before 3.3V and 5V.

So, there's only one main non-rectified voltage after the high frequency transformer, then this line is distributed to two lines. One to feed the "custom DC/DC modules" and the one which is the biggest one I guess is the 12V which doesn't have a "custom DC/DC module", I think it goes to the 4x power MOSFETs near to the "big yellow taped transformer".

And that's why the 12V is rated at 54A.


650v amp ratings.png


But what also I don't understand is how all these lines are rectified? I think the many capacitors and also the aluminum capacitors are the reason, so much more capacitors are the substitute of rectifying diodes.

Another question, why the total power is summed at 650W where the ratings are 120, 648, 3.6, 12.5 watts which is 784.1W?
 
Last edited:

Hi,

The coils have been discussed before:
* input filter
* PFC
* high power path
* standby power path
* some smaller maybe for just for reducing ripple voltage (LC filter)

Why the power is not added up:
Because some depend on each other and some don't. And it depends where in the circuit the limiting happens.

Klaus
 

Is my guess right that there's no diodes and they are using caps instead?

If that's true then I should takeout the caps, but the soldering material is very strong and I only have 30W soldering iron.
 

Is my guess right that there's no diodes and they are using caps instead?
No, that wouldn't work. I think the "custom DC-DC module" is just advertising hype for "a coil on a small PCB to save space" as the name implies it is already DC entering it and it is therefore already rectified by diode(s) and filtered with capacitors.

The power rating is confusing but what it tells you is the individual rating of each output when used alone. The 650W limit is the total the power supply can manage across all the loads, meaning they can't all be run at full load at the same time.

Did you try the resistance tests I mentioned in post #20 yet ?

Brian.
 

No, that wouldn't work. I think the "custom DC-DC module" is just advertising hype for "a coil on a small PCB to save space" as the name implies it is already DC entering it and it is therefore already rectified by diode(s) and filtered with capacitors.

The power rating is confusing but what it tells you is the individual rating of each output when used alone. The 650W limit is the total the power supply can manage across all the loads, meaning they can't all be run at full load at the same time.

Did you try the resistance tests I mentioned in post #20 yet ?

Brian.

Yes but not entirely, there are many SMD low ohms resistors, but there are ones with numbers; like, 417 and reads low resistance. But, obviously that because of the short which affects the resistors, capacitors and the MOSFETs.

What I want to do now is to take those aluminum caps, and go from there by taking out components until I find the faulty one. But the soldering is so strong. I would give it a try now and see what can I do or I should go today and buy 60W.

What you suggest to me 100 or 60W for this work? I consider the soldering here is so strong, it doesn't melt so easily.
 

I use a 45W soldering iron without problems.

Please do those resistance checks - all I need is the resistance from the +12V to 0V in both directions. I appreciate you have done the buzzer check but you must understand that the buzzer makes sound not only for a complete short but for any low resistance. Try probing a good 10 Ohm resistor for example and it will say it is shorted. In the 12V output it is normal to find low values so the buzzer may be sounding when there is no fault there. A resistance test is more accurate, it gives an actual reading of the components and will tell if a short is really there or you are seeing normal operation.

I strongly suspect you are looking in the wrong place and the fault lies in the switching circuit.

Brian.
 

I agree with Brian. You may be looking in the wrong place, and you're certainly using the wrong tool (buzzer) to debug this. If you don't have an oscilloscope, I suggest maybe you should just buy a new power supply.

'If your only tool is a hammer, all your problems are nails.'
 

Hello, good to come back after all this time.

I see my thread as closed. Can I continue posting here or I better open a new thread?

Thank you,
 

Yes, I have done them this morning.

I did the resistance test between the+12 and GND and it's 0.1 ohms as it really like shorting the DMM probes.

2017-09-25 07.42.35.jpg

OK, I succeeded to take out this capacitor this morning, thanks to my new 54W soldering iron :)

So, I thought I would cut the path to the rest of the +12V rail but of course the caps are connected in parallel, so where can I cut the way from the high frequency transformer, I think this way I can determine whether that fault is before or after the high frequency transformer.

So, what should I do now ? Take out the high frequency transformer ? Is there another less pins to desolder component ? Like one of the diodes near to high frequency transformer ?

Regards,
 

You are clearly not understanding how these power supplies work. The high voltage side is completely isolated from the low voltage side - it has to be to make the PSU safe.

If you are really measuring 0.1 Ohm between +12V and GND (yellow and black output wires) it can only be a fault on the low voltage side.
Almost all ATX supplies use nothing more than a single or dual rectifier and one or two electrolytic capacitors connected by an inductor. The inductor is VERY unlikely to be faulty and even if it was, it could not cause a low resistance reading. That only leaves the two capacitors or the diode(s). Obviously I can't be sure from the photograph but my best guess would be one of the capacitors under the black sleeve on the heatsink fin is responsible, if not those, my second guess would be the devices on the heatsink itself. There is a remote, unlikely chance that the cooling fan is shorted if it is connected across the 12V output but in most cases it isn't.

I would say in >90% of ATX failures, the culprit is one of the capacitors just below the IC in the photograph but that wouldn't account for the short you are seeing.

Brian.
 

OK, I took all the caps on the way.

But I really don't know what is the shape of the path? Which lines I'm dealing with.

I saw a new reply on the email, but I couldn't see it in the thread, maybe after I post this reply it would appear.

It says:
davoud has just replied to a thread you have subscribed to entitled - PC power supply short circuit - in the Power Electronics forum of Forum for Electronics.

This thread is located at:
https://www.edaboard.com/showthread.php?t=370386&goto=newpost

Here is the message that has just been posted:
***************
As friends said,
You must first check the double-diodes. Double-diodes packet are similar to the MOSFETs that you shown or slightly larger. I guess one of diodes is burnt.
Double diodes usually mounted to a heat sync. please check components mounted to heat sync at the corner of photo.Attachment 141456 (https://www.edaboard.com/attachment.php?attachmentid=141456)
***************

Where are these diodes precisely?

20170926_103202 (2).jpg
 

Changing parts randomly will get you nowhere.

I already told you in post #33:
my best guess would be one of the capacitors under the black sleeve on the heatsink fin is responsible, if not those, my second guess would be the devices on the heatsink itself
.
The diodes are the tab devices fixed to the heatsink with the black sleeve on one of it's fins. Follow the yellow wires to the capacitor the either directly to one of the diodes on the heatsink or through a coil first then the diode.

Brian.
 

Those are not diodes, those are power MOSFETS. Their numbers are 023N04N.

Their datasheet info:
Features
• MOSFET for ORing and Uninterruptible Power Supply
• Qualified according to JEDEC1) for target applications
• N-channel
• Normal level
• Ultra-low on-resistance R DS(on)
• 100% Avalanche tested
• Pb-free plating; RoHS compliant



Another test, I discovered:
1. The outputs of the high frequency transformer are all shorted together and to the GND.
2. The purple +5V rail is also shorted to the GND.
3. Those 3-pin devices attached to the heat sync are power transistors 023N04N.


20170926_140723 (2).jpg
 

I find it REALLY hard to believe that you have as many "shorts" as you claim. You are insisting on trying to debug a relatively complex circuit with a primitive continuity tester and random part removal. This is a fool's errand. I think you need to get a schematic, an oscilloscope, an understanding of switching power supplies or just cut your losses and buy another supply. If your goal is to learn about power supplies you're going about it the wrong way.
 

I know I need a schematic and an oscilloscope, but I can't find the schematic online if you can help me in this problem. And I don't think I need an oscilloscope, at least for now.

Your third point is "an understanding of switching power supplies", well SMPS units are a lot of models and this one I have is one of the new models and complex to me. And learning to me is not a matter of watching videos and search for documents about SMPS, I need to investigate the problem in the practical aspects and I need to try and try until I learn and find the problem.

I heard from a doctor in a US university who teaches electronics on YouTube, that you can't be an engineer that easy! You may destroy a lot of devices until you learn something.

So, learning by looking into documents is not the way to go, I have to practice and try.

Because, before I didn't know anything about power supplies, and how would I know at least something! Of course by looking into the unit and learn little theoretical information and compare that to what I have.


One thing for sure I know, is that I won't be good in troubleshooting that easy, I have to learn by doing in practice and learning in theory.

And this unit is not a simple electronic circuit, it's a complex system even if the problem is easy for other people to determine and fix, but it's difficult to me.

I know learned a lot of things in electronics and embedded systems, but fixing problems is difficult to me, but I fixed a lot of problems in programming microcontrollers which are basically easy and beginners problems but I learned by posting online and watching videos and reading articles and documents.

But fixing a modern PSU is hard to me, it's really not that easy.

If I'm loosing this PSU, then it's OK not to throw it and try to learn something from it. That's my goal.
 

All switch mode PSUs are relatively complicated but they still follow the basic rules of electronics. One short circuit is unusual but now you claim there are two and in different parts of the circuit, either you are extremely unlucky or measuring things wrongly.

If you look at the schematic for any ATX PSU you will see the low voltage side is almost identical across all models, the main differences are on the high voltage side.

The purple +5V output is the standby supply and it runs quite independently from the rest of the circuit so for it also to be shorted is very strange. That +5V is a low power feed to keep the clock running, provide power for the standby switch and sometimes to keep USB powered up but it isn't derived from the main PSU transformer. By necessity, it has to be there whenever the AC is connected or there would be no way the remainder of the PSU could be switched on.

I'm still thinking you are looking in entirely the wrong place for the fault and your buzzer/meter is giving you misleading readings.

The secondary pins of the transformer should measure as shorts to ground, the resistance of a few turns of thick coper wires inside it is almost zero. Between there and the +12V yellow wires is one or maybe two diodes (sometimes two in one three-pin package) so if there really is a short there, it can only be the diode(s).

Bearing in mind that the black wires are ground, see if you can trace the circuit from the yellow wires back to the transformer and post if for us to see.

Brian.
 

I know I need a schematic and an oscilloscope, but I can't find the schematic online if you can help me in this problem. And I don't think I need an oscilloscope, at least for now.

Your third point is "an understanding of switching power supplies", well SMPS units are a lot of models and this one I have is one of the new models and complex to me. And learning to me is not a matter of watching videos and search for documents about SMPS, I need to investigate the problem in the practical aspects and I need to try and try until I learn and find the problem.

I heard from a doctor in a US university who teaches electronics on YouTube, that you can't be an engineer that easy! You may destroy a lot of devices until you learn something.

So, learning by looking into documents is not the way to go, I have to practice and try.

Because, before I didn't know anything about power supplies, and how would I know at least something! Of course by looking into the unit and learn little theoretical information and compare that to what I have.


One thing for sure I know, is that I won't be good in troubleshooting that easy, I have to learn by doing in practice and learning in theory.

And this unit is not a simple electronic circuit, it's a complex system even if the problem is easy for other people to determine and fix, but it's difficult to me.

I know learned a lot of things in electronics and embedded systems, but fixing problems is difficult to me, but I fixed a lot of problems in programming microcontrollers which are basically easy and beginners problems but I learned by posting online and watching videos and reading articles and documents.

But fixing a modern PSU is hard to me, it's really not that easy.

If I'm loosing this PSU, then it's OK not to throw it and try to learn something from it. That's my goal.

I don't think you're going to learn much without knowing what you're fixing. You're just randomly replacing parts without even understanding the function of the part. If you replace a part and your PSu starts working, what have you learned?
 

Status
Not open for further replies.

Similar threads

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top