Broken PC Power Supply

[Spekkio]

Yesterday when I woke up, my computer was shut off. I usually keep my computer on most of the time. *sigh* made me so angry at the time.
Because I've also had problems with my Neato XV Signature Pro robotic vacuum cleaner, that I have now left in for repair.
Felt like EVERYTHING was failing and breaking at the same time, even if it was only two things.

I just had a hunch that It must be a broken power supply.
I just bought a new power supply and my computer works fine now. And I'm not really interested in repairing the old one.
I might scavenge some parts from it if they work. Because I have been thinking of building my own switched power supply as a project.
But I have some questions in the end of this post about why it might have broken like this.

I was hoping that I got a fun thing to troubleshoot and look around, measuring signals and things like that.
I glanced over all the capacitors, they looked fine. No leaking or swollen caps.
I've heard the NTC Thermistor is a thing that can break and cause problems? I have not really checked if it is OK yet.
The fuse is OK, strangely enough.
And then the error was pretty boring from a troubleshooting perspective, it was quite easy to spot.
A black area on the inside of the power supply chassis, close to that I found this completely blown IC

Sorry for the very large picture.



And I also found this little bit of IC falling out.



Guessing from this picture, it is a TNY278 "Off-Line switcher"

Data-sheet: http://pdf1.alldatasheet.com/datasheet-pdf/view/139812/POWERINT/TNY278PN.html


I wonder what could have caused this to happen?

The power supply that I had was at 600W. I wonder if it might have been to weak to power my computer.

I do have a pretty good graphics card my my computer. An Asus GeForce GTX 780 Ti (It was new in 2013 I think).
Three harddrives, and 5 Fans. A pretty old sound-card, Creative Audigy with one of those front panels with lots of outputs and inputs, MIDI and optical etc.
A Intel(R) Xeon(R) CPU E31240 @ 3.30GHz CPU, that runs pretty cool
And ECC memory.

coretemp-isa-0000
Adapter: ISA adapter
Physical id 0:  +38.0°C  (high = +74.0°C, crit = +94.0°C)
Core 0:         +33.0°C  (high = +74.0°C, crit = +94.0°C)
Core 1:         +34.0°C  (high = +74.0°C, crit = +94.0°C)
Core 2:         +37.0°C  (high = +74.0°C, crit = +94.0°C)
Core 3:         +39.0°C  (high = +74.0°C, crit = +94.0°C)

I bought a new power supply now at 850W. A little bit more expensive, from EVGA, fancy luxury stuff, lol.
Hopefully of better quality, and enough to power my computer.

But I wonder, what does the "off-line switcher" really do?
What functionality does it have?
What could have caused this to just blow up like this? It must have happened over night. So I never noticed it.
The only thing that looks broken is this blown IC from what I can see.

[mariush]

The chip takes care of "5v stand by", the chip along with the small diodes and the very small transformer near it are always on, independent of the rest of the power supply. 
The power supply outputs 5v at up to 2-2.5A and the chip could be damaged if there's excessive voltage coming from mains, or from overheating or some other reasons. 

This 5v stand by voltage is always on and going into the motherboard and from there into the chipset and bios because it makes things like "wake computer from lan" or "turn pc on when pressing a key" possible. Also, without it, the motherboard can't send the signal to start the pc to the power supply, when you push the "power on" button.

As for the fix, check the diodes near the chip for shorts or open, check resistors, replace the diodes where needed, then replace the chip and maybe also check the fuse which could be tripped/broken.

The rest of the power supply may still be as good as new, so I wouldn't throw away the power supply in the trash.

Oh, and if you're really curious, here's how a chip similar to yours  (TNY264) looks inside ... as you can see lots of stuff there but the biggest area is taken by a big mosfet :



image courtesy of zeptobars.ru , they decap lots of chips and explain their insides from time to time : http://zeptobars.ru/en/

[wraper]

This TNY278 is just for +5V standby power supply. I'd say just check a few parts around it (diodes and electrolytic cap first of all) and replace the IC itself.

[Spekkio]

the chip could be damaged if there's excessive voltage coming from mains

Just a stupid thought. Could my ESD mat on my desktop actually cause this? A large voltage discharge for some reason.
The mat is connected to ground with an 1MOhm resistor between mat and ground.
There is no ESD protective mat on the floor where the computer stands, it stands on on some plastic feet.
The computer is connected to the same ground point as the mat.

Maybe a little bit far fetched.

Otherwise I suppose it can be overheating. There has been some really warm days lately.

And my computer might have needed more than 600W with that graphics card.

EDIT:
Another thing I thought of, was that I have accidentally shorted the D+ and D- pins, or maybe also other pins on a USB connector a few times because I forgot to unplug it once when soldering on a prototype board with the USB cable connected to it and the computer.
nothing happened to the computer at the time, didn't shut down or anything.

[MagicSmoker]

The TinySwitch, TopSwitch and similar ICs from Power Integrations are incredibly easy to use but have a reputation for being unreliable.

[Spekkio]

Still, I am a bit curious if my stupid idea about my desktop ESD mat could cause something like this. I bought it recently and this happened now. There is no ground connection in my wall outlet. The ESD mat is connected to the same ground point as the computer in the power strip. Is this bad, I just have this hunch that is it wrong and stupid of me?

From what I have learnt. Having ground connected deep down in earth would make a good zero potential voltage. Just the power strip ground is not much.

My first idea was to connect the ground for the mat to my radiator. But I wasn't sure. That is not the same ground as the computer or my measuring equipment that I have plugged in. Since there is no ground pin in the wall outlet. In my mind my inituiton says that it should be connected to the radiator that most possibly will have a zero potential and suck up all the charges, but there should still be a connection between computer/equipment ground and esd mat ground?

There is still however 1MOhm between the mat and ground. So I seems very unplausible for me that it have caused the IC to blow up, or maybe even impossible. But still something that I can't get out of my mind.

I feel stupid asking this. I should know

Edit:
I found a thread about how to correctly install esd mats. I am reading it now. I think Dave had a video about it. I've forgotten...

But still, please if you have any anwers to my questions here. Simply, could my ESD mat setup have caused my PC power supply to fail?

Skickat från min GT-I9305N via Tapatalk

[MagicSmoker]

...From what I have learnt. Having ground connected deep down in earth would make a good zero potential voltage. Just the power strip ground is not much.

Ground isn't ground without a connection to the ground (i.e. - a rod - or several - in the earth)... ahem.

Simply, could my ESD mat setup have caused my PC power supply to fail?...

Anything is possible, but this is so unlikely it is safe to say the answer is no.

[Spekkio]

Ground isn't ground without a connection to the ground (i.e. - a rod - or several - in the earth)... ahem.

So, In any case I still need a better connection to ground

[tautech]

This TNY278 is just for +5V standby power supply. I'd say just check a few parts around it (diodes and electrolytic cap first of all) and replace the IC itself.

+1
Forget about any ESD issues, they won't be the cause.

[Spekkio]

Thank you all for the answers.
It gave me some new insight. I'll take a closer look at the components nearby. I'll probably use working parts from it for building my own powersupply.

I wasn't certain my esd mat was correctly installed. I think I'll write another thread about it. There are some things around that, that I feel I need to get in good order.

Skickat från min GT-I9305N via Tapatalk

[Kleinstein]

ESD unlikly the reason for blowing the supply. It's more likely a spike from the grid - this could have been a thunderstorm a few weeks ago, or just from a short somewhere in the neighbourhood.

Usually repair on PC supply is difficult as there are many parts that can be wrong as well. Also prices are usually moderate to low and a bad behaving supply can do quite some damage. So unless its an obvious error and the PC is rather old anyway, I would buy a new supply.

So It's at least two chips burnt. First thing would be to identify the chip on the board and check for other damage (diodes, caps). If you can't identily U501 chances are slim.

[wraper]

ESD unlikly the reason for blowing the supply. It's more likely a spike from the grid - this could have been a thunderstorm a few weeks ago, or just from a short somewhere in the neighbourhood.

Usually repair on PC supply is difficult as there are many parts that can be wrong as well. Also prices are usually moderate to low and a bad behaving supply can do quite some damage. So unless its an obvious error and the PC is rather old anyway, I would buy a new supply.

So It's at least two chips burnt. First thing would be to identify the chip on the board and check for other damage (diodes, caps). If you can't identily U501 chances are slim.

What two chips? At the moment there is only IC obvious to be burnt and already identified  TNY278P (U501). And Most likely it is faulty only by itself or together with some surrounding component.