Help ! My PC died suddenly. Could be the PSU, I guess.

:help:
Any ideas of what failed and / or any suggestions how to carry out troubleshooting / faultfinding will be appreciated.
This is going to be a real challenge for me.
It is so much easier to assemble a new rig then to try to locate and fix the fault in the existing system when it dies.
Please, see My System for the hardware details.

1. What happened
Last Sunday the machine worked as usual.
I did not do anything exciting or unusual with it.
I checked email, did some easy work on spreadsheets.
I switched the machine off via Windows in the usual way.
On Monday morning I wanted to check the weather forecast so I went and pressed the machine's main switch button.
Nothing happened. :shocked2:
I tried again.
The machine remained silent. :sleep:

2. Other symptoms
Only yesterday (Tuesday) evening I found few spare minutes to examine the PC in more detail.
This did not reveal much.
The "start-up" sequence runs as follows.
With the PSU switched on the mobo's LED is alight. A good sign, I suppose.
While, and soon after, pressing the ON button the following happens (it all lasts for c. 1s) :
- most (probably all) of the fans (system's via fan controller, CPU's) make a quick starting move and then they stop,
- several LEDs (on the system fans and within the fan controller) make a quick flash, NO RAM LEDs alight.
Then the system remains silent and the mobo LED goes off.
The boot up sequence does not initiate at all.

3. Recent upgrades
Only few upgrades were performed since the initial assembly and until the GPU upgrade in March this year (so c. 2 months ago).
These included added sound card and upgraded RAM in terms of both the size and speed.
Then in March I replaced the old Asus EN6600GT-TD 128MB with XFX GF 8800 GTS 640MB.
I was a bit concerned if my 600W PSU could cope with the extra load but a brief analysis suggested that it all should work fine.
And it did.
Until yesterday.

4. My best guess
By elimination, I think, the hard and optical drives must be OK as well as most other peripherals (e.g. the sound card).
If only they were faulty, the boot up sequence should initiate, I think.
Graphic card might be at fault but I would hope that the system should "kick off", possibly without producing any picture.
I feel that the PSU (Enermax Noisetaker EG701AX-VE 600W) or the mobo (Asus P5AD2-E Premium 925XE) are faulty.
I guess, either there is a short circuit somewhere or the PSU gave up with delay after the GPU upgrade.
I suspect that some sort of protection circuit in the PSU failed and it switches the PSU down soon after the load is applied.
These are all guesses.

5. My initial plan of action
I would start with testing the PSU but I am not certain how to do that most efficiently.
Testing not loaded PSUs may not reveal any faults, I suppose.
Any advice on testing PSUs, ideally without disassembling the entire rig, will be appreciated.
I do have a multimeter.
If it is the GPU that causes the overload then it might be worth putting the old 6600 (which I retained) and testing the system then.
Any other ideas will be appreciated. :secret:


Regards

Robert

Have you checked and removed all optical drives or anything else which is powered by the system which wouldnt be needed for a boot up?

So basically remove everything apart from the CPU, 1 Stick of RAM and the GFX card in - make sure all the connectors are flush to the board so on. And try and boot it then.

It does sound like a short of some sort if it powers up and drops right after, but start with the basics and strip it all out and start from scratch.

I would

1) Double check your CPU Coolers fan is connected and able to spin freely

2) Disconnect all non neccesary components from the motherboard and PSU i.e. hard drives / optical drives / case fans / PCI cards / USB devices etc. & try again.

3) Clear CMOS & try again

4) Swap out graphics card & try again

5) Try with one memory stick and cycle through sticks until you've tried them all individually.

6) Borrow a half decent PSU - no need for anything more than 350w with everything discconected and a 6600GT and try that.

7) If none of the above helps it's probably your motherboard.

Thanks Recoil.

So far I had no time to do anything about it.
I am trying to get the strategy right to avoid spending too much money on new hardware and spending too much time on repairing. If it is the failed PSU then getting a new one, well, would not be a disaster. If it is something else, especially the mobo, then I will be forced to build a new rig, i.e. mobo / CPU / PSU would need to be all replaced and this is going to be time consuming and expensive.

It sounds like I have to get back to basics and nearly strip off the entire rig. Wow. That is the last thing I want to do right now but it seems to be unavoidable.

Quote:

---

Originally Posted by rmaciag

3. Recent upgrades
Only few upgrades were performed since the initial assembly and until the GPU upgrade in March this year (so c. 2 months ago).
These included added sound card and upgraded RAM in terms of both the size and speed.
Then in March I replaced the old Asus EN6600GT-TD 128MB with XFX GF 8800 GTS 640MB.
I was a bit concerned if my 600W PSU could cope with the extra load but a brief analysis suggested that it all should work fine.
And it did.
Until yesterday.

---

Any decent 600W PSU should cope with that spec. I'm not a fan of the Enermax series (other than the Liberty which seems decent), but it should be fine. If it couldn't I certainly would have expected instability (particularly under 3D load) long before the PSU damaged itself or another component.

Do you have a case speaker? Is it hooked up? Any beeps would be very helpful. That there is some power indicates that there is hope yet, but there's obviously a problem. Cross your fingers it's not the motherboard, as a replacement may render those RAID arrays irrecoverable.

its generally the only way to fault find unless you got another full rig - which you can swap parts with - which most people dont...

Borrow another PSU if you can, that's what my money is on.

re: Case Speaker
No, I haven't got one.
Intially I did not include one in my spec, then I searched for a basic one and I could not find any decent one locally.
At the end I forgot about it.

re: RAID0 arrays
Yes, it is my main worry at the moment.
I do not mind how long the repair takes but it would be worth recovering the HDs.
Not that there is something critical on them but still there are some files that are worth recovering that have not been backed up recently.
Plus the whole setup was a results of days of improvements.
Any advice on migrating RAID0 (if possible at all) would be appreciated. I have not started searching for this kind of info yet hoping that the system can be brought to life.

True, I do not have another system to swap the parts with. Well, at least I cannot do that with either with Dell portable or the old Deskpro.

The soonest I can make a serious attempt of faultfinding is going to be over the coming bank holiday weekend (yes, I do not have any other plans, it is sad, isn' it?).
This gives me a couple of days to get ready and to get necessary hardware.

It is sad again but I will not be able to borrow a PSU, I think.
I am the only quasi-PC-maniac around that I know. The PSU in Deskpro is a joke, c. 200W so I cannot use it. If I buy a c. 350W one just to find out that it is a PSU indeed then this would not be a wise buy. Then if I buy a modern 1000W monster and it turns out that this is the faulty mobo ... then I could still use it for the new rig ! Hmmm ... I need to think about it ...

Have anybody used that device ?
Is it worth getting one in my case ?
£10.30 + P&P from Ebuyer Antec ATX Power Supply Tester
Reviews sound "promising" ... well, at least this appears to be a cheaper option than buying a new PSU straightaway.
What do you think ?

You should be able to test the PSU by simply plugging in something like a hard disk and/or a dvd drive and then shorting the relevant pins on the block that normally plugs into the mobo. If it's an old PSU then it may not even require a load to start up. My Emermax will start up with nothing else plugged in just by shorting the correct pins.

It sounds like disconnecting CPU, mobo and GPU leaving everything else still connected to the PSU and then shortening correct pins on the PSU's mobo 24 pin connector. I think I can do that. The PSU should start then which should be easy to notice by spinning fans (in the PSU and on the case). I hope I got it right. Otherwise, pls, correct me.

Here is the update and a couple of further questions.
At last yesterday afternoon I found some time to carry out some testing.
I have a mixed bag of news to report.
I think I managed to isolate the fault - it seems that the XFX GF 8800 GTS 640MB is triggering the PSU's protection to kick in. :surprised:
It seems that the PSU is fine. :)
Luckily the mobo and the CPU are OK too so my RAID0 arrays survived intact. :)

How I tested.

At the end I did not use the PSU tester but I followed the advice given in this thread (many thanks to everybody). :bowdown:

I disconnected the mobo, the 8800 GTS and the CPU from the PSU leaving all other components connected to power.
I shorted the PS_ON pin of the PSU's mobo plug to GND.
The PSU started, both the system and the PSU fans started, the HDs made the spin too.
I measured all the voltages on the PSU's mobo, CPU and GPU plugs.
All the voltages were fine including the 12V on both rails (but under lower load, of course).
I removed the shorting from the PS_ON pin, the PSU went into standby with its fans still running and then I powered the PSU down.
Then I connected everything back and tested again finding that the fault was back.
I repeated the whole procedure again to confirm my findings.

Next I removed the 8800 GTS leaving all other components including the mobo and the CPU connected to the PSU.
I pressed the ON button and … the system started !
I did not allow it to boot up without the GPU so I powered the PSU down.

I replaced the 8800 GTS with the old Asus 6600 and powered the system up.
The system started up as normal. :rolleyes:
I loaded correct drivers and the system is working again. :rockon2:

What should I do next ?

I am still not 100% certain if the 8800 GTS failed or if the PSU's protection circuits developed a fault.
What do you think ?
The PSU appears to be working as normal now while I have no other rig to test the 8800 GTS.
Anyway even if I had another rig then I would be a bit apprehensive to slot the faulty 8800 GTS into it
in case the other mobo / CPU / PSU combo could not handle the faulty card in the same safe way as the original rig. :crazy:
What would you recommend to do ?
Shall I return the 8800 GTS to the retailer ? :help:
The card is a bit expensive so I do not want to leave the issue just like that.
Your advice will be appreciated.

Robert

Can you confirm that you have the 20pin + the extra 4 pin connector attached so that all 24 pins on the main power feed are attached, and that you are using the PCI-E power connector for the card otherwsie the graphics card might be overloading one of the 12v rails. If so then I guess you'd need to RMA the card. If you print out this thread to show your suppliers the testing you've done I'm sure they'd be happy to swap the card out for you.

Thanks for the post.
Both the 24 pin mobo plug and the 6 pin PCI-E plug from the PSU have been connected correctly since the card had been installed in March.
The whole setup has worked fine for 2 months without any signs of instability.
The PSU's manual can be found here :
http://www.enermax.co.uk/support/doc...aker_ax_en.pdf
Each of the two 12V rails is 18A max rated.
I do not know what happened a week ago when the card suddenly started to trigger PSU's protection circuits on startup.
Temperature damage ?
There were no warning signs of that.
The case is huge.
It has two 120mm intake fans in front cooling 4 HDs,
one 120mm exhaust fan at the back, one 80mm exhaust fan at the top,
double 120mm fans of the PSU and the custom CM Stacker cross fan pumping air over the RAM sticks and the CPU area.
The system was not overclocked.
To be honest, I hardly ever loaded the GPU either.

Test the GPU in another machine.

I will double check this but I think mobo / CPU / PCI-E power plugs draw from the same 12V rail (12V1).
One test I can still perform is to connect the PCI-E to the second rail via Molex-to-PCI-E converter (I have one).
I have 4 HDs, 2 optical drives and system fans & controller on the second rail (via SATA and Molex power plugs of the PSU).

EDIT

I checked and PCI-E is on rail 1 while CPU and mobo are on rail 2.

Final test :
I disconnected all components from the PSU and tested the PSU with the Antec tester.
The PSU appears to be perfectly OK, all green LEDs illuminate on the tester.
Then I connected just mobo and PCU (both on rail 2).
I switched on the system and booting sequence initiated.
The 6600 produced a picture on the screen.
I switched off the PSU, removed the 6600 and replaced it with the 8800.
I connected the PCI-E power connector (rail 1).
I tried to switch on the system but the same fault occurred as previously.
I carried out just one more test.
I connected PCI-E power via Molex connectors (but still on the same rail 1).
The fault was still there.
I feel I cannot do much more than that as I have not got another rig to test the card with.
The rail 1 was only loaded via PCI-E and the system still could not boot up.
I am applying for RMA.