Why has this happened to my poor PSU??!!

luminous · Post by **luminous** » Fri Jul 01, 2005 11:47 am

My PC has stopped working with some very weird issues. What looks like a graphics card type corruption comes across my display in Windows. Everyone at first thought that it was the graphics card.

Anyone, to cut a long story short, after a lot of component swapping it has been confirmed that it is a mobo problem.

But what has shocked me is the state of my PSU's mobo connector when I took it out of my PC. Just look at the state of the connector!! All of the red wires show burning.

http://andy.photos.cn.com/c599851_1.html

What could have caused this?? Was it something that I have done

I modded the PSU so that it had only 1 slow turning fan instead of 2 fast ones. The inside of the PSU looks fine when I retrieved my fan, and the PSU still works. Although I will not be using it now.....

lenny · Post by **lenny** » Fri Jul 01, 2005 12:12 pm

Holy s**t!!

From your photos it seems like its the 5V connectors that are charred. Since you've isolated it to the motherboard, I guess then some malfunction is placing a lot of load on the 5V. Not sure if a capacitor failure could do that. Perhaps you can check the voltage regulators on the motherboard as well.

When the system was working, did you check the voltages? Were any of the rails out of spec / fluctuating? I recall a post where someone has a Speedfan graph showing 5V all over the place. That's not you by any chance, is it?

Edit: It's not - found that post by runcible here. Probably unrelated to your problem.

POLIST8 · Post by **POLIST8** » Fri Jul 01, 2005 12:35 pm

Damn dude. That's nasty.

How long after you modded the PSU did you notice problems?

I'd replace both the PS and MB just to make sure you don't fry anything else.

It wasn't a Yum-Cha brand, so it probably was the result of modding (or bad wiring in your walls).

alglove · Post by **alglove** » Fri Jul 01, 2005 2:37 pm

Were you playing around with any 12V/5V/7V mods, by any chance?

CX23882-19 · Post by **CX23882-19** » Fri Jul 01, 2005 2:57 pm

You'll need to replace the motherboard and the PSU. Once the ATX connector on the motherboard is scorched, it will quickly burn any new PSU's connector you plug into it.

Often this is a result of inadequate connection - over time, especially if you do a lot of plugging/unplugging the wires will work loose in the connector, causing arcing and a hot connector which results in that scorching.

chylld · Post by **chylld** » Fri Jul 01, 2005 3:26 pm

CX23882-19 wrote:You'll need to replace the motherboard and the PSU. Once the ATX connector on the motherboard is scorched, it will quickly burn any new PSU's connector you plug into it.

Often this is a result of inadequate connection - over time, especially if you do a lot of plugging/unplugging the wires will work loose in the connector, causing arcing and a hot connector which results in that scorching.

i agree with this comment, i don't see anything but arcing as the cause of those burn marks. but on the 5v wires only???

*baffled*

although it is mighty tempting to use this as a case against the 7v trick - "look what happens when you send power back along the 5v rail..." That's if you used it at all, of course

qviri · Post by **qviri** » Fri Jul 01, 2005 3:56 pm

*switches his fans to 5V instead of 7V*

Hmm, they still cool adequately, might as well go with that...

wim · Post by **wim** » Fri Jul 01, 2005 6:06 pm

cool!

thanks for the pics

brigitte · Post by **brigitte** » Fri Jul 01, 2005 9:25 pm

This exact same thing happened to me once without any modification of my old Zalman PSU (400A version). Dead PSU, dead motherboard and the graphic card had taken a beating and caused intermittent boot failures with the new board. The PSU was connected to an APC surge protector. I never new what caused it, the insurance covered the damage on account of an electricity surge.

So don't automatically blame the 7V trick!

Mr_Smartepants · Post by **Mr_Smartepants** » Fri Jul 01, 2005 11:52 pm

That type of damage is either caused by arcing (loose connectors, corrosion, etc.) or possibly a failed buffering capacitor. Since this affects all the 5V lines, this is a definate possibility. Capacitors are sometimes used to buffer (or smooth out) the square-wave output into more of a smooth sine-wave output.

Either way. Your PSU is toast! Do NOT reuse it for anything. Your mobo ATX connector needs immediate close scrutiny. Any evidence of arcing (discoloration or corrosion) would require a mobo replacement.

Elixer · Post by **Elixer** » Sat Jul 02, 2005 3:27 am

What PSU were you using? I know cheap psus have been known to fry motherboards from a very high voltage spike when they fail. Maybe this occured?

luminous · Post by **luminous** » Sat Jul 02, 2005 8:40 am

Wow, thanks for all the replies people. My computer access is somewhat limited at the moment, so I can only post back at infrequent intervals until I get stuff sorted.

The PSU was a Chieftec, you can see exactly what model in one of the pics (I forgot which one it is without looking myself)

PSU was airflow modded over 2 years ago, and has taken heavy use since then. Has also had its lower fan intact blocked so that heat from the CPU did not go into the PSU.

Voltage rails always appeared to be reasonable, although I did not check them recently.

I have not used the 7V trick at all, so its not that one either

I have only plugged, unplugged the connectors maybe 10 times at most (for the PSU). Only plugged into the mobo once.

The machine is always run with the left side panel off, as I find the mid frequency noise less irrating than the lower frequency noise I get when the panel is on (so should be enough airflow).

My fans are all controlled by a fan controller that runs off a molex.

The only naughty thing that I have done to the mobo is to cover 3 of its coil in silicone cos the whinned a lot. I could touch them with my finger before I covered them, so there were not that hot before. I have posted about this with pics under coil whine.

Mobo is defo toast, no point looking at it, it just will not work. It hung on for a while when underclocked on 166Mhz FSB, but as soon as I tried to up it to 200Mhz with a lower multipler it died completely.

RIP my FOURTH Asus Mobo in that system:
1) Horrid coil whine
2) Broken ATA controller
3) Horrid coild whine, random lockups
4) Graphics corruption in windows, system hanging, and coil whine.

I'm only gong to give you one chance at this......

Now that my mobo is out of warranty, is the replacement made by Asus or Abit?

peteamer · Post by **peteamer** » Sat Jul 02, 2005 8:58 am

Abit vote here, mines been completely (less stupid passive northbridge cooling experiments...) rock solid for 4+ years...

Though active cooling on the northbridge may be worthy of note... or attention......

(just go A64...

. About to order that way myself...)

Pete

luminous · Post by **luminous** » Sat Jul 02, 2005 9:02 am

You are right about Abit

I was planning on using Zalman's Northbridge cooler (I have one already). I take it that this may be a problem from what you have just said!!

I was thinking about A64, but it seemed to be a little too expensive. A new mobo, CPU and HSF....

After all, the HSF I would have gone for would have been a Reserator...

Will do that project next year when the dual core A64's have come down in price

....I hope

peteamer · Post by **peteamer** » Sat Jul 02, 2005 9:49 am

I was planning on using Zalman's Northbridge cooler

Would prolly be alright... mine failed cause I only removed the fan, leaving just the orig. heatsink

...

After all, the HSF I would have gone for would have been a Reserator...

What about a WACC resevoir? As good as/better... depending on which one you get... and you get to choose the (better) pump...

Pete

luminous · Post by **luminous** » Sat Jul 02, 2005 9:55 am

Not sure what you mean by a WACC reservoir...

The thing that I liked about the Reserator compared to just about all other water cooling is that it is fanless. Most of the others need some sort of rad fan that defeats the object in terms of silence.

The next PC that I intend to build will have only one very slow 120mm case fan that is soft mounted. Everything else will be fanless. Does your idea fit with that?

P.S. - you tried it without any HSF

oops

peteamer · Post by **peteamer** » Sat Jul 02, 2005 10:22 am

WACC resevoirs here. Summer discount on now

...

Some crap descriptions of performance are available... I can tell more if you want, nobody seemed interested at the time

...

Gotta polish a car... will be back in an hour(ish)...

P.M. if you like.

Pete

luminous · Post by **luminous** » Sat Jul 02, 2005 10:31 am

They do not look a nice as a Reserator, but do they work like one. Do you need a rad at all when you use one of those systems?

Hope the car polishing goes well.....me? I'm having a beer

peteamer · Post by **peteamer** » Sat Jul 02, 2005 11:02 am

luminous wrote:Hope the car polishing goes well.....me? I'm having a beer

Beer well done (lots) already...

...

Thoroughbred 'B' XP2400+ (10% overclocked) (NOT cool) and GeForce3 Thingy vid card water cooled, CPU cooled to 35-36 in ambient of ~20with NO fan...

Reserator works well... but WACC is better, for less... as far as I can tell...

More later...

Pete

P.S. Can't compare looks to a Resorator.. haven't seen one... Suffice to say girlfriend hasn't objected to WACC yet...

luminous · Post by **luminous** » Thu Jul 07, 2005 10:46 am

Well I found out the answer.....and its very embarrising

What can I say????

Well, I have one of those heatsinks that has to be bolted through the mobo. It is the type that has lots of different washers for different purposes....well I mixed em up

I used the metallic washers against the mobo

One of them managed to short a connection on the mobo......

What is really amazing is that it has worked like this for 6 months!!! It appears that it has only stopped working when the electricity has actually burnt through the washer breaking contact. The mobo in the immediate area has melted, so there was no route for the power to go through....

How many people have managed to burn through a washer!!

Ah, well. At least I now know what I did.....a warning to all of you!!

So that little experience has cost me a mobo and a PSU at a minimum, only time will tell what else I have damaged as I continue my rebuild.

peteamer · Post by **peteamer** » Thu Jul 07, 2005 10:56 am

luminous wrote:...

.. Whoops...

..

Just keep repeating " Experience is a learning curve and a 'Good thing'..."

Pete

lucienrau · Post by **lucienrau** » Thu Jul 07, 2005 11:11 am

Don't feel bad, most of us have done something silly to our systems at one time or another. I'd much rather it be the mobo than the chip.

In my time building personal systems (just for me & friends & family), I've burnt out 1 chip, cracked 2 (early Athlons were a complete pain to mount the heatsink on correctly), shorted a HD and fried a PSU (will never buy a yum-cha psu again).

That's over 18 years though so it's not too bad. It would help if hardware manufacturers labled things rather than just dumping them into a bag.

Pilot · Post by **Pilot** » Thu Jul 07, 2005 11:43 am

Two days ago, I hardly burnt my A7N8X-X a silly way :
Because of the coil whine you mentionned, i tried to wrap the coils with my hot glue gun.

Time to close and remount the PC, the glue was cold, so ok for the launch : I started the computer and waited for windows to boot as it's when the drivers trigger that the coil whine appears (NVidia cooling patch).

=> It was great, almost no whine that couldn't be dampened with gedicoustic. Only, 5mn later while I was doing something else, the computer reseted and froze during startup : I instantly stopped the PSU.
I spent almost half an hour to drop the hot glue, hoping the mobo wasn't fried.

By now, it's working and that merry whining coil is back...for now.

luminous · Post by **luminous** » Thu Jul 07, 2005 11:55 am

Thanks guys, makes me feel a little better now

The outlook is not too good for me though. I just powered up the new machine, and well, how shall I put it. The original problem is still present. Graphics corruption all over the place in exactly the same way as before.

So that is at least one other hardware componet that is toast

Just going to be a question of swapping out the GPU for a known good one and see if that solves it. Only things left after that are the ram and CPU. Pretty sure the ram is OK as it memtests fine.

Oh arses......

" Experience is a learning curve and a 'Good thing'..."

" Experience is a learning curve and a 'Good thing'..."

Nope, tis not working...

PositiveSpin · Post by **PositiveSpin** » Sun Jul 10, 2005 12:55 pm

I looked at the WACC offerings, as you suggested. The reservoir / radiators do look good. I had a look at their CPU block, too, and now I'm not sure about them.

They say "All WACC H20 Cooling Blocks come with our own design of high performance aluminium based redial heat-sink cooling cores" - not the most convincing of statements, given that the photos clearly show a COPPER block, and it has no radial (I assume "redial" is meant to be "radial") features at all.

Surely they should know the difference between aluminium and copper?

Viperoni · Post by **Viperoni** » Mon Jul 11, 2005 2:30 pm

I had similar but not as harsh situation happen with an Enermax 350w and a Epox 8kha+.

The PS simply couldn't handle the load on the +5v rail and instead the MB tried to draw more current from the 5v rail in order to make up for the sagging voltage... which caused the heat buildup in the connectors.

I sold the board a few months after noticing the damage for a super cheap price, and it still works fine for the guy I sold it to... more than a year and a half later.

wim · Post by **wim** » Sun Aug 07, 2005 2:05 am

check it out, more 5V burns. this is from my own system, it's a seasonic ST (revA3) which has recently been put out of service when i got a fanless psu. up near the plug those red wires are charred brown and the wire has actually gone very stiff and brittle. a couple of centimetres back it's bendy and normal.
the funny thing is this psu never failed or anything, the seasonic was only 1 year old and was going strong, i just changed it because i wanted something quieter. i only noticed the burns when i was sifting through my spare parts box the other day, and it made me remember this thread..

so - what can be the cause of this phenomenon, given that the system was p95 stable, not particularly powerful, and coasting along fine? i don't use 7V fans so that can not be related, and i don't recall any shorts while tinkering. i do use s3 suspend all the time, but i think that doesn't use the red +5V wires..

perhaps the connector has a significant resistance and it gets hot, like a resistor under current, and with no airflow because it's encased in that plastic rectangle cell? but then why only on +5V (okay, lower voltage means higher current for given power - but don't most computers draw much more power from 12V)? could properly soldering the wires into their pins prevent the burns from happening? *confused*

Why has this happened to my poor PSU??!!

Why has this happened to my poor PSU??!!

WACC CPU block