Is my motherboard killing PSUs?

Hi, the system in question is my EP35-DS3R/QX6700 system in my signature.

Three months ago it started shutting down spontaneously and making odd sparkling/crackling noises. I pulled out the Bluestorm 500w I was using at the time and noticed the AUX12V 4-pin connector by the CPU socket was blackened on one side of the cable connector (the side containing the two pins closest to the edge of the motherboard). I replaced that Bluestorm 500w power supply with an Epsilon 700w unit I had laying around and the problem was solved.

But last night it spontaneously shut down on me again, and when it was running I was hearing slight crackling noises from the area of the PSU again. I haven't pulled out the AUX12V connector to see if it is scorched yet, I'll do that tonight. For now I've turned the system off. I crunch rosetta@home on this computer full time so it is always under a full CPU load.

I need some help guys , how should I troubleshoot this problem? Of course I don't want to keep throwing PSUs at it if the board is in fact killing them (can a motherboard even kill a PSU?)

Thanks in advance.

Well, stop throwing sub par PSU's at the problem for one... (doubt thats the issue though)

Only thing you can do is check the pins again and make sure the socket is solid and attached well to the mobo. Look for bulging caps, other scortch marks, etc.

Have you physically moved the tower recently?

it's not your mobo dude, it's the POS power supplies you're using. The PSU should be on the top of your budget consideration when you build a serious system, i.e. overclocking and crossfire. Sooner or later, one of those cheap *** psu's will burn out and take everything with it.

So, my suggestion, don't even turn on your PC again until you get a real psu. You should probably have ~650W unit that's 80plus certified. Antec, Corsair, BFG, Enermax, and PC P&C are all reliable brands and a good start. 500-650 will be fine, as long as it's a good supply. You shouldn't be spending less than ~$100 for a decent psu. Get the best one you can afford.

My Asus P5K-Deluxe (P35) is a Ram killer

I burned thru about 20gb's before I finally caught on to its ways

Thanks for the heads up guys. I have a Corsair 750tx in my Maximus Formula rig but I don't want to go through the hassle of ripping that out of the system for testing, breaking zip ties and whatnot, it's too solid a system to screw with.

I'll be checking for scorch marks tonight on the board and PSU connectors when I get home. I haven't moved it in recent months other then tilting it against my desk at a 45 degree angle to install/remove parts, the system has only been run in an upright position. In this system I'm running a mildly OC'd 65nm quad, 8800GT, and four 7200rpm hard drives. Do you think a 600w unit would be acceptable? This is my file/web/media/dns server BTW, and I'll have it crunching full time, but the graphics card is usually idle since the F@H gpu client doesn't support my OS and I don't play games on it.

Thanks again!

PS: I bought those Fortron PSUs before I knew about sites such as johnnyguru; the review of the Epsilon 700w was quite bad when I eventually read it.

A good 400W unit would be enough. 600w will be fine. Most important is to buy a quality unit. A good 400W unit is better than a cheap 1000W unit.

I say it may be worth trying your Corsair unit on that mobo. I would bet that Corsair would shut its self down before accepting damage to its self. If it works I would go with any Corsair unit over 450Watt since it runs at full load most of the time.

Bluestorm 500w were/are decent psu's .. is you connector on the board loose at all ? id pull the mobo an inspect closely front and back for small bulges in/on the pcb.. it will kinda look as if you put some clear nail polish on the board and let i dry .. the pcb slightly raises like a bubble .. at first glance you might not think it anything but close inspection you should notice it .. you could also try with a multi meter to test the psu ..but thats only if you have confidence on it not frying anything if you turn it on again .. dont think jumping the psu would tell you enough but you could just to see if your at least running no lad at spec.. i know its not gonna give you load voltages but you can at least see if your starting out where you should be .. if your psu is not under warranty pop it open an give it a looking too

I have seen motherboards kill both PSUs, CPUs, motherboards, and HDDs. All were different scenarios, but hte possibility that the board is doing that is there.

Have you tried powering up the PC on a different wall socket? Different UPS or surge protector?

Post pictures of the PC if you can.

I would try a corsair or antec or something. I don't know much about the PSU you have been using, but those do have protections on it that protect the motherboard and other components from getting damaged by a bad power source, yet it kills the PSU in the process.

I call these problem scenarios 'gremlins' because they are soo hard to figure out at first and each time you test the machine it kills whatever you put in it to fix it lol.

Well I did a little bit of testing and experimenting tonight and here is what I found:

First, I unplugged the ATX12V aux CPU connector and looked it over for scorch marks, there's none in the pin sockets on the cable connector or the motherboard connector. I plugged it back in.

Next, I unplugged all of my hard drives and tried to power the system, it POSTed. So I powered down, plugged all four back into power and tried to power the system, it almost immediately shut down. Next, I unplugged one of my hard drives (a Seagate 7200.10 250GB) and tried to power the system, it POSTed and booted into the OS fine. Now, I've shut down again, plugged the fourth drive back into power and tried to power it. It POSTed with the original configuration and is running as I type this.

EDIT: Here is a low quality photo taken by the awesome iPhone.

was the connection loose at all ??

A dodgy molex connector? Seen to many of them. Like the ones on fans with a molex pass-thru, try to push them in and the molex prong will pop out! Has only happened with cheap fans tho. Coolermaster, Thermal take, yateloon, Delta etc seem to have good molex connectors. seen a lot of sub-par PSU's with miss-aligned molex prongs too!!!

The four drives are SATA drives and I don't think any connectors were loose but they are all jammed close with the right angled SATA connectors jammed in as well, it's possible. A short time after having the four drives connected I tried to reboot and couldn't get it rebooted again until I disconnected the power to one of the drives. I think i'll squeak by with it for a few days until my replacement PSU arrives and give that a whirl.

yeah ..good choice on a new psu !!

I air canned the chunks of dust out of the inside of the Fortron 700w but it didn't help it power the system reliably. I replaced it with a Silverstone 560W last night and the system is running fine again, much how it was when I replaced the last power supply. My older Bluestorm survived in this configuration for over a year, actually with a higher overclock too, so I'm hoping the short lifespan I got out of the 700w powering this system is a result of that power supply and no other hardware. I was not able to power a bench system with it either, though the new PSU powered that bench fine last night too. I'm going to take the 700w to my friend's house and we might try to power some of his older spare computers with it, hook it up to the Kill-A-Watt, see if it can even power an Athlon XP system.

Is the Epsilon over 2 years old? Even if it is I would shoot FSP an email and tell them about it and in a nice way let them know that Corsair, Etc offer much better warranty's. Worth a shot IMO to see if they will work with ya.

ScottinIndy said:

Is the Epsilon over 2 years old? Even if it is I would shoot FSP an email and tell them about it and in a nice way let them know that Corsair, Etc offer much better warranty's. Worth a shot IMO to see if they will work with ya.

Click to expand...

It's three and a half years old, I think I killed my hopes of a warranty when I mangled the case trying to take it apart. Today I took it apart again and looked at the insides, this time I saw two bulged capacitors with significant amounts of brown crust on their tops, I didn't catch those the first time. My guess is those blown caps are at least part of the problem with this PSU.

Yep, the bad caps are most probably the main problem with that psu.

At least you aren't fighting trashy Gigabyte boards burning up like I am. So far I had the 4 pin connector melt on a P35-DS3R and yesterday I fried a fet on a P35-DS3L. And that was with 2 totally different psu's too. I guess those cheap Gigabyte boards just can't take a 200 watt cpu load while testing heatsinks.

A motherboard cannot kill a PSU.

Even if you short the voltage outputs, the PSU should just shutdown.

Even if it doesn't, other things (rectifier, regulator, etc) will blow up before the filtering caps. At high load, the caps are actually charged to a LOWER voltage (hence higher ripple), which doesn't bother the caps at all.

Caps usually blow up because the INPUT voltage (from the wall) is too high, which charges the caps to higher than their maximum voltage rating, and the dielectric breaks down and starts conducting, and give you the magic blue smoke.

Surges from the wall can certainly kill it, though. Are you sure the power line is clean? Are you using a UPS/surge protector?

Your situation seems similar to what happened to an ASUS socket 478 mb that I had 7-10years ago. I cleaned it out really well and within a short period of time noticed that it didn't seem to be running as well as it should be. I checked it out and didn't find anything obviously wrong with it so since it was still working well enough for my needs I just left well enough alone and continued to use it. It wasn't until 2-4 months later when it was really running bad (I don't remember specifically what it was doing, but think I was leaning toward the psu), that I knew I had to deal with it. That's when I found that the AUX12V 4-pin connector had actually melted onto the mb socket. Because of a similar instance after cleaning a wall phone/answering machine where water dripped inside and apparently got into the wiring (when it rang it literally sounded like it was dying, even when nobody was calling it would make weird noises, which lasted for years), it dawned on me that I had probably caused the problem when I soaked all of the fans in water when I cleaned out the computer. Even though I thought that I had given them plenty of time to dry out (apparently water seeping into wiring takes years to dry out-please refer back to the phone/answering machine situation), obviously I hadn't. So, hard/expensive lesson learned... water in wires + electricity don't work well together. The good news is that if the plug hasn't melted onto the socket and the pins don't look corroded and black/burned at the bottom of the socket (like mine), then you'll probably only need to replace your fans.