Here's a head scratcher.....

[ohio_grad_06]

Ok. So doing some work for a guy who'd previously bought an iBuyPower gaming PC. I'll give the backstory.

Originally, his power supply died, taking his graphics card(GTX 580) with it. Replaced the power supply and graphics card with a new Radeon 7970, and a Corsair CX 750 watt Bronze rated power supply. He gets it back, gets on it for 2 hours, and BAM, he hears a squeal, and turns it off. Apparently the water cooling setup that was originally installed best we can determine had failed, leaking onto the motherboard, and killing it.

Replaced the motherboard and CPU with an AMD FX 8350 and new motherboard(ASUS M5A-970 or something to that effect, it was like a 95 dollar board, kind of a middle of the road board). We put that all in, install the new graphics card and power supply again, ran stress tests on CPU things seemed good, sent it out. He gets on it for 45 minutes(tried to game for 5 minutes I think), and I get a call that he had seen a spark, and turned it off and the computer would not start. Checked it out, and there was a pronounced burn smell, as well as black marks on the wall. A VRM on the new video card had burned, I mean literally you could see where it had popped and the black finish on that part of the card would rub off easily.

Troubleshooting again, card also took the motherboard with it.

Exchanged the motherboard for a comparable Gigabyte motherboard with better power phasing, this board is basically a newer version of the board I run in my personal gaming system. Replaced the board, CPU again and graphics card with one just like the one that blew out.

Assembled everything, all seemed ok except for an electronic whining noise. Swapped power supply for a Cooler Master v700 Gold rated 80+ certified unit(Jonny Guru if you have heard of him, does power supply reviews etc, had reviewed this unit from what I saw on Google and basically seemed to say how great it was). We also exchanged the 7970 graphics card for an Nvidia GTX 770(better performing etc). Reassembled, formatted hard drive, clean install of Windows, everything seems great now.

Only question I have is when I press the power button, the fans and lights turn on for about 2 seconds, the computer actually turns off. If you listen, you can actually hear the power from the power supply/case actually click like the power button is being pressed, the computer like I said goes completely off. All fans spin down and computer seems off, but then you hear the click again as though you'd just pressed the power button again, and it boots. It appears rock solid stable, will run for hours doing updates, will endure stress tests with no overheating, things seem fine.

For a quick rundown here's the system currently.

AMD FX 8350-4 ghz chip(I disabled power management features like it's speedstep style features to allow it to go to low power mode etc in the bios, even disabled turbo core which is AMD's auto overclock feature)

Gigabyte 970a-ud3p motherboard(Has AMD chipset, I like this board because it has 8+2 power phasing so better power phasing for the VRMs, etc, set up for overclocking--I've got an older version of this board which is going on a year old and works perfect)
8 gb generic 1333 ddr3 ram(leftover from old setup)
2TB Hitachi hard drive(I don't like Hitachi drives either, but this is again left over)
Blu Ray Drive(Left over)

Cooler Master v700 power supply(Chosen because it is 80+ certified with a gold rating which I understand is supposed to be about the best you can buy)
Nvidia GTX 770 2gb graphics card(manufactured by EVGA--granted I did purchase this at open box for cost savings, but it was marked as a return to the vendor which means to me someone had bought it, returned it, EVGA apparently got it back, tested it as being good, and sent it back to the retailer).


Cooling wise the setup is this.

1. Cooler Master Hyper 212 EVO tower style CPU cooler with 2 140mm fans(I think) set up in a push/pull configuration---one fan pushes air into fins of the cooler, another fan is on the other side pulling air out the fins of the cooler.
2. 120mm rear exhaust fan pulling air out right behind the second fan on the CPU.
3. 240 mm or so case fan on to the top of the case(sits directly above CPU area and pulls air in from above).
4. 120mm case fan at the front blowing cool air across hard drive.
5. 2 120 mm case fans on side panel again to cool hard drive I think.
6. Power supply is bottom mount so no overheating on that end.

Case from what I've researched, is an NZXT Phantom. Link.

http://www.microcenter.com/product/...antom_410_Mid_Tower_ATX_Computer_Case_-_Black

His case is not exactly like that, but is similar, he got the computer originally about 3 years ago, and it may have been an iBuyPower design they got, not sure.

Otherwise now everything seems rock solid, can run crysis demo on it, stress test and it seems very happy otherwise. I am thinking I'm just done messing with this computer, it seems to work, I've had this PC in and out for 3 weeks and want it out).

The good news is that everything I buy I get through my local Microcenter, and put 2 year warranties on 99% of what I buy there, so any parts needing replaced have not come out of my pocket, no return shipping etc, just take it in and swap it. If you all have Microcenter nearby and don't go there, START!!! The money saved on this setup alone is worth it.

But anyway, as said it does the following.

1. Press power button
2. Computer powers on for 2-3 seconds, powers off.
3. Fans spin down and stop
4. Computer powers back on, boots
5. All seems right with the world.

Do I fight with this PC more or leave well enough alone? Lol.

 

[Krynn72]

Wow, you describe almost exactly an issue I had with my PC. To show how close my situation was to yours, I'll post a quick recap.

I wake up one day and try to power on my machine. Got display artifact moving around depending on what was moving on screen, stuck at low resolution, etc. So I replace the video card and it seems to work ok. About an hour of use and suddenly I smell electronics burning. I look to my left and smoke is literally pouring out of the top of my case and I see through some vents the flickering glow of fire, so I pull the plug from the surge protector and turned my window fan to exhaust to get the smoke and fumes out of my room.

Next morning I check damages, and see that the PSU had one of it's huge capacitors blow out (Never buying Zalman again). It also destroyed two of my five hard drives (burns all over them, they must have been what actually caught fire).

I replaced the PSU with a good Corsair one (possibly even the same as the one you used) and the system worked again... for a bit. After starting a game for about 10 minutes, the machine shut off and rebooted itself. In my case it didn't ever get to boot up again, it would power on for approx 3 seconds, shut off, fans would all wind down to a stop, then restart and keep cycling like that over and over. I took everything out of the system so all that was plugged in was the PSU, MB, and CPU and it kept happening, no error from the MB about not having RAM.

After replacing both the MB and CPU it was still happening. Now I was tearing my hair out. So I disassembled everything, and started plugging things in one at a time.

Thats how I discovered that one of my four RAM modules was the culprit. These were the only parts from the original build besides some HDDs, but my testing before never really suggested they had issues. That sure taught me to ever assume the RAM is ok.

It was an expensive PSU failure, costing me a GPU, two HDDs, and a RAM stick. It also possibly (likely) damaged the original MB, and maybe the CPU, but I can't be sure because I don't care enough to test it anymore now that my machine is all working.

But yeah, check the RAM

 

[ComputerRepairTech]

I had a MSI p35 motherboard with that issue. I never did take the time to properly diagnose it. The board ran fine for years though ill admit i rarely powered it down. It still works fine today still does that same issue you are describing the brief power down. I suspect its the load on initial power on but ive never heard any click.

 

[ohio_grad_06]

It is a click as though you had pushed the power button again. You literally hear the power get cut to the board and back on. I think I may leave it alone. If the board ever dies I've got 2 year replacement plans on all parts.

On a side note though, I've noticed it sometimes hangs with a green screen. I realized the ram is 1333 speed and that the bios has a ram multiplier that made the ram run at 1334. Manually downclocked ram to 1066 speed and started prime 95 blend test. So far it's run more than an hour with no error. Got this one at home to work in so thankfully I can keep tabs on it.

 

[ComputerRepairTech]

It is a click as though you had pushed the power button again. You literally hear the power get cut to the board and back on. I think I may leave it alone. If the board ever dies I've got 2 year replacement plans on all parts.

On a side note though, I've noticed it sometimes hangs with a green screen. I realized the ram is 1333 speed and that the bios has a ram multiplier that made the ram run at 1334. Manually downclocked ram to 1066 speed and started prime 95 blend test. So far it's run more than an hour with no error. Got this one at home to work in so thankfully I can keep tabs on it.

1334 wouldnt be an issue, should not need to download. dunno about the green screen.

 

[ohio_grad_06]

Yes my plan is if it does it again return the video card for a similar model, pick up a new set of ddr3 1600 ram and nuke and pave, then allow Windows to use mostly generic drivers except for video, and not allow it to install catalyst control center. The green screen is weird. Can't seem to turn up much on Google. Just random.

But ram was about only thing not tested. So far after over an hour prime 95 is stable at these settings. I think the video card is factory overclocked so maybe that has something to do with it.

 

[JustMe]

My EVGA x58FTW motherboard w/i7-920 does the same thing. You press the power button, you hear the fans start spinning, a couple of seconds later the whole system powers off then starts up again as if you hit the power button again.

I could never figure out what was causing this so I let it go. That was in April 2009. The system is still in daily use today - no problems.

 

[B Trevathan]

Only question I have is when I press the power button, the fans and lights turn on for about 2 seconds, the computer actually turns off. If you listen, you can actually hear the power from the power supply/case actually click like the power button is being pressed, the computer like I said goes completely off. All fans spin down and computer seems off, but then you hear the click again as though you'd just pressed the power button again, and it boots. It appears rock solid stable, will run for hours doing updates, will endure stress tests with no overheating, things seem fine.

Sounds like a shorting/grounding issue, probably a capacitor or diode starting to go bad somewhere. After you run it for a while and then shut it off for just a very few seconds then back on, does it do the same thing?

Since the case is one of the few things you haven't replaced you might want to check/replace the power switch.

 

[ohio_grad_06]

Ok, so here's the deal. We had Church. Messed with it a bit after getting back from services. Ran GSmart Control short test on the hard drive which showed some old age and possible prefailure symptoms, but nothing in red no errors on drive so should be healthy enough to run. Started Memtest86 which has found no errors so far.

As I said, the ram was from the original machine that had liquid damage to the motherboard. Not sure how good the ram is, but you can tell the ram they used was the cheapest they could find at the time. No heatspreaders, used 4 2gb sticks of ddr3 1333. So far Memtest 86 is not finding errors.

The case, I don't know about the power switch, maybe it was the way I hit it, it has not done the power on/power off thing the last couple of times so tenatively that may be done with.

The only remaining thing I can think of is I've seen this system lock up with a green screen a couple of times with no error codes, or a black screen. Again no error codes just random.

I was using previously a Radeon 7970, which I never experienced those symptoms with, however I did get an electronic whine from the system, which lead me to change that for the GTX 770(they didn't have anymore 7970's or any other high end Radeon cards for that matter left). But the 7970 never gave me the weird screens. So I'm thinking things were stable, the most recent changes I did were the power supply and video card due to the electronic whine that concerned me.

What we've done brief recap.

1. CPU and Motherboard are both new.
2. Changed power supply for 700 watt 80+ certified unit with Gold rating.
3. Changed video card in case electronic whine had come from it. Changed from Radeon 7970 to GTX 770.
4. Reinstalled OS with fresh format.
5. Changed power settings so that CPU should not use any power saving modes or turbo core to auto overclock.
6. Turned off Hibernation and maybe even sleep modes
7. Disabled AMD Catalyst Control Center from starting on bootup as that seemed to interfere with the Nvidia drivers at one point.
8. Tested hard drive on gsmart control short test which appears ok.
9. Set memory settings manually in bios to 1332mhz, this Gigabyte seems to like to run the memory 1mhz or slightly above what the bus says, which may slightly overclock the ram.
9. Running Memtest86 which seems to suggest ram is ok also.

Going to finish Memtest86 for a while, and maybe do some gaming on it to make sure things look stable and restart a few times to make sure things seem to play nice. If so, giving it back. If it locks up anymore, I think I'm going to swap the video card for a brand new model(non open box), nuke and pave it, install only drivers for board, and try not to install Catalyst Control Center, and basic Nvidia drivers. Will let you know.

Anyone have any other suggestions?

 

[lan101]

Whenever I think ram could be culprit, I just put a new or a known good module in and see if problems go away. Especially if there's like 4 modules. I remember a few years back a pc had 4 512 MB's. I felt one was the culprit. Called the customer about a 2GB stick and it was a done deal. Better than all the time spent running memtests and checking different modules.

I know it may not always be feasible, but to me it's better trying that out.

 

[ohio_grad_06]

Yes I'm really beginning to dislike that machine. I ran memtest and it ran ok for a good while, rebooted went to open the crysis demo to test. Locked up with a black screen. To elminate any doubt about the hard drive, I'm tossing in a spare 320 gb seagate that i'm pretty certain is a working drive. The drive in the system was a hitachi, which I just haven't seen many hitachi drives that impressed me honestly. Going to nuke and pave the 320 gb and see if windows does not respond better. If I'm still getting weird issues, I've got some ddr3 1600 gskill in my personal box that I know is good that I can toss in to try. If that does not do it, it's about got to be the video card. Just at a loss you know. Feel like I've hit this PC from every angle and am just having a time with it.

Like you said though, I can't assume ram is good or any parts from before liquid damage to the old board because you don't know what it affected, and the old power supply had failed, possibly it affected remaining components in ways I don't know.

 

[ohio_grad_06]

Happened again with a Seagate hard drive that I'm relatively certain is good. Swapped in a known good set of gskill ddr3 1600 ram and will see if crashes occur anymore. If not that then the new gfx card I'm guessing.

 

[ohio_grad_06]

Yes I worked late in the night on this one. Same thing occurred with known good ram.

Board CPU and power supply all new. Tested with different hard drive and also known good ram. Still got black screen that was just a hard lockup. Only thing not brand new was video card which I bought open box to save a little money.

Ram I tried was not new but was a known working set. Hard drive I got same issues with his drive and a tester. The blu ray drive should not affect anything. Anyone have suggestions before I return the card?

 

[CLC]

 

[tek9]

You tried booting in and out of the case? Nothing shorting out?

 

[ComputerRepairTech]

you said its locking up with a black screen while gaming...well thats not likely ram or hard drive. thats usually video card...possibly psu or mb but usually video card. Does the event log show anything from nvidia or anything?

 

[ohio_grad_06]

Have not tried booting outside the case, but don't think grounding is an issue. I've put together how many of these things and never had that trouble, plus it will run fine and just do this randomly. I did buy the card as open box, it was one that was sent back, supposedly tested good and sold as open box. Which the card works much of the time, but these lockups are random. They seemed to only start when I changed Video cards from the 7970 with the whine to this card. So I guess there is my answer.

As far as event logs, that's the weird part, no event log created except where I have to manually power off the box, no stop codes or error codes, just black screen with non responsive mouse and keyboard.

Like the cartoon, how I felt lol.

 

[ohio_grad_06]

Black screen btw is random. Not necessarily during games but sometimes in the desktop, running Windows experience index etc. Just random. You can load a game sometimes it works sometimes not. I'm thinking video card since I feel like I've eliminated every other component. The other thing that still sticks to me is that I don't remember it acting out with the old card installed. Back to Microcenter. Sigh.

Just glad I didn't feel with newegg. I've changed practically every component twice. Can only imagine the trick fees. Microcenter is about a 40 minute drive. Oh well.

 

[redmon]

I feel for you.......

 

[LifelineIT]

Ever watch House, MD? It's never lupus? Except when it is?

I think what you have is a series of unfortunate events that may or may not be related, but I also had a similar issue once.

You know what it was?

It was a goshdarned $2 SATA cable. No joke.

But, FWIW, I've seen bad RAM test fine in memtest for a day or more, and I think that you chased bad parts with used/unknown parts and you muddied up your water pretty good. Your PSU blew your RAM as well, your RAM ate your next board and possibly your next video card, OR your PSU blew some caps or caused a microfault, and that's causing a ground fault somewhere under load, and/or it's eating components. I realize you're on a new board now.

As for your switching issue, that sounds to me like a bad switch. AFAIK all switches now are intermittent-on, and what you're getting is a bridged connection somewhere, probably inside the switch body. This bridge COULD be happening at other times as well, but it's probably happening when you depress it. One of my own rig currently has the power button wires stripped out and dangling, I have to short it when I turn it on, what's that they say about mechanics with crummy cars?