VoodooGlowDoll

Distinguished
Oct 26, 2010
4
0
18,510
I'll start with the quick, TLDR version and simply ask; what is the best way to thoroughly test a power supply for any possible failure?




For those more interested in my first-world trials and frustrations the long version is as follows:


The main PC I use most at home is a gaming/web browsing PC that I occasionally use for design & AV work. Almost all of my personal files are kept on a personal server and work files are stored using cloud services and accessed with a laptop. This week I bough a second XFX Radeon HD 6970 in preparation for a few upcoming games, this addition makes the specs of my main/gaming PC:


Win7 X64
I7 2600K
Asus Maximus IV Extreme P67
2x XFX ATI Radeon HD 6970's in Crossfire
16 GB G-Skill DDR 3 1600
OCZ ZX Series 1250w PSU
OCZ Revo X2 120 PCIe SSD (OS, design tools, some games)
Western Digital Raptor SATA 3 HDD (storage, some apps & games)
SoundBlaster PCIe Sound Card


I bought an OCZ ZX Series 1250w PSU two months ago, played some games, had some fun and everything seemed to work fine until two days ago when I installed a second Radeon 6970. I installed the card in PCIe slot 3 but it did not show up in the BIOS, GPU-Z, or CCC, so I swapped the video card and sound card. After that the video card showed up but the sound card didn't. Without changing anything I ran the Metro 2033 benchmark tool (installed on Raptor HDD) and it started, showed good FPS but crashed during the third loop. Removed the sound card completely and placed my PCIe SSD (OCZ Revo X2) in PCIe slot 3 and booted to "no boot device". Determined PCIe slot was likely DOA since I had never used it before and the board is less than 6 months old. Moved the SSD back to the 4x slot and removed the sound card completely.


Left the sound card out leaving both video cards in crossfire and the PCIe SSD and the Raptor drive hooked up. I was busy for a couple days but used the PC to browse the web, and play an hour or so of BFBC2 (all files on SSD) and sync new podcasts to my ipod ea morning (app on the Raptor, mp3's on network). When the weekend came I attempted to access some files on the Raptor drive and noticed extremely long load times. Checked windows system log and found numerous disk errors over the past two days relating to the Raptor HDD. Ran chkdsk for 12 hrs, repaired hundreds of files but only reached 11%. Rebooted PC and recovered the small ammound of data on it I needed but did run into a couple files that wouldn't transfer, confined HDD failure via SMART. Removed the HDD and tried to play BFBC2 (installed on SSD) but it caused the PC to BSOD after less than 10 min. Removed the old, working 6970 and moved the new one to it's slot using the same power cables as the working card. Ran BFBC2 and pc crashed again this time with no BSOD. Placed card in new PC resulting in the same failure. Determined card to be DOA.


At this point the possiblility of a bad PSU has been in the back of my mind for a while so I hook up a basic "yes or no" psu tester Removed new 6970 and installed the old one back in it's original slot. I was able to play BFBC2 using on-board sound for over an hour. Shutdown and tried to install the SoundBlaster to it's orignal slot but it wouldn't show up in the windows device manager regardless of what slot I put it in. Installed it into another PC and it wouldn't show up there either. Determined sound card to have failed.


So, if you're keeping count, the components that have failed in the past three days are; the motherboard, the HDD, a new video card and the sound card. All this happening so close together makes me go back to wondering if they're all symptoms of a PSU failure so I pickup a more detailed CoolMAX PSU testerbut all voltages show as being within normal limits, nothing is over or under the specified limits. In-fact, as I write this I'm running the PC with the original 6970, and the PCIe SSD with no issues.


HDD's fail, video cards arrive DOA, mobos's have bad slots from time to time but having that soundcard suddenly fail worries me. Considering the circumstances and the number of components that failed in such a short time frame the possibility of a PSU failure is really nagging at me despite the tester. I'm hoping someone can recommend a better way of testing the PSU than hooking it up to another PC, increasing the load, and seeing if it gets cooked.
 

beenthere

Distinguished
Cards are powered by the mobo VRM so they would not die from a bad PSU. The mobo might go bad if the PSU was over-volting. You could check this with a digital volt meter. There is no other easy means to properly test a PSU without placing it under load and checking the voltages to see if they are within specifications.
 

VoodooGlowDoll

Distinguished
Oct 26, 2010
4
0
18,510
Is there a non-software method of checking the voltage outputs from the PSU while under load? I can check the voltages controlled via the mobo like CPU and RAM via the ASUS RC Bluetooth but those aren't the numbers I'm worried about. Is there some other type of tester I'm unaware of that sits between the PSU and the mobo connectors?
 
No, but if you are handy with a soldering iron, you could take a main power cable extension and splice a main power cable from a dead PSU to it, then plug a power supply tester into the third leg of the assembly.

Because the third leg is not carrying any current, you only need one wire of each color. I suggest leaving the green wire out because plugging the green wire into the tester will turn on the PSU.

This will let you monitor the voltages while they are under load.