quixadhal Posted April 8, 2012 Share Posted April 8, 2012 Hi guys, I recently upgraded my CPU from an AMD Phenom II X3 720BE (which ran stable, overclocked to 3.2GHz) to a Phenom II X4 975BE. Oddly enough, I got random crashes and resets when this new CPU was mostly idle, even when not overclocked at all. I RMA'd it for a replacement, and the new one seems to be having the same symptoms, leading me to start looking at the rest of my system components for something to blame. :) What brings me here, is noting in the CPUID Hardware Monitor program's display, that the -12V rail seems to vary quite considerably from one moment to another, while the other rails all seem pretty solid. Currently, it's sitting at -10.50V, but has been as low as -11.33V, and as high as -7.10V (when the CPU was underclocked by the BIOS by power saving being on). Until switching to this new CPU, my system has been rock solid, but since that rail seems to be reading so far off of normal (compared to all the others, which only vary a few tenths of a volt), is this normal, or a sign that my PSU may be failing? Link to comment Share on other sites More sharing options...
peanutz94 Posted April 8, 2012 Share Posted April 8, 2012 Those readings are wrong. Your computer would not even run if those values were true. The only way to know for sure is to use a multimeter and check the voltages at the PSU connectors. Can you post a screen shot of the monitoring progra that you are using to read the voltages? What are 3.3v 5v, and +12v reading Link to comment Share on other sites More sharing options...
quixadhal Posted April 8, 2012 Author Share Posted April 8, 2012 Heh, that's why I was asking. The -12V is the only one with the strange readings, here's what it reads currently after a fresh boot. It looks like power management dropped the CPU clock, and the -12V rail and vcore both drop along with it. http://i302.photobucket.com/albums/nn96/quixadhal/weird_voltages.png Here's newer versions of the monitoring software, since I noticed they were a bit out of date. http://i302.photobucket.com/albums/nn96/quixadhal/weird_voltages_2.png Link to comment Share on other sites More sharing options...
peanutz94 Posted April 9, 2012 Share Posted April 9, 2012 The -12v rail is for standby, and i could totally be wrong about this, but i don't believe it could cause random restarts. Maybe coming out of sleep or hibernation,or maybe at cold start, but not after it's been up and running. I hope someone else reading this will correct me on this or at least let me know if im on the right track. The rest of your voltages look great! What i find odd is that my MB does not show a -12v reading. I'm wonder if thats an AMD thing? Have you checked your event logs to see if there is a cause of the crashes? Link to comment Share on other sites More sharing options...
quixadhal Posted April 9, 2012 Author Share Posted April 9, 2012 Yeah, in cases where I get anything in the event log, it seems to be the kind of issue one gets when trying to overclock too far (0x3b and 0x1e blue screens), but... I'm not. This is at stock settings. I even tried bumping the vcore and nb-vcore a notch to see if maybe the motherboard were misreading it and sending a slightly too low voltage, but it makes no difference. What's odd is, the previous CPU I had was overclocked a fair amount (from 2.8 to 3.2GHz) and was rock-solid stable. That's why I started looking at the PSU, since the 720BE was a 95W CPU, and the new 975BE is a 125W one. So, last ditch effort, I turned off all the power management things, fixed the speed and voltages to auto in the bios, removed my usb 3 card (which is the only addon card besides the video card), and we'll see what happens. This is the second CPU that's had the same symptoms, so I doubt it's a defect there. If it only crashed while I was stress testing, I'd say it was overheating, but I've literally had it crash or freeze while I'm typing in a web page with the CPU idling at 4%. I think my two options are going to be to buy a new motherboard, or to return the CPU and force them to exchange it for a different model, in the hopes that the MB manufacturer lied about it being compatible. Thanks for the help, hardware troubleshooting isn't my strong suit and it is appreciated! Link to comment Share on other sites More sharing options...
Corsair Employee RAM GUY Posted April 10, 2012 Corsair Employee Share Posted April 10, 2012 A lot of AMD MB's don't use the -12 Volt rail at all its just there from the ATX Spec but I would check in the BIOS and see what it shows. If it is unused on your specific MB you may get odd readings in a program like this because it may not be terminated (grounded) properly if its not used to power something on your MB. Link to comment Share on other sites More sharing options...
quixadhal Posted April 10, 2012 Author Share Posted April 10, 2012 I'm loathe to jinx things, but I *might* have figured out the problem. My RAM modules are rated at 1.5 volts, but are supposedly spec'd to accept up to 1.8 volts if you want to overclock to try and get more precise timings. My motherboard only goes DOWN to 1.6 volts, so I never worried about it since I'm not overclocking to the point where the RAM timings are that important. On a last minute hunch, after the last crash, I thought MAYBE having a faster CPU might be stressing the memory subsystem more than the slower one was, so I bumped the voltage from 1.6 to 1.65, and *crosses fingers* it hasn't crashed since then. Does that sound at all plausable? Link to comment Share on other sites More sharing options...
Corsair Employee RAM GUY Posted April 10, 2012 Corsair Employee Share Posted April 10, 2012 Yes it does and I would check with the Memory manufacturer to see what they suggest and the memory frequency should be set no higher than DDR1333 with the CPU you have listed. Link to comment Share on other sites More sharing options...
Recommended Posts
Archived
This topic is now archived and is closed to further replies.