Hi, I've had my pc a good few months now but have recently started experiences system crashes whilst in games, especially graphically intensive ones like Far Cry 3 and Sleeping Dogs. I haven't added any new hardware recently and the only things I've installed are the official Windows updates and a couple more games. The screen goes black, the video to the monitor cuts out and the monitor powers down, and there is usually a looping sound, at which point I am forced to restart - it looks very much like the typical black screen issue. I have cleaned and reinstalled/updated drivers to the latest catalyst beta ones, both underclocked and overclocked my card using Trixx, checked for cooling issues inside the PC and replaced fans, and have used Trixx to disable ULPS, all to no avail. I really don't want to have to RMA the card, and definitely not buy another, as I'm living in Korea now and I've had terrible experience with the local retailers, and the mail service ^^;;; Is there anyone who can tell me what's wrong with my system or why this issue has suddenly arisen? I read elsewhere here that it may be related to my PSU or some other conflict in the machine but I really lack the expertise to determine what. Anyway, here are my specs: CPU AMD FX-4100 @ 3.43Ghz RAM 2 x 4gb Samsung PC3-10700U DDR3-1334 MB ASrock N68-VS3 FX GFX Sapphire 7850 2gb OC Ed. PSU "Low noise Rio power supply R-6000LW v.2.3" - 600W HDD 2x Seagate 500 GB NTFS Sata600 7200rpm 16mb cache OS: Windows 7 Ultimate 64bit Video Card Drivers in Use ( and tried ): Catalyst 13.11 beta (latest), Catalyst 13.10 Motherboard Drivers Used: Apps/Games causing issues ( and if they've been patched ): P1.80 10/31/2012 (latest) All major components check 100% after thorough testing. System clean of viruses / spyware etc. Other things tried: Checking for heat issues/replacing thermal gel/cleaning fans (everything cool now) disabling ULPS through Trixx Overclocking through Trixx / CCC Underclocking through Trixx / CCC Please help!!!
Well given that cards reputation normally I'd say RMA it, however in your case the PSU could be to blame. It's an unknown brand lacking 80+ cert so something tells me it was sitting on the cheap & nasty shelf when you bought it. However having said that, a bad PSU usually restarts the PC instead of just causing a black screen - does it ever shut down completely? If it was me I'd open a ticket with Sapphire first.....
Yeah I really don't know anything about the PSU or the brand, though living in Korea there are so many brands here that I've never heard of before. I did find this page about it though, and gleaned the following info from that site - http://img.danawa.com/prod_img/500000/727/517/desc/prod_1517727/add_1/RIO R-6000LW V2.jpg ATX 12v v2.3 Natural RoHS ?? DC Output - +3.3v / 30A , +5v / 30A , +12V1 / 18A , +12V2 / 18A , -12V / 0.8A , +5Vsb / 3A As for system shutdowns - no, I've fortunately never had one, and have been using this PSU for over a year now, and it was powering a Radeon HD6850 before I swapped in this 7850 around August/September last year. What feels weird to me is just how sudden this issue cropped up, without any apparent triggers/changes to my system. Also, I filled out a Sapphire support ticket 3/4 days ago (no replies yet) and posted in the Sapphire support forums the day after - also no replies...
Yeah Sapphire can be pretty slack at replying.. I opened a ticket some time ago and it took 2 weeks to get an answer. Might pay to remind them again.... They have acknowledged a black screen issue with the 7870's so it probably affects the 7850 as well.
I see - then are you saying that it's unlikely to be the PSU? I dont want to install a RMA card or a new card (seriously considering going nvidia now) and bork that one too...
Oh so you've had the 7850 since August? Atm it's hard to say between GPU or PSU but one of them is to blame. Run OCCT PSU test and check the 12v rail log.
I just finished a one hour OCCT PSU stress test and got the following results No errors but the 12v rail did fluctuate between 11.43 and 11.54... though I've no idea what that really means or how significant that is... ^^;;
I think you've discovered the source of your black screens lol. Those 12v rail numbers are appalling.
+11.40 V to +12.60 V is within the +/- 5% tolerance for ATX spec. Also, software is notoriously inaccurate for PSU voltage readings. You need to test with a multimeter to make sure that is the actual voltage your getting.
Well it's accurate enough to know when there is a problem. The sortware may report it but it's reading from hardware.... Even without a multimeter it's obvious his 12v rail is bad since even with no load it read 11.5v - and this was only running in windowed mode which draws nowhere near the power of running fullscreen. OP run the same test again in fullscreen DX11 and see what happens....I bet your rig crashes.
I don't have a multimeter to hand so in the meantime I'll try again at DX11 fullscreen for an hour. Assuming something does go wrong or looks wrong with the PSU, would it then be safe to just replace it? Or could the graphics card have been damaged by the PSU too..?
Nah the card should be fine... It might pay to check with another application though, like AIDA64 just in case the reading is low like PhazeDelta suggested. If you go into AIDA64 preferences and enable shared memory it will send readings to OCCT (select AIDA64 as monitor in OCCT options). Oh btw OCCT has a GPU error test you can run if you want.
Thanks so much for all your help, PhazeDekta and PillMonster. Ive been struggling alone and completely unskilled thr past week or so and it finally feels like Im close to an answer. anyway, I just tested again...three times.. I set Aida64 as monitor like you said but forgot to enable fullscreen, so I started again and within 3seconds of the initial idle period, my PC hung, completely unresponsive and I restarted. To make sure it wasnt a fluke, I tried another scan, same settings and it crashed in the same way at 1597 seconds. Oh and according to Aida64, the 12v voltage varies between 12.25 (idle) and 11.88 (the last value when it crashed). Unfortunately Ive no logs to show you but its clear my PC doesnt like PSU tests. Im doing a GPU test now (windowed dx11 aida errorchecking) ... Does this mean its just the PSU at fault?
Any way you can borrow a power supply from a friend and test your card on that? IF it does the same thing on that power supply, then the card needs to be sent back.
Well that's alright it just means a few tests might be needed to determine the cause. Worst case scenario you could RMA the card and see if the problems persist, then replace the PSU if needed. I'm guessing a bad PSU since black screen crashes tend to show up straight away, even if the card is new. A crappy power supply though could run OK for a while before going bad.... Running the error check in OCCT wouldn't hurt. Then maybe we could get a sensor log with GPU-Z later too.
I see - but which error check in OCCT do you mean? Also, I've done GPU-Z sensor tests whilst playing Far Cry 3 and even when I crash to a black screen the sensor logs don't show any change. e.g. GPU Core Clock [MHz] , GPU Memory Clock [MHz] , GPU Temperature [캜] , Fan Speed (%) [%] , Fan Speed (RPM) [RPM] , GPU Load [%] , Memory Usage (Dedicated) [MB] , Memory Usage (Dynamic) [MB] , VDDC [V] 2013-04-06 08:14:30 (in-game, no visible issues) - 920.0, 1250.0, 61.0, 50, 2990, 90, 734, 61, 1.138 2013-04-06 08:15:29 (just prior to crashing) - 920.0, 1250.0, 59.0, 50, 2995, 94, 734, 61, 1.138
Hey I just reread your 1st post and noticed the driver version. There have been reports of random black screens occurring with the early 13.1 drivers, I even experienced it on my own 7950. Upgrade to the latest set (13.3 b3) and test again.