Hi, I have an Hd Radeon 7870, with my drivers up to date, and I've been having signal issues since I purchased my PC a year and a half ago. My problem is the following, at random times my screen freezes for a second, the music stops and the signal is lost, then the screen turns black and then it recovers and a message says that amd software stopped working but it was recover appears on the screen. Also, after some time playing a game (it doesn't matter wich one) the screen freezes the sound is lost and inmidiately the screen goes black, the PC still works, I can hear it working and if I turn it off and restart everything works fine, but this happens randomly some times is after five minutes of playing some times after an hour... Same as with the sudden losses of video, and is pretty weird because i have an HDMI switcher connected with a console and a TV reciever and this problem doesn't happen with them, well the TV reciever has lost the signal very few times but that isn't related to the PC issue. Things I've tried: -Connecting it without the HDMI switcher -Two different HDMI switchers, one with power supply -Different cables -Different drivers (I have updated my drivers a couple of times since I purchased the PC) -Different HDD ( I have an SSD now) -Changed the motherboard I haven't overclocked my system, but I have tried the OCTT test and the signal was completely loss. I honestly don't have any idea of what I can do now, I've seen a couple of threads with problems like mine, but there is never a definitely solution, so I hope you people can help me. Thanks.
Hi Extraordinary, thanks, I'm going to do it now, would you mind telling me whats the purpose of that change? thanks
With you saying that the screen goes blank and then the driver recovers / stopped responding and recovered, those are the same symptoms as an unstable overclock My old 7870 was factory OCd - Im wondering if yours is too that it is not stable at the OC it came with, raising the vCore can stabilize an unstable OC Just give it +10-15mV to start with You could also try dropping the Core clocks down a bit and see if that stabilizes it too
Settings > Unlock Voltage Control - make sure you're using the latest Beta from here http://forums.guru3d.com/showthread.php?t=387818
Ok the voltages was: Core voltage 1138 by default, now is 1250 Memory clock was 1208 now is 1000 (note here that I honestly have no idea of what i'm doing. Now that I've changed the voltages shouls I run OCCT or some similar program to see if the issue persist? Thanks again
You can leave the clocks as they were, just raise the Core Voltage slightly then run the same tests that was crashing it earlier Best to change one thing at a time or you won't know what fixed it if it works
Ok I have changed the core voltage to 1300 (I can't put it higher) I'm going to run the 2hour gpu stress test of OCCT that gave me the error yesterday, I'll keep this post updated
1300 is a bit high for stock clocks if 1138 was default voltage 1200 would be more than enough to stabilize stock - too high can cause problems too, and heat
Ok at 1300 the test has stopped in 2 minutes so, yeah, I better lower it a bit. Oh, my fault, when I press the reset button in MSI Afterburner (I suppose that reset is there to leave the default settings) It leaves ir at 1219 not 1138, my fault, so using 1300 maybe wasn't that high afetr all. Should I lower the clock?
If default is 1219 - raise to 1250 Like I say 10-15mV will likely do if that is the problem - GPUs are very sensitive to voltage increase, even though to you and me 1.219v to 1.250v is almost nothing, to a GPU it is a decent increase and enough to stabilize OCs etc Just leave everything as it is when you hit reset, then raise Core Voltage to 1250 - set it at that and test it
Just do what you normally do when you get the problem, play some games, you could run things like Kombuster and Furmark to stress the GPU Yea set Afterburner to start up with Windows and load those settings, you can save them to a profile at the bottom of Afterburner, if you tell it to load settings at boot you shouldn't have to mess with it any more unless you want to overclock your GPU Depends on the rest of your system to how much power you need from your PSU, Id say 500w is borderline Do a calculation with the rest of your system here http://extreme.outervision.com/psucalculatorlite.jsp EDIT - Raise the vCore slightly more? Check event logs and see if anything else is going wrong Check the PSU calculator Test the GPU in another machine Anyone else got any ideas?
I have entered all the data in the PSU calculator: -Intel Core i5-3570 3.4Ghz Box -Motherboard P8Z77-V -Arctic Cooling Freezer i30 CPU Cooler -G.Skill Ares DDR3 1600 PC3-12800 8GB 2x4GB CL9 -Seagate Barracuda 7200.14 1TB SATA3 -Samsung 840 EVO 120GB SSD -Antec Three Hundred Two -Antec VP550P Basiq 550W -Sapphire Radeon HD 7870 GHz OC Edition 2GB GDDR5 -LG GH24NS90 DVD 24X -Logitech Desktop MK120 -Cooler Master 120x120mm 1200RPM And acordingly to PSU calculator and if everything is ok I shouldn't need more than 400W of power to run the PC (maybe I'm mistaken my knowledge is limited). I'm going to try rising 10 points the vcore and lets see if that works, so far changing it to 1250 was enough to keep it alive an hour and a half. I'm going to record it using speedfan log again. I have the results of the last test for the GPU temp and I can say that it was stable between 50-52ºC with some random peaks of 60-63ºC (and I say some because they weren't frequent. EDIT: It even reached 70ºC EDIT2: Raised it 10 points, the signal was lost at 1h:40min, the temp was constantly at 50-52ºC but just after the signal was lost and I had to turn off the PC it seems that it raised up to 60-65ºC
I got ideas....but doubt this guy wants to hear them.... I'd grab a better power supply and RMA the graphics card while possible.