delly Posted October 26, 2013 Share Posted October 26, 2013 So it all worked beautifully after hacking the registry for Win 8.1 support, I spent hours creating custom curves (v2.4.5110), adding profiles, beholding my beautiful creation! Then reboot, play game, BSOD after 2 minutes. WTF. Reboot again and load up Corsair to check fans are responding to my custom curves.. Not budging, temperature hits 70, rpm just sits there at 900rpm! And no its not the wrong group before you start typing. So I read all the support forums, uninstalled every other AI Suite or program that might even be thinking about monitoring fan speeds but to no avail. Create new profile, ah.. now it works again. So then I tried changing the fan group to water temp and then back again to cpu avg temp and it springs into life! So the workaround is everytime I start windows I have to open corsair and for each fan flick the groups around. Looks like a bug to me, incidentally restarting service did not fix. Another thing that might be relevant is I am using the R4 case and its hardware fan controller with 3 settings, this switch has no effect either (apart from preventing the GTX being controllable with CL). One last thing although its not specifically relevant to this issue, the avg core temp will spike by 10deg every few seconds under load (haswell?) resulting in the fans spooling up (when it works :p:) and down constantly, is there anyway you could implement a kind of smoothing algorithm that perhaps allowed configuration of the duration that a temp would need to remain to trigger rpm changes? Its like my machine is schizophreniac! And no I cant use the water temp, it is just too restrictive (and dangerous). Screenshot: I have applied group flick fiddle (GFF) on the second fan (Fan 4) and you can see it is responding (it has roughly the same curve). Last thing please add a way to copy a custom curve it is so painful having to copy them. http://img43.imageshack.us/img43/6518/g1v6.png Link to comment Share on other sites More sharing options...
peanutz94 Posted October 26, 2013 Share Posted October 26, 2013 Have you tried dropping the CL program into your start menu? I know it's probably a dumb question, but i had to ask. Link to comment Share on other sites More sharing options...
delly Posted October 27, 2013 Author Share Posted October 27, 2013 Thanks for at least replying peanutz.. I didnt make absolutely clear but yea its running at startup but wont engage a CPU Temp based custom curve (water temp works). Hence my trick to flick the groups I described which triggers the curve.. This bug gets through testing because its a pain to test restarting windows, a quick manual test will show that creating a CPU temp curve works fine. Classic bug. Its a bit of a buggy app, I get that, but really as a programmer myself this is 'mission critical' stuff and there should be some intermittent polling at the very least to detect error situations and 'in code' force the curve to be active, it is not ok to just call the curve activation code at UI hook points or service startup, anyone who has written windows code knows this is dangerous.. there should be no chance the code can happily churn along with the crosshair way off and temps rising (see pic) without at least a warning or something. This bug needs a ticket with the developer ASAP or you could damage peoples hardware. For anyone else facing this problem then I can confirm that this bug does not affect Water Temps, I am using this till the bug is fixed, perhaps its better anyway, its certainly smoother. I'd bet my bottom dollar its a coding/threading bug.. hmm might just have to crack out the old .net Reflector to check this dodgy code. Link to comment Share on other sites More sharing options...
wytnyt Posted October 27, 2013 Share Posted October 27, 2013 regarding your 10 degree spike temps-this is normal as the cpu is simply processing commands it will always do this and the fan fluctuation is due to this and having the fan curve too sensitive this and using ''load'' to spin fans will produce same results these coolers are designed to use the h80I water temp to curve the fans as the cpu is just too erratic and controlling temp of the water will always produce better cooling results... Link to comment Share on other sites More sharing options...
peanutz94 Posted October 28, 2013 Share Posted October 28, 2013 egarding your 10 degree spike temps-this is normal as the cpu is simply processing commands Not at full loads wytnyt. Idle temps sure, but a 10c spike every few seconds at full load isn't right. They should all be close to the same, within reason, if they are all running wide open during a stress test. Link to comment Share on other sites More sharing options...
wytnyt Posted October 28, 2013 Share Posted October 28, 2013 thought it was at idle but rereading shows under load yeah that isnt normal,thx for the catch Link to comment Share on other sites More sharing options...
delly Posted October 28, 2013 Author Share Posted October 28, 2013 regarding your 10 degree spike temps-this is normal as the cpu is simply processing commands it will always do this and the fan fluctuation is due to this and having the fan curve too sensitive.. 1. the curve is not too sensitive, it was the 10 degree spikes that caused the rpm change, if the temp goes from 50 to 60 and back to 50 for a fraction of a second I don't consider that an average even if it is whats being reported but I do admit its a difficult situation to cater for. A dampening option is a good idea IMO. 2. as has already been pointed out these fluctuations are happening under load, so I did some further testing completely removing all overclocking and the spikes are gone... I suspect it could be adaptive voltage, I've had quite poor system stability with my 4.5Ghz OC even when its at low temp it will crash. I'm reading now that Haswell has OC memory issues. Link to comment Share on other sites More sharing options...
Recommended Posts
Archived
This topic is now archived and is closed to further replies.