Core i7 42 C at idle with thermalright ultra 120
Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee
Core i7 42 C at idle with thermalright ultra 120
I kind of don't like this result, and wonder if something i can do to bring it down.
I used arctic silver 5, is it the setting time they have mentioned in the instruction that i have to wait? even at 38 after setting time, it is still kind of high for me.
I applied arctic silver 5 on both CPU and heatsink, now i think of it (and other turorials), they always seem only apply arctic silver 5 on cpu, applying on both CPU and heatsink, is it too much?
And btw the CPU fan speed is at max ( the fan come with my thermalright ultra 120 extreme 1366 rt is not pwm fan , what a dissapointment)
I used arctic silver 5, is it the setting time they have mentioned in the instruction that i have to wait? even at 38 after setting time, it is still kind of high for me.
I applied arctic silver 5 on both CPU and heatsink, now i think of it (and other turorials), they always seem only apply arctic silver 5 on cpu, applying on both CPU and heatsink, is it too much?
And btw the CPU fan speed is at max ( the fan come with my thermalright ultra 120 extreme 1366 rt is not pwm fan , what a dissapointment)
You may have put on too much TIM if you put it on "both" the CPU and the heatsink. This will affect your temps.
Redo it, fully wiping off the TIM, and put a very thin layer only on the CPU. You can use the line method that the instructions describe, or another method, but make sure the end result is a very thin layer of TIM.
Redo it, fully wiping off the TIM, and put a very thin layer only on the CPU. You can use the line method that the instructions describe, or another method, but make sure the end result is a very thin layer of TIM.
i just did that, take it out, clean up and reapply only to CPU, the result is the same(actually maybe little bit worse).
I also noticed that however, my thermalright cooler can be "twist" left and right a little bit even when all 4 spring screws fully tightened, is it by design or my install is at fault?
another side questoin, i use speedfan to read the temperature of mobo, what is the diff between CPU and Core* temperature?
why my CPU temperate is almost always 10+ higher than Core?
thanks in advance
I also noticed that however, my thermalright cooler can be "twist" left and right a little bit even when all 4 spring screws fully tightened, is it by design or my install is at fault?
another side questoin, i use speedfan to read the temperature of mobo, what is the diff between CPU and Core* temperature?
why my CPU temperate is almost always 10+ higher than Core?
thanks in advance
As for your question on core temps verses CPU, you should look at your core temps and not the CPU temps. Your CPU temps should be lower than the temps of the cores, so perhaps it's the way you're reading the temps. But regardless, I recommend you use the CoreTemp application to look at the temps of the cores themselves, not speedfan.
As for your actual temps, the real test is what are your temps at load, not idle. If there is a real problem, that's how you'll know. The idle temps are somewhat irelevant and can be misleading. Suggest you load up Prime95 on all 8 cores and see what CoreTemp reports.
(One thing about the TRUE - it normally needs to be lapped to get the best performance.)
As for your actual temps, the real test is what are your temps at load, not idle. If there is a real problem, that's how you'll know. The idle temps are somewhat irelevant and can be misleading. Suggest you load up Prime95 on all 8 cores and see what CoreTemp reports.
(One thing about the TRUE - it normally needs to be lapped to get the best performance.)
Take RealTemp to monitor your core temperatures, load your CPU with prime95 (HT on) and tell us your load temps after 15-30 minutes of prime95.
Also TRUE without lapping and washer mod is just an ordinary heatsink, good but not stellar.
Also TRUE without lapping and washer mod is just an ordinary heatsink, good but not stellar.
thanks guys, ran prime95 for more than 15 mins, coreTemp applicaiton shows all core# range from 69 to 71 C.
My mobo's (asus p6t) monitor software report CPU temperature of 75C.
is that normal for stock TRUE? remember i have CPU fan at max speed, it is very load, i will definitly want to lower the speed for my taste.
also since you guys have exp with TRUE, can you "twist" the heat sink slightly left and right? or is it truly secure in fix location? I just want to make sure my install is okay.
My mobo's (asus p6t) monitor software report CPU temperature of 75C.
is that normal for stock TRUE? remember i have CPU fan at max speed, it is very load, i will definitly want to lower the speed for my taste.
also since you guys have exp with TRUE, can you "twist" the heat sink slightly left and right? or is it truly secure in fix location? I just want to make sure my install is okay.
Mine idles about 42 on average. That's however at 4.2ghz and 1.2v with a TRUE 120 with the asus fan manager on quiet. Not sure of RPM but I can't hear it with the case closed.
Last edited by Hellspawn on Sun Aug 02, 2009 4:29 pm, edited 1 time in total.
Yes it is able to be moved left and right rotation somewhat. I'd prefer that it be locked down, but...fatdragon wrote:thanks guys, ran prime95 for more than 15 mins, coreTemp applicaiton shows all core# range from 69 to 71 C.
My mobo's (asus p6t) monitor software report CPU temperature of 75C.
is that normal for stock TRUE? remember i have CPU fan at max speed, it is very load, i will definitly want to lower the speed for my taste.
also since you guys have exp with TRUE, can you "twist" the heat sink slightly left and right? or is it truly secure in fix location? I just want to make sure my install is okay.
Well load temps of 71 on an i7 aren't terrible. But they aren't great either especially at stock and the fan on max. I suppose that your temps remain the same even when you press down on the TRUE and apply pressure while checking your temps to ensure that it isn't a mounting problem?
I think that with an active 1200rpm fan and with a mild overclock mine were in the high 60's, but I had also undervolted as well.
Maybe you got unlucky with your CPU, but more likely, assuming you've done everything else properly with the TIM and the mounting, that heatsink needs to be lapped. So you may either just have to put up with this, lap your TRUE or exchange it for a new heatsink like the Prolimatech Megahalems.
I think that with an active 1200rpm fan and with a mild overclock mine were in the high 60's, but I had also undervolted as well.
Maybe you got unlucky with your CPU, but more likely, assuming you've done everything else properly with the TIM and the mounting, that heatsink needs to be lapped. So you may either just have to put up with this, lap your TRUE or exchange it for a new heatsink like the Prolimatech Megahalems.
According to Chris Hare's web site, Core i7 is supposed to operate at a maximum "cover temperature" of 67.9 degrees Centigrade. So I think 71 is no good.
Are you using the "Extreme" version of Thermalright Ultra 120? I have been doing some calculations for my own Core i7 build, and they show that this cooler should give you a max of 50 degrees Centigrade at maximum load. So something is wrong.
I am planning to go for Prolima Megahalems myself.
Cheers,
Uday
Are you using the "Extreme" version of Thermalright Ultra 120? I have been doing some calculations for my own Core i7 build, and they show that this cooler should give you a max of 50 degrees Centigrade at maximum load. So something is wrong.
I am planning to go for Prolima Megahalems myself.
Cheers,
Uday
The SPCR tests of Ultra 120 Extrme with stock fan at 12V showed a thermal rise of 0.33C per Watt of CPU power. So, if the benchmark programs managed to use the full 130W of Core i7 processing power, then the temperature should still rise by only only 20C over the ambient. 70C, or even 62C, means to me that the cooling system is not functioning as intended.Kaleid wrote:Bear in mind that programs like prime95 put a heavy stress on the system which is unlikely to be matched even closely by real usage.
My CPU hits about 70C with prime95. But I have never seen it go over 62C when encoding or gaming.
Doesn't the power consumption increase with overclocking, especially with OCs > +50%?reddyuday wrote:The SPCR tests of Ultra 120 Extrme with stock fan at 12V showed a thermal rise of 0.33C per Watt of CPU power. So, if the benchmark programs managed to use the full 130W of Core i7 processing power, then the temperature should still rise by only only 20C over the ambient. 70C, or even 62C, means to me that the cooling system is not functioning as intended.Kaleid wrote:Bear in mind that programs like prime95 put a heavy stress on the system which is unlikely to be matched even closely by real usage.
My CPU hits about 70C with prime95. But I have never seen it go over 62C when encoding or gaming.
If you think you are overloading the CPU by 50% through overclocking, you can expect 30C over the ambient. That still doesn't give you 70C unless your ambient temperature is 40C.Rucker wrote: Doesn't the power consumption increase with overclocking, especially with OCs > +50%?
There are tons of reports of overheating i7's around the net. My best guess is that some of these coolers are not functioning correctly at higher loads. Even though SPCR stress-test the coolers with a Pentium D processor that has a similar thermal envelope to i7's, the CPU power consumption in their tests has only been 67W, i.e., half the rated power. Perhaps SPCR should do a stress test with i7's to see which of their coolers can really deliver.
Higher temps do not necessarily mean that the coolers should do worse. The basic Physics says that when the temperature gradient is steeper, heat transfer should take place faster. I would expect that the stock Intel coolers which don't use heat pipes (or do they?) won't be fazed by higher temps. But the heat pump coolers, which use hydro dynamics to transfer heat, can run into turbulence effects at higher temperature gradients.
I think people should calculate the expected temps from the figures given by SPCR reviews, and return the coolers that do significantly worse than that as being "defective". 130W is not an out-of-the-world power consumption for a CPU. The coolers should be able to handle it. It will be good to document in this forum which coolers are prone to being defective.
As K.Murx has pointed out in the other thread, one should also check the cooler temperature. If the CPU is hot and the cooler is also hot, then it is an air flow problem. If the CPU is hot and the cooler is NOT hot, then it is a cooler problem.
Uday
I have an original C1 i7 920 and P6T Deluxe, with a TRUE and nexus pwm fan at 900rpm.. and mine idles at 38C with the cpu stock, get's to 69C when encoding x.264 videos for an hour or so. That's actual temp, not mobo reported temp.
Using CPU-Z to find out what revision your cpu is would help greatly.
Using CPU-Z to find out what revision your cpu is would help greatly.