Sign in with
Sign up | Sign in
Your question

First overclock temperature readings - higher than expected?

Last response: in Overclocking
Share
April 16, 2012 2:12:49 AM

I just tried out my first overclock of my i5-2500k with Cooler Master 212 Hyper Plus at 3.60 GHz (I'm starting out with baby steps). My max core temp readings on HWMonitor on full load with Prime95 came out to around 56 C. Idle temps range from around 30 C to 33 C across all cores. Based on other temp readings that I've come across, this seems a little high for a mild overclock with an aftermarket cooler. This was my first build from a while back, and despite the fact that I put a good bit of research into proper thermal compound placement, I'm wondering if maybe I did a poor job of it. I'm also curious about the fact that my FANPWM0-2 percentages are all at 0% on full load, given that I seem to recall different settings around the time of my build (could be mistaken, though). Basically, are these temp readings higher than expected, and should I look into those FANPWM readings?
a b K Overclocking
April 16, 2012 2:28:12 AM

1) check to make sure your fan profile is set to a more aggressive cooling ratio.
2) how is the airflow in your case?
3) temps look pretty good, but my guess is the cpu fans are set to cool at a certain threshold and you haven't pushed the chip far enough to hit that yet. Remember that systems provide more cooling to the cpu the hotter it gets.
a b à CPUs
a b K Overclocking
April 16, 2012 2:58:48 AM

Please describe your TC methods.

Check that all your fans are running! If they are, you can safely ignore the 0% readouts.
Related resources
April 16, 2012 4:23:00 AM

kelthic said:
1) check to make sure your fan profile is set to a more aggressive cooling ratio.
2) how is the airflow in your case?
3) temps look pretty good, but my guess is the cpu fans are set to cool at a certain threshold and you haven't pushed the chip far enough to hit that yet. Remember that systems provide more cooling to the cpu the hotter it gets.


Looks like my fans are under normal control by the motherboard, I can manually adjust the settings but the default options are low, normal and disabled (basically free to run at max power all of the time). All of my fans are running at apparently normal rates, it's just the FANPWM that's constantly 0%.

Airflow in my case should be pretty decent; I'm sure my cable management isn't the best, but I've got a 120mm case fan in front of the HDD bay, a fan on the Hyper 212 pulling air through the heatsink, a side fan (80mm, I think), and another fan (also 80mm IIRC) in the rear top of the case. I do have an AMD 6950 in there in the top PCI slot, but the temps for that maxed out at around 58 C.

Something curious happened when I upped the multiplier. It ran at 3.8 just fine, maxing at around 57 C on full load, but when I upped it to 40x, the system failed to boot intially and reset the multiplier to stock. I bumped it back up to 39x and it ran fine, package temp maxing out at about 60 C and none of the core readings above 58 C. I didn't meddle with any voltages, so I'm not sure why it failed to even start up at 4.0 GHz. Still no change in FANPWM, but apparently that's not much of a concern (and possibly a good sign).

kajabla, sorry, what do you mean by 'TC methods'?
a b à CPUs
a b K Overclocking
April 16, 2012 10:42:19 PM

TC = thermal compound. You say you think your methods were good, but I'd like to check that.

No 2500K should fail to hit 4.0ghz.

Is the side fan pointing in like it should be? Stick it in the front if you can.
a b K Overclocking
April 16, 2012 10:51:36 PM

Quote:
bumped it back up to 39x and it ran fine, package temp maxing out at about 60 C and none of the core readings above 58 C


Something is wrong here. How much Thermal Paste did you use? There should also be no play in your heat sink. All push pins should be fully inserted to make sure that there is good contact between your cpu cooler and cpu.
a b à CPUs
a b K Overclocking
April 17, 2012 2:13:27 AM

The 212+ does not have push pins. Only the awful stock Intel sinks have pushpins.
a b K Overclocking
April 17, 2012 2:20:04 AM

kajabla said:
The 212+ does not have push pins. Only the awful stock Intel sinks have pushpins.


true. sorry about that.
April 17, 2012 3:13:48 AM

I'm not sure if I found the article on here or anandtech (might've linked to a different site from here), but the article suggested parallel lines on the copper piping of a cooler like mine for the best spread. It's been months since I built the system, but it's very possible that I was too generous with the arctic silver. I don't know why it failed to boot at 40x, I would've thought that'd be more of an issue if I'd played with the voltage (I haven't touched any of that) or if it was really unstable at 39x. If my temps are in fact significantly higher than expected of my set-up, I might have to reseat the cooler and reapply the compound when I get a chance (some of my building materials are out of town).
April 17, 2012 7:21:18 AM

Update: I set the multiplier to 40x again, this time it's had no problems booting up. Max temps after a few hours of prime95 are about 58 C on a particular core and 61 C on package temp. Still no movement in FANPWM but my CPU fan speed has definitely peaked at a few hundred RPMs more than at 36x, I'm assuming it's just the way my BIOS is set to control fan speed. I'd imagine my temps are still higher than expected, so I won't push it higher unless I decide to reseat the cooler (and maybe it'd be best for long term processor health if I did that regardless).
April 17, 2012 6:18:08 PM

Nah, I built this thing about a year ago, thinking back on it. It's almost certainly been running over 300 hours.
!