Export thread

So I built a mini-iTX gaming system not too long ago, and for the most part it has been great. However, I'm now concerned that I may have a fan problem, but various tools do not agree with one another.

There are two fans in my case (a Lian Li), a front side 140mm fan and a 120mm top fan, both in-take, with vents opposite.

On top of this, I have the stock CPU fan that came with my i5-3570K.

I first noticed a few days ago, that the top fan only seems to run intermittently - either visibly slow or sometimes not even running, so I ran the ASRock M/B Tuner that came with it, and it told me that the M/B was running at 34C and the CPU at 40-ish, which is far from bad but hot enough (at less than 5% load) that I was wondering if something was wrong with the fan controller.

Just to get a second opinion, I downloaded HWMonitor and Speedfan to see what they said, and here's where it got weird. Both of them said that my CPU was running at 57C at almost no load! I was about to assume that the stock fan must be dying, but then I noticed that the CPU core temps (and the package temp) listed by HWMonitor were actually lower than the overall CPU temp listed, and were actually more or less in line with the ASRock utility.

So, I'm still confused about why my top fan is turning off, but now I'm concerned about the CPU stock fan, because 57C is a pretty serious temp at to be running at more or less idle, but the tools don't agree.

Thoughts?

Slight update: Playing the Sleeping Dogs demo with everything turned up for about 10 mins resulted in a "CPUTIN" temp of 127C, and a Core "Package" temp of 81C according to HWMonitor. CPU Load was around 50%

Thinking it might be averaging your cores, maybe? Dunno enough about the situation to guess at it, I assume the sensors are technically all correct somehow, just not sure how they are arriving at their numbers (moving average? Instantaneous?). Sounds like something is definitely up, though. Sorry I can't think of a way to be of more help right off the bat.

--Patrick

Reading up a bit more, it seems like w/e is in the "Value" column in HWMonitor is the real-time temp. HWMonitor's "CPUTIN" temp value in that column is always the same as the "CPU" temp value in Speedfan, while the Core Package temp value in that same column in HWMonitor is the same as the "CPU Temperature" in AXTU.

Various places seem to say that the temp of the individual cores should always be slightly hotter than the temp readout of the "CPU", so it appears some sensor is either not working right, or some software needs to be re-configured.

And while it's the top fan that spins down and stops, all the tools agree on the M/B temp being 35C at idle, which is a tad warm but not overly so. So maybe the chassis fan doing so is normal, even if I don't know which temp read-out to trust for the CPU.

What I would focus on is the individual core temperatures. They're the ones you can get an actual reading off without knowing how they formulate any sort of aggregate temperature. If one of your cores is running at critical temperatures and none of the others aren't it would lead one to believe that it's taking too much voltage. If they're uniformly running too hot, it would point to a cooling issue.
Also, if your CPU was running at 127C, that's WAY above critical ranges, which are anywhere from 60-100C degrees depending on your CPU if I recall correctly.. The last time I had running temps that high, the fins on my CPU needed a serious cleaning. Dust had settled in between the cooling fins and eventually started gunking up the fan.

*note to self... spray out the case when I get home... Especially the radiator for the CPU*

Bowielee said:

Also, if your CPU was running at 127C, that's WAY above critical ranges, which are anywhere from 60-100C degrees depending on your CPU if I recall correctly.

Click to expand...

Critical ranges on even the recent Sandy Bridge Extreme processors is still only about 70C for the chip, not sure where to find stats for the individual cores.

--Patrick

Bowielee said:

What I would focus on is the individual core temperatures. They're the ones you can get an actual reading off without knowing how they formulate any sort of aggregate temperature. If one of your cores is running at critical temperatures and none of the others aren't it would lead one to believe that it's taking too much voltage. If they're uniformly running too hot, it would point to a cooling issue.
Also, if your CPU was running at 127C, that's WAY above critical ranges, which are anywhere from 60-100C degrees depending on your CPU if I recall correctly.. The last time I had running temps that high, the fins on my CPU needed a serious cleaning. Dust had settled in between the cooling fins and eventually started gunking up the fan.

Click to expand...

I'll air-spray it out, but if it's the core temps I should be worried about, I probably don't have much of problem since they're mostly pretty reasonable and within a few degrees of one another. Maybe switching to a non-stock CPU cooler will do the trick.

EDIT: Also, they're all a good 60-70 degrees C short of the TJ Max, so they're well and away low enough to be acceptable, if still a tad warm for the load.

PatrThom said:

Critical ranges on even the recent Sandy Bridge Extreme processors is still only about 70C for the chip, not sure where to find stats for the individual cores.

--Patrick

Click to expand...

I believe laptop CPUs have a higher critical range... I can't remember why that's the case or where I heard that, but I recall hearing it at some point in time...

probably...

maybe...

don't quote me.

For the folks in the know, since I know very little about non-stock cooling fans, what's the difference between the CoolerMaster 212 EVO and the 212 Plus? The 212 series in general seems to be the go-to of choice for 1155 CPUs, but we're beyond my understanding of the field now.

Evo has slightly higher CFM, would need to dig further to see if there are any other big differences. At work now, can't dig.

--Patrick