PDA

View Full Version : same CPU temps idle & load (?)



pb126
01-04-2010, 12:10 AM
I ran the Real Temp test program with prime 95. I have a quad core 9450 build, built with help from the SuperPC forums. The 4 core temps never changed one iota during the test run. I recently changed from the stock CPU cooler to a new Arctic Freezer 7 cooler (after a year with the stock cooler). The temps are reading 41,41,37,37 and have never changed in the week that the PC has been on since the cooler change. It's conceivable I didn't get the cooler pins in right, they seemed to have trouble getting locked in place, although in the end it seemed I got it.
I assume something is wrong and the temps should change.
Any suggestions welcome.
Thank you and Happy New Year.
Regards,
pb126

The Wise Monkey
01-04-2010, 03:12 AM
That seems an issue with the software rather than the hardware installation. Check the temps directly in the BIOS or use a different programs such as Core Temp or SpeedFan to confirm.

pb126
01-04-2010, 03:52 AM
Thank you Wise Monkey. Core Temp and Speedfan give me the exact same numbers, 41,41,37,37. The BIOS gives me one number, 39.
Could it be my temperature sensors are "stuck"?
Added detail: the mobo is Asus P5KC and I am running W7 32.
Thanks.

The Wise Monkey
01-04-2010, 04:58 AM
Hmmm, strange. It could be that the sensors aren't working correctly, which could cause the problem.

You could always try disconnecting and re-connecting the fan, but don't worry too much about it if you can't get it to work - you should be able to tell how hot the CPU is roughly by listening to how fast the fan is spinning.

pb126
01-04-2010, 05:03 AM
Fan stays at 2550 with no variation whetehr idle or load. Was stuck there for the whole Prime95 test. I wonder if I did something strange during the cooler installation.

zburns
01-04-2010, 11:02 AM
Lets assume you have a hardware problem such as a defective sensor (that is effectively "shorted" and only has one output) that is feeding both "fan control" and "temp readout" including the BIOS temp readout. Then, seems to me, you would have your experience. When I say "defective sensor", I mean the sensor plus its associated "IC circuitry -- a shorted output would be in the circuits, not necessarily the "sensor". I said it the way I did for brevity.

The above would explain why you see what you see, if the assumption is correct -- ie. one sensor feeding fan and readouts.

As a further simple check, you could check temp readings in BIOS with system on idle, just turned on. Then use browser to go to large file sites (anything with large pictures or large photos). The temp should change immediately.

If the temp stays the same and the fan speed sounds "always same speed", first assumption above may be correct. Suggest you call Asus Support, get their comments. Ask how many sensors are on your mobo; then ask how you can get a reading from another sensor, if it is possible (or a fan running off another sensor -- I think I am reaching with this comment).

If you have experience using Task Manager, you can see your CPU % usage under idle and load. When the CPU shows some load increase instead of idle, the temp should be going up, at least a little.

pb126
01-04-2010, 11:07 AM
Thanks Zburns. Since it is the CPU temp which is never changing, wouldn't it be the Intel CPE sensors which would be defective and not the motherboard sensor? Thx.

zburns
01-04-2010, 01:18 PM
When I put CPE sensor into Google several different ways, did not get any "first order" or "first level" definitive statement that linked CPE to Intel or to your problem. Not sure that matters. My comments to you are simply my logic.

This paragraph inserted as an addition after I have finished this post; it is a long post just trying to help by explaining my reasoning. But this is a major point I should have mentioned first. Maybe you just have a poor
connection somewhere that is causing almost an open circuit; so just go back thru and check all connections. Did you bend any pins or something like that, etc.. But if you go back thru everything, make sure you are using static control.

The temps are read directly from your CPU, the electrical signals probably go thru some logic circuits on the Mobo. If fan speed is not changing and temp readings are not changing, the problem could be common to both fan speed and temp readings (obviously). If the CPU temp signals leave the CPU and go thru mobo logic circuits, to me it is more likely a mobo circuit fault (a shorted transistor -embedded in a logic circuit - that is forcing "a non changing output" to be seen). This is just a guess, but it is a resonable and logical guess.

If you think it is your CPU, then you have to ask, (1) is it more likely that CPU temp output has failed or (2) more likely that a mobo circuit fault has occured. I would think Intel's QC testing on CPU's is 100 % computer testing and probably faultless.

The one thing I did not mention as a cause is the possibility of static elec damage. This is the time of year when static electricity would be worst. If you exercised proper static control methods (the operative word would be "rigorous"), the odds are in your favor; however, if you did good static control some of the time but not all the time, then, assuming a circuit fault is the problem, it could be either the CPU or the mobo depending on "static control circumstances". If there is a circuit fault, whatever the component is, you will not know the cause (ie mfgs defect or some cause beyond the mfg).

Good static control means using the wrist strap as Rob describes; better to also have an anti static piece of foam(also grounded) as part of your anti static setup to lay your parts on as you handle them between the anti-static wrap and the computer case.

I am only saying "static" is a possibility; the individual builder knows what approach he used as he did the build. The really cold weather just happens to magnify the problem.

Back to the problem, the reason I suggested you call Asus, is what WM said about the fan increasing in speed as the load goes up. You say that is not happening. To me that says fan and temp readers are getting the signal from the same source. If that source is some integrated circuit, it is probably on the mobo, but the integrated circuit would get its input data from the CPU. If the CPU temp output is ok but the mobo int. circuit is "faulty", then your fan and 'CPU temp reader' would both be seeing a "no change in temperature" indication. The Asus tech support guy should instantly recognize that fact. OR, he (Asus Tech Spt) will have a better suggestion or approach.

All of the above also assumes your interconnect wiring is correct, etc.; but if you think about that, CPU is on the mobo, the circuits that feed data to your monitor are on the mobo. This means your problem, more likely, is confined to the mobo.

The Asus tech should have a better grasp of the problem than what I have said because he should know the mobo. Hope this is of help. Final Coment!! I guess in your post, you meant CPU sensors; same comments, regardless!!

zburns
01-04-2010, 03:04 PM
One other comment. Is your new system wired so that the fan speed is supposed to vary with temp? Does your mobo book say this!! The reason I bring this up is that on my 1.75 year old system, I am not sure that my fan varies with temp. I do not do games; and I do not really load up my system; my impression (based on my hearing) is that my CPU fan speed (Zalman) is constant. I assume this is the case, just is not an issue with me and I have not bothered to find out. (My Zalman is on a 3 way speed switch provided with the Zalman fan -- that's probably it, 3 speeds)

Sometimes when I put it in the sleep mode, something happens (noise level goes up; it is rare but it means a boot failure next time I boot up) and I hear a lot more noise which I assume is a fan ( I hope not a hard drive). This is a small and infrequent issue because otherwise my system runs well, consistent and predictable.

pb126
01-05-2010, 10:49 AM
zburns thanks for the extensive comments. Very helpful. Sorry about the CPE thing, I did mean CPU.
I spoke to Asus, following your post, and they figured I might indeed have an issue with the motherboard and the temp sensors. They gave me an RMA# so I will send the PK5C back. In the meantime I have a Gigabyte board (the EP45-UD3R) which I will try, with the same CPU I have now, and see what happens. I will post results once this is done. Thanks again for your help.