Overheating?

[Jomapil]

Hi there.

When I reboot my PC it appears the following warning:

" CPU Over temperature error! "

Must I take it to the technician or can I do anything?

[pepegot1]

If you had to ask the above question-take it to a Tech before it becomes a wreck!

[classfour]

It could be malware - many going around, killed two or three lately that were indicating hard drive failure.

If you don't know how to kill malware - take it to an expert.

[Jomapil]

Thank you pepegot1 and classfour for your suggestions.

Good afternoon!

[Nelson37]

Does this message appear BEFORE, or AFTER, the OS, or Windows, begins to load?

Did the PC shut down automatically or lock-up, immediately previous to this next boot-up?

How long, approximately, was the PC running before the last shut-down?

Are you running any temperature-monitoring software?

Simple test - Boot to the BIOS setup, usually by pressing F1 or F10, watch the screen. This will bring you to a menu, somewhere in there will be a "Monitoring" or "Health" section which will show you temperatures, WITHOUT Windows or any
software being involved. They should reach a maximum in 5-10 minutes, 120F is fine, 140F is getting a bit warm, 160F it's time to do something and you need to shut down. These vary a bit by processor but those are fairly close for yours. Cooler is always better. Usually the alarms in BIOS are set for higher than 160F, though some can have separate Warning and Shutdown levels. Sounds like your message is a Warning.

Start with simply blowing the dust out of the heat-sink and fans, a can of compressed air, vacuum cleaner hose, and clean paintbrush in alternation.

If temperature levels rise rapidly when you monitored them, then likely the CPU fan is not running properly, or may need to be re-attached. Report on temps and minutes to reach max temp, if it's a steady rise, etc.

If it runs just fine in the BIOS, for 10 minutes, then let it go for another 20 or so. If no problem, then that points to virus or faulty temp monitoring software. Boot to Safe Mode with Networking, DL, install, UPDATE, Update again, then Run Malwarebytes.

[Nelson37]

While in BIOS, get familiar with key functions and the EXIT option, most expecially EXIT WITHOUT SAVING. You do not want to make any changes in these screens, and most especially do not want to Save any. Just be careful and read the screens.

[SLK001]

Modern CPUs will not give you any warning about overheating. If the core temperature reaches a certain temp (~85C, IIRC), it will self-protect and just shut off - not safely, mind you, but like the plug is pulled (or battery removed). If you are getting warnings, you either have a temp monitoring application running, or malware.

Do you have a temperature monitoring application on your PC?

[lacywest]

Is the PC the one in your specs ... I can see ... if not ... what model is it ??

When do you see this message ... in WIN XP or when you first boot up in the bios post area ... what temperature is shown in BIOS ... if your BIOS shows that

[Bjs]

Leaving a system running in bios provides very little load on the processor ... thou some bios's are equipped with voltage monitoring ... if a specific V value is running higher than expected this can cause increased temperatures.

As the os loads greater resources are required which will cause temperature's to increase ... usual common faults are ailing psu, failing cpu fan, improper heatsink to cpu compound installation (or quality of).

Malware is doubtful due to the error.

[Nelson37]

This is incorrect. BIOS temperature loading will exceed that of all but the most thorough burn-in programs.

It will also show the designated temps for any temp alarms, if present.

There is an error MSG. It is not yet certain that there is any hardware error at all. If it ain't hardware, it's software.

[SLK001]

It appears that ASUS motherboards have this built into the BIOS - if a temp limit is triggered, the next boot will pop up this warning. This should cause the operator to carefully check to see if the warning is justified - temp limit properly set, proper airflow across the heatsink, fan working, proper use of thermal compound, etc.

This is a SERIOUS warning and should not be ignored. While it IS possible that some form of malware is manipulating the warning "flag", I would suspect a hardware issue.

[Jomapil]

Thanks Nelson37, SLK001, Iacywest and Bjs for your answers.

That warning was the first thing to appear on the display.

But since 20JUL it no more appeared. I switch off and on several times during the day, I have the PC on for 16 to 18 consecutive hours and nothing happens. Do you think there is no problem and it only was an error of OS?

[Nelson37]

It sounds like you are saying the msg appeared BEFORE Windows loads, so therefore it is NOT an error from Windows.

That means it must be coming from the BIOS, requiring you to check the alarm or warning settings there.

Basically, you need to do what I told you to do the first time.

If your CPU gets too hot, it could be totally destroyed. You must decide if this is a problem you want to take some action to prevent. Or not.

[lordsmurf]

My first action would be to remove and remount the CPU.
The old heat paste is probably powder.

And the fan may not be functioning properly. Watch it, see if it spins constantly.

Turn off the computer when not in use.

[Nelson37]

The first thing to do is to get some freaking actual numbers which will provide useful information to diagnose and solve the problem.

Enter the BIOS setup and read the damn temperature numbers! It's not rocket science.

Jonapil - I see the PC has run up to 16 hours with no problem since the incident. Do you understand that without supplying the info I requested, which was "how long had the PC been on before the error?", the 16-hour statement is TOTALLY meaningless and gives NO USEFUL DATA? You have an indication of a TEMPERATURE problem. TEMPERATURE is measured in NUMBERS. These are available with a few keystrokes. Get Them, Supply Them Here, or stop bothering people with guessing games.

[hello_hello]

I'd go into the BIOS after rebooting to see what the actual CPU temperature is, particularly after receiving one of those warning messages.

Under "normal" usage a CPU shouldn't get overly hot. Maybe try installing HWMonitor to look at the CPU temperature while Windows is running. Every so often something will go wrong with software or a device driver etc and it'll start utilizing the CPU way more than it should and for no apparent reason, it doesn't necessarily have to be malware... I recall once having to update the USB, ADSL modem drivers on someone's PC for that very reason. HWMonitor should also hopefully tell you if the CPU fan is working and how fast it's spinning.

My BIOS lets me adjust the temperature at which the motherboard sounds an alarm. Every so often on a warm day it'll hit the temperature it's currently set at while the CPU is working hard, converting a video or something like that, so I spend half an hour listening to the alarm going on and off. Every time it does go off I remind myself to increase the alarm temperature in the BIOS next time I reboot. One day I'll remember to do it....

[fritzi93]

Yeah, the temperature reading is right there in the BIOS.

The only time I encountered anything like the OP's problem, I had just cleaned out the case of an older computer (still have it). Apparently I jarred one of the heat sink fan posts loose. (Shitty Intel stock fan). The heat sink was still lightly contacting the CPU and it wasn't immediately obvious what had happened. Give a light pull to the heat sink fan, each of the four corners, and you'll know. And BTW, the heat paste was mostly dried up, naturally.

Re-applied good paste, re-attached the fan and it's been good for a couple years now. I have the shutdown temperature in the BIOS set at 60C. Every time I clean out that case I check that damn heat sink.

[Jomapil]

I rebooted the PC just now and I saw the temperatures are:

CPU temperature - 70 ºC ( But it is increasing ).
MB temperature - 34 ºC.

And is it possible to see the SETUP without reboot?

[Nelson37]

Shut it down, NOW!

No, it is not possible to get into BIOS without a reboot.

You have a definite and very real CPU overheating problem. The failure to check actual temps before this may very well have already done serious damage to the CPU.

Check fan RPM in BIOS, or use Speedfan in Windows. Possible it just isn't spinning fast enough.

Blow out the dust, check the fans, test, then re-install the heat sink with new paste if necessary. Check the mounts as mentioned above.

This reminds me of my step-daughter and her first car. Destroyed within 24 hours.

[SLK001]

A modern CPU will issue an immediate shutdown command (RE: turn off power) before it can be damaged. This is hardcoded in the machine code for the CPU. IIRC, the shutdown temp is around 85C.

Jomapil, either you have a bad CPU, or your cooling system has a flaw in it, or your voltages are way out of spec. The first thing that I would do would be to check the seating of the heatsink and the application of the thermal paste (should be Artic Silver).

[hello_hello]

Jomapil,
I skimmed through the post and couldn't see where you mentioned which CPU you have, but it should shut down or, in the case if Intel I think, slow itself down before it lets you cook it. Of course it'll still pay to get the temperature under control. If it was me......

First check the CPU heatsink is still mounted properly and the fan is working...
I'd install HWMonitor and SpeedFan. Both will report CPU temperature and CPU core temperature, which will normally be five to ten degrees higher than CPU temperature. If the difference is more than 10 degrees (at least for Intel CPUs) it's probably an indication your CPU isn't being cooled properly. It's a good idea to install two monitoring utilities to ensure they agree, then reboot and check the BIOS CPU temp to make sure they're both getting it right.
RealTemp is another CPU monitoring program designed for Intel CPUs and Intel's method of rating the maximum temperature of their CPUs.

Speedfan also shows CPU loading. If, as the temperature increases the CPU is working hard, and you're not actually running any CPU intensive tasks, then you need to find out what's keeping the CPU busy. If the CPU isn't working hard, then you probably need to re-seat the CPU heatsink or replace the fan.

For the record, Intel use a CPU temperature called TJ Max, which is supposed to be the maximum average temperature of the CPU cores... or something like that. Once TJ Max is reached, I think Intel CPUs begin to slow themselves down. RealTemp should show you the current temperature of each core, plus how much "headroom" you've got before you hit TJ max. According to RealTemp, the Intel CPUs in the two PCs here have a TJ Max of 100 degrees, which roughly equates to a maximum "CPU temperature" of around 90 degrees, as long as the CPU is being cooled fairly efficiently (ie core temperatures don't exceed the "CPU temperature" by more than around 10 degrees). 90 degrees also happens to be the maximum temperature (from memory) I can set in the BIOS before the CPU temperature alarm goes off.

[Nelson37]

CPU temp protections will shut it down before it catches on fire, but not before it has exceeded recommended operating temperature and shortened it's lifespan. There's immediate damage and long-term damage.

[hello_hello]

CPU temp protections will shut it down before it catches on fire, but not before it has exceeded recommended operating temperature and shortened it's lifespan. There's immediate damage and long-term damage.

I'd like to see some actual evidence to confirm that. Why wouldn't CPUs shut down to prevent long term damage in preference to immediate damage? Or in the case of Intel and TJmax, why wouldn't the TJmax temperature, where the CPU begins to throttle it's speed, be the temperature where throttling begins to prevent long term damage, rather than the point at which it's throttled to stop it going up in flames? The former would make more sense to me.
I'm not sure I've ever heard of an example where someone's been consistently running a CPU at temperature "x", only to have it stop working two years later because of accumulating long term damage. Are you aware of any examples where this has actually happened?

Anyway..... when it comes to Intel the way I understand it there's no real recommended operating temperature as such. They provide a Tcase value (60 degrees for the CPU in this PC) which I understand to be the maximum expected CPU temperature under "normal" operating conditions, ie checking email, watching a movie etc. Not "full load" conditions. This seems to be confirmed in a post from an Intel employee here, and in the case of the CPU in this PC.... I don't use it for much heavy lifting these days but for years it ran for hours on end at 10 to 15 degrees higher than it's rated Tcase temperature (still at least 20 degrees below TJmax) and five years later it's still going strong.

I do wonder if there's really a "safe" operating temperature which is below "maximum" operating temperature, why the CPU manufacturers seem to be keeping it secret.