Tom's Guide | Tom's Hardware | Tom's Games
![]() |
![]() |
![]() |
I wish I was getting the temperatures that you guys are getting. I have a AMD Athlon XP 2200+ (The TBRED) and its idle temperature is 65 and in games, it gets to 73. I have spent a fortune on new cases and Cooler fans but with no difference. The guy I bought the computer from tells me this is a normal temperature, but I think he is talking Bulls**t.
Has anybody got any advice as to what I can do to bring my CPU temperature down. Oh and by the way, this is the temp in Winter! Does anybody know if the new Athlon XP 2400+ runs cooler than the 2200+?

DANG!!! 65C idle?! You're on shaky ground there Rawdon. True, some (though few) would argue that these are acceptable temps, but having temps in the 70s is a threat to the life and longevity of your Athlon. A few questions:
1) What are you using to get your temperature readings?
2) How many case fans do you have, where are they placed, and are they blowing air out or sucking air in?
3) What kind of case do you have?
4) Is your CPUs Heatsink/Fan seated and installed correctly?
5) Is your Heatsink/Fan AMD stock or another manufacturer?My initial reaction is that your Heatsink/Fan probably isn't mounted correctly.

I just worked on a XP 2200+ or was it a 2400+ I can't remember. Upon startup its around 45C and at full load its 55C.
There are 2 mini exhuast fans.
What kind of thermal compound are you using? Silver is a good conductor and using copper heatsink would dispate heat better.
Very important that the heatsink touches the CPU tight and evenly. Some people hit it at an angle and then sometimes it catches the white plastic that says: SOCKET A

To first reply to KyleD, Iam using the motherboard temperature reading in the BIOS, I have 2 case fans (One at front sucking the air and one at the back exhausting the air), I have a normal Goldwin PC Case, I have checked the heatsink twice and it is installed correctly and Iam using a Evercool Copper Core Heatsink (CUD-725) which is rated to a Athlon XP 3000+.
To reply to RanmaP, Iam using standard silver paste with a Copper Core heatsink. I have also checked the seating and it is very tight. A strange thing is that eventhough the CPU is running so hot, the CPU Fan is only running at 2600RPM when its full RPM is 4200. Could this mean the motherboard is sending the wrong signals to the Fan?

A fan running at half-speed isn't going to cool as well as it does at full speed, but that still wouldn't explain 60+C CPU temperatures...I'm a little puzzled.
It's possible that your motherboard isn't giving your CPU fan all the juice it wants, which would result in the lower RPMs, and the simple solution for that would be to connect the fan directly to the power supply instead of through the motherboard (You may need an adapter) I've heard those Evercool's can get real loud at full speed though...
The other thing that comes to mind is that some motherboards have adjustible fan speeds for fans plugged directly into the motherboard built in to the BIOS...so there may or may not be a setting to adjust that.
Other then that, you seem to know your way around your HSF, so my last suggestion is that you ensure the CPU core and Heatsink bottom are clean and particle free before you add a new layer of thermal interface compound.

As I said in my initial message, I have spent a large amount of money on heatsinks, the copper one Iam using now is my forth. The first one I had ran at 6200RPM with a slight difference of about 3 degrees from the current copper one (But hell, was this thing loud. A 747 Jumbo Jet could have been outside and I would not have known!)I did try to connect through the PC Powersupply, but it ran even slower without the sensor telling it to run faster! I have also checked the BIOS, but there is nothing to adjust the Fan speed. As far as I know, Gigabyte does not put this feature in their BIOS anymore.
Iam about to buy my third case, which I have been told is designed for intensive gaming which has been designed with 4 Fans and maximum heat dissapation (a huge $60 worth of aluminium), but I have my doubts. Some other friends have told me aswell that if the VCORE is stable at 1.660 Volts, the temperature does not matter. Is this true?

Let your fingers tell you whether your heatsink is really 65 degrees C which is hot (100 C boils water).
MB temp reporting is full of sh_t at times. That's why some proud owners claim below room temp CPU reading. If your system runs stable at these temp your MB sensor could be at fault.

*LOL* I know what Real Cool means... I do that alot too. Even if I know its hot, I'll touch still it.
Usually it's not hot enough to burn you.
Uncomfortable? Yes... but not blistering.

I have tried what was suggested and the top of the chip does seem to be very hot (Not blistering but it is painful to hold my finger there for more than a few seconds). My initial thought of the MB sensor being faulty has been thrown out the window, as I tried the processor in an old FSB 133 MB(also Gigabyte) and the temp at idle was 64 degrees and about 72 degress at Windows. I am afraid to test this thing at full load incase I burn out the CPU and void my warranty.
Another very intresting thing occured when I put a older Duron 1200 in my new MB. The CPU ran at 27 degrees and yet on the older MB with the same heatsink, it ran at 59 degrees. Can anybody explain this? I really get the feeling that Iam missing something really obvous here! Because surely what the temp is in one machine should be the same in the other.

go to google and type in tak. they make processor heatsink/fan, they have several models. i have seen the double up and running on a athlon a friend of mine has. it dropped his temp 10 degrees!!

Question
"Another very intresting thing occured when I put a older Duron 1200 in my new MB. The CPU ran at 27 degrees and yet on the older MB with the same heatsink, it ran at 59 degrees. Can anybody explain this? "Answer
"MB temp reporting is full of sh_t at times."

May be this would help.
I have lapped enough heatsinks in the last three years (all due to the damn heat CPU are generating recently). Thermaltake, Thermalright, Stock, AVC, Coolmaster, none has a true flat base. They are either concave or convex. When use with P4 having a much larger heat spreader the problem "may" not be a profound, but with AMD Processor this could be a much bigger issue.
It does not take much to find out if your heatsink base is flat or not. This is another reason stock HSF's use thermal "pad" to make up the difference.

If a Duron 1200 will actually run at 27°C then it will have the processing power of an AMD486DX100.
The temps have to be incorrect.
I hate to say this, but, I believe the interface between the cpu and heatsink is suffering. You should not be running this hot.
Something is dirty, poor fit, or improperly installed.
Suggest you recheck everything. I've had huge temp swings on my system between pads, paste and improper installs.
Good luck,
Skip

To ask Real_Cool a question, you are saying that thermal paste should be avoided and I should get a heatsink that has a thermal pad fitted on it? And as far as I know, every motherboard, even branded are full of S**T.
And to answer SkipCox, I had the same thought when I switched on the systems and began monitoring the CPU Temp. I reset the machine and looked in the CMOS, which stated that it was running at 1200Mhz, and Easytune 4 in Windows confirms this.
I have checked the CPU Cooler for dirt and washed it with alcohol, I have reinstalled it 4 times, but Iam interested to know what you mean by 'poor fit'. Do you mean that the Cooler bracket does not squeeze the cooler against the CPU?

Good Thermal paste is preferred over thermal pad provided the sink is really flat. Thermal paste does not and will not transfer heat as good as solid copper or aluminum.
Thermal pad is like doubling the thickness of kitchen tin foil with thin layer of thermal transfer compound (dried) coating on both sides. Because of it's thickness, via compression, it compensates some of the imperfection of the sink base surface.Yes, Asus, Abit, MSI, Gigabyte, Albatron, etc are name brand, and their temp reportings are all full of it. How do you think over $ billion industry is created in the last three years? Heatsink and Fan.
They are all from a place called Taiwan. You see, Asus.com.tw, msi.com.tw, abit.com.tw

hi i have a xp 1900+ on v266 jetway board with a geforce 4 64mg ddrram video card 2x cdroms 2x20gb hdds 2x case fans and a thermaltake k7 volcano 11+ xaser editon in a cheap crappy case and i only get to 37.5c.its all about were the air comes in and were it goes out. i have 1 case fan blowing in from the back straight accross the front of the proccessor fan and my other sucking out at the side panel. i had problems to start with (43c) but i just moved the case fans around until i got this result. so if i were you i would just try moving my case fans around as each case circulates the air different.I also noticed that my temp went down more when i put my case on my desk than in the desk.

One of my P4's clocked from 2 to 2.92Ghz, Sandra reports 91 Watts, runs Prime95 for hours without crash and MBM reports 43C with stock HSF at 26C room temp. I am not jumping up and down because I know the senor is not reporting the correct temp. Based on some math i.e. Watts x C/W (heatsink coe) + room temp = CaseT, I say it should have been around 58~60C rather than 43C and I don't care how many fans blow onto it.
Hell, it is running 24/7 for the last 8 months. Motherboard temp reports have fooled a lot of people. If the system runs smoothly, sit back and enjoy.

"One of my P4's clocked from 2 to 2.92Ghz, Sandra reports 91 Watts"
I did not know Sandra reported wattage. Which module is it?

Sandra 2003 beginning November 2002. I am using version 2003.7.9.73. It did not have this feature prior. I found this to be really nice. Simple math based on average wattage from Intel and AMD data with consideration of Vcore and processor speed.
I used to calculate it myself prior to the functions added.
In addition to Wattage, it reports also C/W which can be inaccurate when using air duct drawing cool air from outside the case direct.

Run "Burn-in Wizard", the report/result will show:
Min, Max, average CPU temp, system temp.
Min, Max, average fan speeds.
Min Max, average Vcore, wattage, C/W
Begin/end time stamp.
I normally run 25 to 30 continous runs. But, it may pass the test and fail in Prime95 which I use for stability. Sandra is nice for better temp monitoring with the par chart over time/run.
You know what I mean Johnoh.

![]() |
![]() |
![]() |

This post is quite old and has been locked from receiving new replies. Please create a new posting instead.
| Ads by Google |