[Tyger]

I would just like to ask for some help as well, luckily I have not burned out any chips or friend anything for the past 4 months that I have owned this computer.

My questions is, I have 5 fans, one intake in front, one more intake on side panel blowing on top of CPU, stock AMD hsf over AMD Athlong XP 3000+, 2 exhaust fans out back. Even with this setup, the temp program that came with my motherboard states that my idle temp ranges from 48-50c, and usually during gaming it gets as high as 57, on occassion I have seen it at 62 but that was once. I do not like these temps and I have artic silver compound as per instructions between the heat sink and processor. Is there anything else I can do to bring my temps down?

Make sure all those fans aren't working against each other. The air should move from front to back or vice versa.

[Advertisements]

how do i do that?
I mean, i tried unplugging all the fans except the hsf and one exhaust fan out back and the temps idled at 57, now i have 2 exhausting out back and its idling at 53. I have tried switching what the fans are doing and placing them in different areas but seems like idle of 50 is the best i can do..what else should I try, I have already tried to remove..clean up thoroughly and reapply compound and it dosnt change.
Thanks

[csu266]

how do i do that?
I mean, i tried unplugging all the fans except the hsf and one exhaust fan out back and the temps idled at 57, now i have 2 exhausting out back and its idling at 53. I have tried switching what the fans are doing and placing them in different areas but seems like idle of 50 is the best i can do..what else should I try, I have already tried to remove..clean up thoroughly and reapply compound and it dosnt change.
Thanks

[Doby]

Your Everast report is pretty normal, as for the rise in temp to 54c thats not normal but I don't think it has to do with a problem with your system but a bad sensor or software read. normally you see a rise of 5c to 7c when going from idle to full load.

Download install and run prime95's torture test, it tests the cpu,mobo,psu,ram and heat related problems.

Prime will run your cpu at 100% this will heat up the cpu but if a overheat exsits prime will fail and stop running. If you can run prime for 30 minutes with out fail then you have no overheat issues.

If you can run prime for 6 to 8 hours then all of the hardware I mentioned above has no problems and is totaly stable.

If prime fails though its hard to tell what peice of hardware is bad but at least you know something is

Another thing to consider is using a digital thermometer and place its sensor at the base of the heatsink to get a true reading of your cpu temps. I do this with any new build just to get a idea on how off the sensor/software temp reads are and to be honest there ain't many that are correct, some are off by as much as 15c and they usually read high.

Another thing I have done and checked for accuacy with a digital themometer is the finger test,

Place your finger at the bottom of the heatsink when running the cpu at full load and use this scale

35c to 45c = comfortable warm
45c to 55c= need to pull your finger away after 3 to 5 seconds
55c to 65c = need to pull away after 1 second
70c= burnt finger

I hope some of this helps you but I don't think you have a heat problem because of the fans, hsf and thermal paste you are using.

Rick

[SomeGuy05]

I've had the Prime95 test running for about an hour now, Everest reports the temp steady at 50c-52c, however the CPU usage is steady between 50%-54%. You mentioned it should run it at 100%. Is there any way to boost that?

One of the main uses on this PC is an online game. Both times the CPU has gone out, that is what was running. But like I mentioned, It worked fine for well over a year before the first chip went bad, and ran fine for another week on the second chip. I don't play 8 hours straight like some people. I think I had been playing for at most one hour last time the chip went out.

The other issue that concerns me is the different readings I get using different programs. While idle, these were my results:

Everest: 41c-42c
SpeedFan: 46c-47c
BIOS System Monitor: 49c-50c

Edited by SomeGuy05, 26 August 2005 - 12:02 AM.

[The Skeptic]

Hi SopmeGuy05.

When you put heavy load on your cpu the temp rises very quickly. It also cools down very quickly as soon as you take off the load. This is normal and you shouldn't worry about that. The important question is at what temp the system stablizes. in post no. 13 I quoted from SpeedFan how to identify your cpu temp. I suggest that first of all you follow the instructions and find out which temp represents the cpu. next, start SpeedFan and leave it open. After that start running a heavy program, like an antivirus scan and watch the temp on SpeedFan. If the temp keeps climbing above 60-65c then stop the scan and let the cpu cool of. Let us know what happens. If it keeps heating then I would start suspecting that your problem lies somewhere else, probably an unstable power supply.

[warriorscot]

csu266
A different HSF, you should use the pad on the stock hsf for some reason when you change it the hsf doesnt seem to works as well in some cases. I would just get a better hsf they arent expensive.
But dont post inside someone elses thread its a bit rude, you should make your own thread.
Someguy, it looks like the temp. sensors on the board are not calibrated properly making them read high in bios. I would use a digital thermometer and you can read the real temps when it crahses.

Edited by warriorscot, 26 August 2005 - 05:51 AM.

[csu266]

Next time I will start my own topic, this topic was talking about temps as well so i thought I would also get some advice. But thanks, I will try a new HSF.

[Doby]

I am glad you ram prime and mentioned the cpu usage at 50% this clued me into something, prime95 is a program for stress testing and will run your cpu at 100%, your is running at 50%, there is only one condition that will cause this.

Intel has a feature that throttels back the cpu when it is in danger of burning up, at first I just thought maybe you had some bad luck with chips but I am sure there is a problem because the cpu is throtteling back.

The number one cause of this is heat but the experiance you now have with mounting cpu's, hsf's, thermel paste, the advice you have been given with these components and the quality of the themal solutions you are using suggest to me that heat is not your problem.

2nd cause could be accidental overclock, check this out I can't help much I ain't that familar with intel setups.

3rd, high vcore setting, vcore is the voltage supplied to the cpu if it is set to high this will increase temps, check the intel website for the spec of your cpu then go into bios and check that the vcore is set according to their spec.

4th, flaky psu so go into bios and check your voltages, 3.3v, 5v and 12v each of these can be plus or minus 5% and still be within spec but this is not absolute with a flaky psu because it could be overvolting when you ain't watching the reading, if you are overvolting any of these I would replace the psu with at least a 450w name brand

5th, bad mobo, it could have some kind of short or something that would requirer replacement.

I will be away for the weekend and won't be able to respond to Sunday evening but hopefully someone more familar with intel setups will respond.

Other than checking settings and voltages in bios I would not run the machine.

Rick

Edited by Doby, 26 August 2005 - 01:40 PM.

[SomeGuy05]

Out of your list, the only thing I see that could be possible is the motherboard...

1. (Heat) should not be a problem with the cooling I have installed, and the temps on all other components are very low.

2. (Accidental Overclock) I never overclock. I checked that just to make sure the last time I was in BIOS, OC is off.

3. (Vcore Voltage) The Vcore voltage requirement for a P4 3.0Ghz-3.6Ghz (mine's 3.0) is 1.40v. According to the BIOS meter, it's supplying ~1.38v.

4. (Other Voltages) All appear normal, but like you said I'm not monitoring fluctuations all the time. The PSU is brand new, I got it after the first chip went bad because the old PSU was faulty. The new one is a 450w Cooler Master.

5. (Motherboard) If that's the problem, I guess there's no way to know for sure.... Everything appears normal.

I ran Prime95 for 18 hours. CPU temp stayed steady at 50c-52c the entire time according to Everest, and the prog found no errors and had no stoppages.

I'm not sure there's anything else I can do to find the problem. The second chip ran fine for a week under normal usage (gaming, browsing, etc), so it's basically impossible for me to tell if something is steadily causing a problem over time. I wish it was something instantly noticable so I could immediately tell if a part replacement had any effect...

[Doby]

I duno, you certainly have tried everything recommended and followed through on all the bios settings and voltages and everthing appears normal. Running prime for 18 hours all most always means your hardware is good if not it would error and the program will quit running.

The only thing that concerns me is that the 50% cpu usage when running prime, I pretty sure this means the cpu is throtteling back, I have no other expination for it and if it is, then this, whatever the cause, is why the other two cpu's burnt out.

I am at a loss at this point but I will do a couple searches and get back to you if I find anything.

Maybe someone else has a idea

Rick

[Doby]

I think you should read this

There is also tons of info to sift thru if you google cpu throttle back.

Also if there is a setting in your bios or you are using software to controll the speed of your cpu fan try and disable it and see if the cpu usage when running prime will hold steady at 100%.

Rick

[SomeGuy05]

Thanks again for all the info guys (that was an interesting article in that link as well).

I've been running the rig for 8 days now since the last CPU was installed, gaming, browsing, etc. Everything I have been doing on the system for over a year. So far, no trouble. I've stress tested it with Prime95 (20+ hours, no stoppages or errors), 3DMark, and AquaMark3.

These are the temps I have recorded during almost constant monitoring with Everest:

-----IDLE-----

CPU: 41c-43c
GPU: 35c-36c
MB: 32c-33c
HDD: 30c

-----LOAD-----

CPU: 50c-52c
GPU: 40c-41c
Have not constantly monitored MB and HDD under load as they have not been an issue.

And the final specs on the rig after new CPU and cooling (Mid-tower case):

PSU: Cooler Master 450W w/120mm fan
MB: ASUS P4C800-E Deluxe (socket 478)
CPU: P4 3.0 (Not HT, no OC)
HSF: Zalman CNPS7000B-Cu (92mm fan, almost all copper, fan speed controlled by a direct-wire potentiometer at highest setting [2500RPM])
THERMAL COMPOUND: Arctic Silver 5 (applied exactly according to specifications on their website/manual)
GPU: RADEON 9600XT 128MB
GPU HSF: Arctic Cooling VGA Silencer Rev. 3
RAM: 1.5GB Patriot Cas-2 PC2700 DDR (3x 512MB)
HDD: Maxtor 120GB 7200RPM
CASE FANS: 5x 80mm Cooler Master (2x lower-front intake, 1x center side-panel intake, 2x high-rear exhaust), all unobstructed inside and out of case.
2x 5.25" CD drives
1x Floppy Drive
Sleeved IDE cables added to replace standard "ribbon" shape to help airflow.

Hopefully it holds up this time. If not, I'll probably scream like a banshee and put a bullet in it for causing me the stress and will return to the message boards barring a PC-problem induced suicide.

Thanks again for the helpfull links and info everyone

[Doby]

Your welcome,

Your temps are perfect for that cpu, good luck and I think everthing will be ok now