[XJDHDR]

I see! As I said above, if the standard thermal grease I have ordered doesn't do the job, I will consider using the grease you suggested.

[avarice]

hey hold on before you order any more costly grease.
i ran into a guy who had issues very similar to yours. he had a sink that was not properly seated on he cpu. the clips on the side of the sink that were supposed to secure the sink tightly against the cpu's face were jagged. when he removed em and sorta placed the sink witout any clips over the face of the cpu, he saw a 10-15 degree fall in his temps.
that explains why the aluminum foil could reduce the temp when u used them. the only thing it did was close the gap between the cpu's face and the sink's bottom(i ve assumed that there is some).
i am afraid a grease wont help you solve the issue. its time you change the sink... the whole setup.
there is an easy way how to check if the sink is faulty.
take a thread.. a very fine one.. place it on the face of the processor and then secure the sink over it.
after you've locked all the clips in the sink, if you are able to pull off the string(fine thread) from between the cpu and sink, your sink is faulty..
if the string offers some resistance while coming off the face of the cpu, your sink is alright...
hope this helps.

[crjdriver]

Since there are different ways to install the thermal compound AND the hsf depending on what type of cpu, it would really really help if you posted what you are working on.

Just think about it for a minute or two. Do you think the procedure for installing a water pump is the same for a 1955 Chevy as it is for a 2009 Cadillac??? I do not think so. You are going to have to ge through a LOT more steps with the modern car.

[XJDHDR]

@Avarice
I tried the test you suggested with a strip of Aluminium foil. I sandwiched it between the CPU and heatsink and then tired to pull it out. The foil tore itself and left a piece between the heatsink and CPU. I then tried this test again using a strip of paper instead. The paper didn't tear but there was definitely some resistance as I pulled it out.

@Crjdriver
That is correct. I can do better than explain what I have though. I have attached some pictures of what I am working on to this post. They are low resolution though since I had to take the photos with my phone's camera. In each picture, the RAM chips are there to use as a size comparison. I will also give a small explanation of each picture:
(cpu-overheating-no-apparant-reason-)cpu - It's a picture of my CPU. It's a Pentium 4 3.0GHz HT. It also uses the newer pinless design, as opposed to the older CPU which used pins.
heatsink-underside - The underside of my heatsink. It is perfectly symmetrical so it doesn't matter which way round you attach it to the motherboard.
heatsink - A top view of my heatsink.
motherboard - My motherboard. Illustrates the pinless design used on the CPU.

[avarice]

i know its a bit late to ask this question but still..
you sure you secured the locks after you placed the heatsink? i guess you have to press it to secure the locks and my friend says.. it will give a "click" sound when u do that.

[XJDHDR]

Yes, I'm very much sure that the clips are secured. Am I correct in thinking that your suggested test would also show whether the clips are secure or not?

[avarice]

hey did u say that you could remove off the paper?
i tried it with mine and i pretty much could not. the paper tore off.
i guess the sink is not pressed against the cpu tight enough. by the looks of ur cpu, it seems like an lga775 socket. its pretty common these days. why dont you try plugging in someone else's heatsink onto your motherboard? that way you can zero in on the issue.

[avarice]

Originally Posted by XJDHDR

Am I correct in thinking that your suggested test would also show whether the clips are secure or not?

it is to test whether the sink is pressed against the cpu tight enough so that heat dissipation actually occurs.
sometimes, even after securing the clips, the sink might be loose on the surface of the cpu.

[win2kpro]

Originally Posted by avarice

even after securing the clips, the sink might be loose on the surface of the cpu.

If that were the case, the machine would fail to boot.

[avarice]

Originally Posted by win2kpro

If that were the case, the machine would fail to boot.

not always. if the chassis intrusion was disabled, bios can load and machine would boot normally.
heck i myself have clocked a processor to 4.8g, used a homemade tube cooler, poured liquid nitrogen into it and it worked.. did not even hook up the sink.
chassis intrusion is not always a problem.. atleast not with modern boards.

[crjdriver]

OK, you have a socket 775 type cpu. After you clean ALL parts with 99% or 91% alcohol, apply a small amount of thermal compound to the center of the cpu heat spreader. Now take the hsf and place it on the cpu however do not lock it down yet. Give it a twist or two back and forth.

Now engage the four push pins. Do this by pushing on diagonally opposed pins until you hear them click; repeat for the other two. Now turn the mb over and make sure ALL four pins are fully engaged. Note I always recommend pulling the mb when installing a hsf on a socket 775 type.

One other item. The hsf must contact the cpu heat spreader with a min of 15lbs of clamping force. Failure to achieve this force will result in a shutdown after a few seconds.

[avarice]

Originally Posted by crjdriver

One other item. The hsf must contact the cpu heat spreader with a min of 15lbs of clamping force. Failure to achieve this force will result in a shutdown after a few seconds.

that was exactly what i was trying to bring out. the clamping force....
guess i ran out o words..

[win2kpro]

Originally Posted by avarice

not always. if the chassis intrusion was disabled, bios can load and machine would boot normally.
heck i myself have clocked a processor to 4.8g, used a homemade tube cooler, poured liquid nitrogen into it and it worked.. did not even hook up the sink.
chassis intrusion is not always a problem.. atleast not with modern boards.

You don't understand the properties of a socket 775 LGA processor. I can guarantee you that if the cooler is not locked down where the pads make contact with the socket contacts with at least 17 (f/lb) of pressure the machine will not POST or boot.

If you check the archives here at TSG I can't tell you how many times I have had to post instructions on how to reinstall a socket 775 processor because the cooler was not installed correctly, and the machine refused to POST or boot.

[XJDHDR]

@CrjDriver
Thank you for that information. My first thought would be Methalated spirits but I need to confirm it's alcohol content first.

@Win2kPro & Avarice
That is some interesting information. I'm not sure what is meant by "chassis intrusion" though. How would the CPU/ motherboard know there is too little pressure though?

P.S. My thermal grease order still hasn't arrived. I got an email from Sybaritic (the company I ordered from) this morning saying that they have given the grease to their courier and it should be here in the next 2 - 5 days. They also gave me a tracking number. As of 10pm yesterday, it was in Johannesburg.

[win2kpro]

Originally Posted by XJDHDR

How would the CPU/ motherboard know there is too little pressure though?

Your problem is not pressure related. If that were the problem, you would not be able to access BIOS. The pressure problem occurs when the pads on the processor do not contact the the contacts in the motherboard socket with a minimum of 17 (f/lb) of pressure.

When heat is not being transferred from the processor heatspreader to the cooler heatsink properly so it can be dissipated by the fan on the heatsink the temperature rises rapidly. Processors have a factor known as TDP (Thermal Design Power). This is the amount of heat that must be dissipated by the cooler for the processor to operate properly.

When the processor heat reaches a point above TDP the 1st thing that happens is the processor "throttles back" so that it is not operating at the specified specification speed. If the temperature continues to rise, it reaches a point called thermal trip which is approximately 15-20°C above TDP based on the processor being used. At that point the processor shuts down completely to protect itself from self destruction.

I would have to have the processor sSpec number to tell you exactly what the specified TDP temperature is for your particular processor. The TDP varies from processor to processor based on the core used in that particular processor and the stepping of the processor.