How to Stop Your Laptop Overheating

Testing the Temperature
Quite a few people have commented that I used way too much thermal grease (compound) and that the amount I used would cause overheating. I disagree with this view. So in an attempt to provide some kind of proof, one way or the other, I decided to monitor the temperature for a few weeks with SpeedFan v4.44. Then to take the laptop apart again, clean it up, using the suggested alcohol wipes method, and replace the thermal compound. This time I would use a thin amount. I would than assemble everything and then monitor the temperature again. I followed the same process as before with the slight suggested changes. This documents that process.

According to the CPU chip specifications for a LF80537 T5500, which is a Intel Core 2 Duo Mobile, the Minimum/Maximum operating temperature (°C) is 0 - 100°C. The maximum operating temperature of the CPU was 65°C before I started this process. After I completed the process the maximum operating temperature was 60°C. It seems I might have applied too much thermal compound but not way too much to cause overheating. Just too much as in wastage. It would seem as if excessive amounts of thermal compound could cause your system to heat up slightly more and if it is running close to the maximum allowed temperature it might cause overheating. It caused a 5°C increase in temperature in my case but this is not definite because it was not a controlled situation. The increase was not enough to cause overheating, but enough to deduce that excessive amounts of thermal paste could increase the temperature. The photos contain a step-by-step guide of what I did.

Notes: Wipes

I found the alcohol wipes worked terrible. It smeared the thermal grease all over and just created a bigger mess to clean. It could be because the previous thermal grease was al dried out and flaked off nicely. This thermal grease was wet and sticky. It smeared too easily. On the GPU there are very small resistors or components and the wipes smeared the thermal grease all over them and then kept getting caught in them.

If I were to do this instructable from the start over again, I would use less thermal grease from the start as it does make a difference no matter how slight, but I would not use the alcohol wipe.

Update 2013

I have been ripping Intel CPUs out of laptops and desktops and replacing the thermal compound (grease) for years now and have noticed that on the older Intel CPUs the thermal grease was used liberally. However, on the newer CPUs the compound is quite thin. I wonder if it really matters that much or is it cost saving to have a very small (tiny) amount of compound in between the CPU and the cooler (heat sink)?

Here is what Intel itself says about how to apply thermal grease to a CPU. What you will read will surprise you because there is no mention of layer thickness and no mention of spreading it around. To tell you the truth it states that you blob everything in the centre of the CPU and the pressure of the cooler (heat sink) on the CPU will determine the thickness, squeezing out all excess. I think I also mentioned something like that in one of my comments. This information comes straight from the CPU manufacturer, and they should know. Read for yourself at:
http://www.intel.com/support/processors/sb/CS-030329.htm (Click "How to Apply TIM" from the menu)

I have included some of the photos from the Intel site for reference.
I have also replaced the thermal grease (compound or interface material if you are from Intel) on many secondary chips like integrated GPUs on motherboards by releasing the cooler (heat sink), cleaning both surfaces, apply thermal grease and replace. Once you know how, you can replace the thermal compound on practically anything.

Update 2014

Many people are still suggesting I am using too much thermal grease. I want to bring the following to the attention of these people. I am using a non conductive thermal grease. It does not conduct heat as well as the conductive thermal grease but does not damage components when it comes in contact with them because it does not conduct electricity. Cooling has to do with getting rid of excess heat, so the thermal grease is only one part of the solution. The heat sink, fan and airflow are the other parts. You could have a CPU that does not have well applied thermal grease but has a fantastic heat sink and airflow across it and the system will never overheat. Here are some links to other resources which I found very helpful in understanding overheating, thermal grease, fans and airflow.

http://www.tomshardware.com/reviews/cooling-air-pressure-heatsink,3058-8.html
http://www.tomshardware.com/reviews/thermal-paste-heat-sink-heat-spreader,3600.html
http://www.tomshardware.com/reviews/thermal-paste-performance-benchmark,3616.html
http://www.bit-tech.net/hardware/2009/02/16/all-about-tim/1
http://www.intel.com/support/processors/sb/CS-028869.htm
http://www.intel.com/support/processors/sb/CS-030329.htm

I hope this helps everyone understand more about what causes overheating and how to solve the problem.

During the course of writing and expanding this article I have constantly been told that I use too much thermal compound or grease. Most of the people that make these statements have no proof either way of how much paste needs to be used. They also do not read the complete article with its comments. I have been doing this since the early 1980's and have never found the amount of TIM (thermal interface material/compound/grease/paste) I use to cause any problems. I normally provide a link to CPU manufacturer, Intel's, website where they explain how to apply TIM. Intel just state that you apply a blob in the middle and let the pressure of the heat sink spread it. I also refer people to Tom's Hardware were there is a much more in depth analysis on the application of TIM.

Although Tom's Hardware is quite thorough regarding the issue you will see that it is not totally impartial and personal the preferences of the tester do come to the fore. Like the statement that "smearing paste on the whole CPU is quite pointless and a thing of the past" because all CPUs have hot spots that generally lie like a sausage in the middle of the cpu from left to right. However, Tom's admits itself that it is not that simple in their statement "The two images below illustrate what a hot spot is, albeit in an over-simplified way. Reality is not as simple; CPU cores may be loaded differently, and there is also the issue of on-die graphics, which may be more or less active than the processing cores." This means we do not really know exactly where the hot spot really is at any given moment. Tom's even admits that the application method is a "philosophical debate". Although Tom's stated previously that spreading paste on the whole core "is quite pointless and a thing of the past" they then discuss Revoltec Thermal Grease Nano which is applied with a brush. A picture is shown where the whole CPU is "painted" with the paste even though it is "pointless and a thing of the past". Tom's states that "these pastes typically fall to the bottom of our performance charts. When you try to apply semi-liquid pastes by brush, typically you wind up with too much, and that isn’t optimal." Then Tom's states that spreading is tedious and runs the risk of applying too much.

I am a firm believer of spreading a liberal amount over the entire surface that comes in contact with the heat sink. The reason for this is that I believe in the the law that states the maximum heat transfer happens when the biggest area is touching the conductive material.

NOTE: NOT ALL TIMs CAN BE SPREAD BY HAND!

In an effort to provide a more believable answer than just my experience and opinion, I have gathered some pictorial "evidence" of how the manufacturers of the chips themselves apply TIM. At www.hardwaresecrets.com they tested the outcome of the amount and method of applying TIM by applying a tiny dot in middle, a small dot in middle, transversal line, parallel lines and spread across the CPU methods. Although the tests are not 100% lab proven the do give and idea of what to expect.

The results in core temperature in Celsius were as follows:

a tiny dot in middle - 55

a small dot in middle - 50

transversal line - 51

parallel lines - 51

spread across - 51

The difference in applying too much paste to just the right amount is 1%. The difference between applying a dot in the middle and spreading it is 0.2% and can be ignored. Read the article as the "correct" dot amount for them is also a bit of trial and error. I will then rather stick to my spreading method.

Make up your own mind based on the this information, the kind of paste you have, the kind of CPU and the heat sink that is to be used.