i7 with aftermarket heatsink still hot

Question

I have an i7 2600 processor running with an after market coolermaster heatsink. I've tried two different aftermarket heatsinks and the stock. The cpu runs at idle around 40-45c in my room that is about 65-70f or 20-21c. The differences between the after market heatsinks and the stock are almost negligible, about 2-3 degrees between them. Referencing this post this temperature is normal. It still seems very hot to me though especially under load it can get as high at 65-70c getting near the max temp of around 72c I believe. I see the 40-45c range when using my computer to do basic things such as surf, use eclipse and visual studio. The cpu runs hot when I play a game such as Starcraft II or Portal II.

My question is: Shouldn't I be getting lower temperatures with an aftermarket heatsink? I was hoping that they would drop my core temp to the 30c range while idle but there is little difference between the stock and aftermarket coolers. I put a new layer of thermal paste on each time I apply a new heat sink.

Answer 1

This maybe a stupid answer but you are cleaning off the existing thermal paste, right? I've had the same behavior on systems where the paste is 1) too much, 2) not evenly distributed, 3) not enough.

Answer 2

The best heat sink in the world cannot make up for poor case-flow.

Have you properly implemented cable management and aligned case fans to promote equal air current throughout the components?

Another factor to consider: Ambient temperature. If your case is next to a furnace, or close to a wall, they can act as radiators/insulators.

Answer 3

The 212 is a good cooler for its price. But on the ones I installed, the base on them looked like they were finished with a beltsander. Being overly analitical, I make assumptions on how much metal to metal contact I am going to get. And in my opinion it was not going to work like I wanted.

So without wasting to much time or effort (20minutes) to perfect it, I flattened out the high ridges on it. A quickee lapping job

I put 600grit high quality metal oxide sandpaper down on a perfectally flat surface, and moved the base of the sink in the opposite direction of the ridges. Knocked down 1/2 of the ridges height. Then replaced the sandpaper with a microfiber rag, and a normal (slightly abrasive) car wax. Keeping the base perfectally flat, not taking off any more than nessisary to just knock down the high machining ridges. Once Finished it was not mirror or perfect, but it would get a lot better metal to metal contact going.

The I cleaned off all the wax and junk, (dont want any grits in there :-) cleaned it with alcohol and a tech spray. Then fully dried it , which includes the micro porosity of the metal.

Before applying thermal paste, I like to "Permiate the porosity" of the 2 metal surfaces. So I use a lint free microfiber rag, and rub the thermal paste into both surfaces, then remove most of that.

Applied AS5 thermal paste , just because that is my today favorite still and I had some. And I used the very thin spread method, because that is my choice of paste spread methods. (especially when the first seating might not be straight down) The 212 can have tracts down it because of the "touch" base , so using the Squished pea or line method just was not going to work.

Putting it on One of the computers cpus, replacing the stock intel, and its mostly dried up junk paste under it, went way over my expectations. the thermally controlled fans didn't even go past thier Idle speeds until I ran some hard stress testing, and even then it was great.

You did not say what Overclock your running on your chip, or if your at stock, and you did not indicate what Voltage the CPU is at?

When it comes to IDLE temperatures, you did not indicate how you have the bios set to allow things Like CE states , and all the other cooling and power saving features. Some features of the bios can put the chip into a very low power state. Even the OS your running could be "waking up" the cpu more or less, or how you have the OS set.

The way you have all of those things set, can make very big differances, when trying to compare your results to others.

Also I agree a lot with Waterwizards, pointing out that you cant get room temp CPU readings, without good case cooling and logical Flow of the air patterns via the fans and all.