Lapping your heatsink for better cooling performance.

[Shock]

So I just recently finished lapping the heatsink on my Arctic Cooling Freezer pro, and realized I should share how to properly lap a heatsink for all of you Audioholics out there. I didn't have a camera to take pictures of the heatsink before and after but the difference is pretty remarkable.

So what you're going to need is some wet/dry sandpaper. I just went to the local hardware store and picked up some 400-600-1500-2000 grit waterproof sandpaper for about a buck a sheet. You're also going to need an extremely level surface, some dish soap, and some water. I like to just fill a container with water and take it to my computer desk or an old coffee table since the whole process does take a long time.

So you're going to soak the sandpaper in water for a few seconds. I use some butter knives to hold the sandpaper open as it wants to fold over on itself when it gets wet. Next you're going to put a VERY small drop of dish soap on the bottom base of the heatsink and spread it evenly with your finger. After this wash your finger in the tub of water and start spinning!

Some people like to move the heatsink back and forth on the paper, but I find spinning allows the weight of the heatsink to do the work. It also is faster than the straight back and forth technique. After a few minutes you'll need to wash the sandpaper again and the bottom of the heatsink and apply new soap. You might not think that this matters but believe me, if you don't want to be there all day you will keep a relatively clean piece of sandpaper.

I like to spend 15-20 minutes with each grit of paper, so the entire process should take 60-80 minutes depending on how shiny you want it. At the end you should be able to hold the heatsink on an angle and use it as a mirror to read your monitor. It should look like a mirror, but only from an angle. If you look at the heatsink straight on you'll notice swirl marks that can only be gotten out with a nice buffing of the metal. This is NOT necessary and is pretty much a waste of time.

After lapping, my temperatures went down roughly 3 degrees Celsius, and my fan speed dropped 100 RPM's under load. It's not a massive decrease in temperature but it's pretty good for something you only have to do once. My AMD Phenom x4 810 idles at 25 degrees Celsius, and under a full prime 95 torture test of all 4 cores running at 100% it hits just under 40 degrees Celsius. Idle fan speed is only 550 RPM, and full load is 1153 RPM. In CPU terms, mine is extremely cold.

I love tinkering with my computer, so if you're like me you can give this a try.

 

[itschris]

Cool stuff.

How are you spinning the heatsink? When you say "spinning" I assume you mean you're just physically turning it in circle over the paper correct? Or are you laying the heatsink on the paper and literally spinning it like a top?

 

[jinjuku]

LOL, enthusiasts have been doing this since the days of the OC'd 486 DX4's. I just would rather spend $50-60 on a high end heat sink and be done with it.

Sorry, I always got a laugh out of that.

 

[BoredSysAdmin]

Hmm, I don't know why, but then you mentioned lapping I thought about this:

Seriously thou OCing Celeron 300A to 450 was the Bomb !!

 

[jinjuku]

Hmm, I don't know why, but then you mentioned lapping I thought about this:


Seriously thou OCing Celeron 300A to 450 was the Bomb !!

I remember the dual Celeron 300A at 450Mhz with the Abit BX6 motherboard. That thing was stable as the day was long with a good HSF.

Ran photoshop like a bat out of hell.

 

[j_garcia]

Not sure why this would even make a difference at all. Yes you get slightly more contact between the surfaces, but it does not make the heat sink itself any more efficient.

 

[Alex2507]

Not sure why this would even make a difference at all. Yes you get slightly more contact between the surfaces, but it does not make the heat sink itself any more efficient.

It's Canada. There's not much else to do.

 

[bandphan]

It's Canada. There's not much else to do.

well there is..

 

[Nemo128]

Not sure why this would even make a difference at all. Yes you get slightly more contact between the surfaces, but it does not make the heat sink itself any more efficient.

The tiny pits and grooves that happen on some HSFs during manufacturing are taken out of the equation by doing this. The problem with these pits and grooves is that they can retain very slight amounts of air in them when the HSF is pressed against the core or heat spreader. The thermal paste is not thin enough to enter the smallest of these pits and grooves. The air becomes a heat pocket, and air to metal thermal conduction is a very slow and very inefficient means of thermal dissipation.

So yes, it actually does increase efficiency by increasing the surface area coming into contact with the dissipation source. Is the benefit worth the amount of time you can spend doing this? I've done it and it was only worth it when I had time to spare. These days I'd be lucky to have the time required to get the sandpaper.

 

[majorloser]

But what about the top of the CPU? It's not exactly a perfectly smooth surface. Not that I'm recommending putting it on some sandpaper.

I can see where this would help to remove anodizied coatings on the bottom of the heatsink or microscopic amounts of alluminum oxides.

 

[BoredSysAdmin]

I bet dollar for dollar, then both surfaces would mirror-like polished - you'd get better performance 100% of the time

 

[Shock]

Cool stuff.

How are you spinning the heatsink? When you say "spinning" I assume you mean you're just physically turning it in circle over the paper correct? Or are you laying the heatsink on the paper and literally spinning it like a top?

I literally spin it like a top, the friction is very very minimal with the dish soap. A good spin can get you a good number of rotations.

 

[AcuDefTechGuy]

Yeah, isn't it easier and more efficient to just get a CPU cooler with a 120mm fan?

 

[Shock]

Yeah, isn't it easier and more efficient to just get a CPU cooler with a 120mm fan?

Anyone who is going to go to the length of actually doing this is going to already have a massive amount of airflow in their case as it is. I can tell you that lapping is easier than removing a backplate and installing a cooler, especially if you have the system fully installed as is.

 

[Nemo128]

But what about the top of the CPU? It's not exactly a perfectly smooth surface. Not that I'm recommending putting it on some sandpaper.

CPUs have had copper heat spreaders for a number of years now, and you can indeed lap the heat spreader as well. The combined effect of both is pretty surprising. Sometimes you have to increase the tension on your HSF mounting hardware though, as the decreased thickness of the two elements leads to less contact by pressure.

I bet dollar for dollar, then both surfaces would mirror-like polished - you'd get better performance 100% of the time

Absolutely, the thing is...

Yeah, isn't it easier and more efficient to just get a CPU cooler with a 120mm fan?

most people that do this as Shock mentioned are already at the limit of the hardware. Lapping is really only used one of two ways.

1. A cheap way to get a couple more degrees of headroom for OCing stability.
2. To extract the maximum possible heat dissipation out of your already maximized cooling hardware.

So either it's a cheapest-method solution or the last step in the most-expensive-method solution.

Lapping the contact surfaces is VERY popular with water cooling users and/or overclocking (OCing). Every degree counts when you're OCing. I miss spending days on tweaking, tucking cables, testing airflow with smoke, stress testing, gap sealing... it's fun stuff, really.

 

[njedpx3]

Cool (literally) - So if I understand by lapping (smoothing and polishing) you aresignificantly increasing the actual surface area of the heat sink. Which makes sense more surface area, better ambient cooling.

Thanks for sharing the process!

Forest man

 

[Nemo128]

Cool (literally) - So if I understand by lapping (smoothing and polishing) you aresignificantly increasing the actual surface area of the heat sink. Which makes sense more surface area, better ambient cooling.

Significantly, varies by heat sink and it's debateable, increasing, certainly. Products like silver-based thermal compounds with thinner compositions also help.

Feel free to ask about anything cooling-related, I love the stuff and try to stay pretty up to date. I do a lot of testing in that area, I just don't get to enjoy it for my own purposes now.

 

[Shock]

Cool (literally) - So if I understand by lapping (smoothing and polishing) you aresignificantly increasing the actual surface area of the heat sink. Which makes sense more surface area, better ambient cooling.

Thanks for sharing the process!

Forest man

Anytime

 

[highfigh]

Not sure why this would even make a difference at all. Yes you get slightly more contact between the surfaces, but it does not make the heat sink itself any more efficient.

So, what part does the heat sink compound play? The minute difference in surface area can't matter. Also, more surface in contact is great but it's the surface area of the radiating area that matters. All the body of the sink does is carry the heat to the fins.

 

[Shock]

So, what part does the heat sink compound play? The minute difference in surface area can't matter. Also, more surface in contact is great but it's the surface area of the radiating area that matters. All the body of the sink does is carry the heat to the fins.

Surface Finish Impact

Here's the part I've been waiting to reveal... the importance of surface finish in relation to the impact on thermal conductivity. CPU coolers primarily depend on two heat transfer methods: conduction and radiation (heat-pipes also add convection). This being the case, let's start with conduction as it related to the mating surface between a heat source and a cooler.

Because of their density, metals are the best conductors of thermal energy. As density decreases so does conduction (of heat), which relegates fluids to be naturally less conductive, and gases as virtually non-conductive. So ideally the less fluid between metals, the better heat will transfer between them. Ultimately though, this means that the perfectly flat and well-polished surface (Noctua NH-U12P) is going to be preferred over the rougher and less even surface which required more TIM to fill the gaps (Thermalright Ultra-120 eXtreme).

Heat radiation is different however, and requires exactly the opposite. Because gases (air) are naturally poor heat conductors, surface area of the heatsink is key to the cooling performance through convection. This type of cooling is what you commonly see in automobile radiators, which utilize large arrays of metal fins to radiate heat to be drawn away by a fan. The same is true for the CPU cooler, which needs as much surface area as possible to optimize it's radiative effects. OCZ and others have recognized that the surface of a heatsink does not have to be the sum of its overall size. By adding dimples and bends, the surface area is increased without growing the overall dimensions.

To sum it all up, science teaches us that a smooth flat mating surface is ideal for CPU coolers so that less Thermal Interface Material is used. Because these coolers are using fans to force air over the heatsinks fins, the overall surface area of those fins should be as large and uneven as possible. In the next section we'll find out just how well all of these principles worked for our collection of test products.

That's taken from http://benchmarkreviews.com/index.php?option=com_content&task=view&id=285&Itemid=62&limit=1&limitstart=12

Both the conductive surface and the radiative surface matter in thermal dissipation.