Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.

ayeying

macrumors 601
Original poster
Dec 5, 2007
4,547
13
Yay Area, CA
Okay, I’m going to admit. At first, this idea seems a bit farfetched and probably completely unneeded. Therefore, I’m going to go ahead and explain a bit of background of this idea and whether or not you agree is completely up to you.

The copper mod was first heard about with the old Dell Inspiron 8500 and 8600 notebook PCs. During those laptops, the high end video card option was the ATi Radeon Mobility 9600 Pro Turbo, a fairly powerful midrange video card of its time. The Inspiron 8500 was Pentium 4 Mobile based while the 8600 was based off the Centrino chipset with a Pentium M processor (one of the early generation chips). The ATi Radeon Mobility 9600 Pro Turbo was a powerful card, because of it, a lot of heat was generated and it was fairly common for this card to overheat and shut down the entire computer. The problem was discovered that the heatsink designed for the ATi Radeon Mobility 9600 Pro Turbo was the same as for the slower, less heat generating cards such as the nVidia GeForce FX5200. The heatsink itself however isn’t a flaw of the design, it’s the heat transfer. The heat generated by the diode and ram chips were carried to the heatsink by a very inferior thermal pad of a few millimeter thick layer. Because of this, using a “thin” layer of thermal paste was out of the question since it would never touch the heatsink in the first place. Therefore, a copper plate was introduced to bridge that gap. This introduction of copper to an aluminum stock heatsink resulted in close to 20+ deg C heat reduction on several accounts.

Now, while the MacBook Air’s heatsink does touch the diode, the heat transfer is still inferior due to its size.

Here’s how the MacBook Air’s heatsink gets rid of the heat created by the diodes. The heatsink, which is essentially a plate of aluminum for the Rev B and Rev C or a heat shield wrapped in aluminum foil for Rev A, draws the heat from the diode and heats up the entire plate. The plate is cooled by a small fan that creates an air flow. This flow is basically sweeping the heat on the plate out the vents in the back.

Aluminum is a great heat radiator, meaning it is great at getting rid of heat fast. However, it’s not the best conductor, which draws the heat from the source.

Now heat is energy and the aluminum plating can only carry so much heat before it is full. The heat generated under full load from either the CPU or GPU can easily exceed what the aluminum plate can handle very quickly.

The copper mod is designed to bridge a gap in which heat can be conducted faster from the source, but still allow enough time for the radiator to radiate the heat away. This is done in which the copper is placed between the Diode and Heatsink. The dense mass of the copper is allowed to store more heat while still allowing the radiator to radiate the heat quickly.

So why not build a heatsink completely from copper? There’s several reasons for that, one is that copper is a terrible radiator. Copper loves to store heat, not radiate it out. Aluminum however hates to store heat. It loves to radiate heat as fast as possible but due to its lightweight and low mass it cannot store that much heat before its full.

Look at it from all the other heatsinks, even the ones designed in the MacBook/MacBook Pros. The base, in which touches the diode are almost always copper. However, the fins at the end of the base are usually a combination of Copper and Aluminum.

This is what I’ve done with my MacBook Air, pictures attached.

I applied a thin layer of thermal paste on the diode, then placed a thin piece of copper on top of it. From there, I applied a thin layer of thermal paste on the heatsink and squished the copper in between. This combination allows the copper (conductor) to effectively draw heat from the source (diode), and still allow the aluminum (radiator) to radiate heat away.

Under 100% load testing (Folding@home and nothing else), the fans have stayed ~4500-4800 RPM with temps ranging around 70-78 deg C

However when surfing the net with Folding@home, the fans will rev back up to 6200RPM.

But even with the fans revving to 6200RPM quickly, it is delayed longer than stock and the fans return back to 2500RPM quicker.
 

Attachments

  • IMG_0075.jpg
    IMG_0075.jpg
    672.1 KB · Views: 779
  • IMG_0077.jpg
    IMG_0077.jpg
    700.6 KB · Views: 728
  • IMG_0078.jpg
    IMG_0078.jpg
    934.5 KB · Views: 953
  • Like
Reactions: zafarella
well can you post instructions like dimensions of the copper plate, where to buy the supplies, etc... especially for those who are less technically inclined? Also how much was everything?
 
Nah of course not, Apple stimulates its buyers to perform these mods :p

Nice mod indeed, but I'm afraid it will be hard to find a perfectly flat piece of copper that has the right dimensions.

That is my concern. The copper shown in the pictures appears to be basically a foil. Usually heat sinks are very flat (sometimes honed to be as flat as possible to allow a very consistent contact/heat transfer, any gaps impede heat transfer). In order to accurately gauge the success of this mod (or any other) would be to precisely time how long the fan runs and temp reduces at a given website, before and after the mod (at the same room temp). Just saying But even with the fans revving to 6200RPM quickly, it is delayed longer than stock and the fans return back to 2500RPM quicker." is basically guessing.

But the whole idea is a good one and very creative. If the data shows an improvement, then congrats!
 
im working on a new heatsink design myself, my buddy will machine it for me out of copper.

im looking into extending the heatsink a bit to reduce the total heat,

the heatsink must be pure copper to make this work well, probe is when copper gets thin its weak, not much room to work with either.

there has to be a better way, directing the air has a good temperature lowering effect for me, only if i could build a plastic duct for it.
 
well can you post instructions like dimensions of the copper plate, where to buy the supplies, etc... especially for those who are less technically inclined? Also how much was everything?

Well, the dimensions are actually a guess. My dad works as a machinist so copper is in plentiful supply for me. I'd guess you can probably find copper sheets online and cut them to size. Hammering them flat is a problem though.

nice work...:)

btw didn't you void your warranty?:(

Not really. It's not glued onto the heatsink like I wanted to. It's just squished like a sandwich. Furthermore, I had apple replace my logic board even with new thermal paste on it.

im working on a new heatsink design myself, my buddy will machine it for me out of copper.

im looking into extending the heatsink a bit to reduce the total heat,

the heatsink must be pure copper to make this work well, probe is when copper gets thin its weak, not much room to work with either.

there has to be a better way, directing the air has a good temperature lowering effect for me, only if i could build a plastic duct for it.

I'd suggest not creating a pure copper heatsink. I've tried it before and the results were worse than using the stock heatsink. Copper itself isn't a good radiator but it is a great conductor. It's literally grabbing the heat away from the diode and storing it rather than getting rid of it. However, after the entire copper is full, then the continuing heat generated by the diode will be blanketed with more heat.

I went with a Copper/Aluminum solution because i wanted a buffer to store the heat while the Aluminum got rid of it.

The copper does look a bit warped, however, it's as flat as I can get it with conventional means. Furthermore, the pressure applied by the "L" plate further flattens the copper.

That is my concern. The copper shown in the pictures appears to be basically a foil. Usually heat sinks are very flat (sometimes honed to be as flat as possible to allow a very consistent contact/heat transfer, any gaps impede heat transfer). In order to accurately gauge the success of this mod (or any other) would be to precisely time how long the fan runs and temp reduces at a given website, before and after the mod (at the same room temp). Just saying But even with the fans revving to 6200RPM quickly, it is delayed longer than stock and the fans return back to 2500RPM quicker." is basically guessing.

But the whole idea is a good one and very creative. If the data shows an improvement, then congrats!

There was a problem with being flat too. For some reason, the stock heatsink has a gap of about 0.5mm even with thermal paste on and L plate applying pressure. I'm not sure why, but with the stock heatsink, the diode isn't really touching the heatsink at all for me. The Copper gaps this barrier and I know it looks like a foil but its a thin piece of copper. It's as flat as I can get with just a regular hammer and flat surface. I'm still trying to find a way to create two more and flatten as much as I can.
 
There was a problem with being flat too. For some reason, the stock heatsink has a gap of about 0.5mm even with thermal paste on and L plate applying pressure. I'm not sure why, but with the stock heatsink, the diode isn't really touching the heatsink at all for me. The Copper gaps this barrier and I know it looks like a foil but its a thin piece of copper. It's as flat as I can get with just a regular hammer and flat surface. I'm still trying to find a way to create two more and flatten as much as I can.

I think your best bet would be to lap the surfaces flat, not pounding it. Make a holder and rub each side of the copper on superfine sandpaper, or if you have access to it, a lapping machine with lapping compound.
 
I had in mind something similar, but instead use copper foil for the entire bottom of the mac. Dunno if there any special glue with good conductivity is available.
 
  • Like
Reactions: Wackery
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.