Modding the HSF : copper, wing, resistor, AS 5

My AA1 is back from death and I start working on the cooling unit. BUT, before he dies, he was silent ! Now, the fan run most time while I haven't touch anything about cooling. It looks strange !!

First of all, I run Prime95 fanless with the original HSF. Using Everest or any software witch read the die temp brings 95°C. So, it's not really possible to go fanless all the time mostly if you plan gaming. If you use AAFC or some other softwares, they repport ACPI temp witch is an average of some temps and the value if far lower (from 10 to 20°C).

1/ I remove the HSF and get the blue pad of. Since space is too big for thermal grease, I cut a 20x20 mm from a old floppy 3.5". I put some AS5 on both side and run Prime95 again fanless : 92°C.

2/ I get some parts from a Thinkpad HSF and get the radiator. A part looking like this |_|_|_|_|_|_|_|. With AS 5, I fix it on the air duck and run Prime95 : 90°C.

Now, I plan to romove a small aluminium part on the back of the air duck to replace it with a large copper one. But I don't have time now... I have the same fan connector from a Thinkpad and I will solder a resistor to seen if I can lower the fan rpm.

I think with some improvement on the air guide, AA1 can run heavy load with fan at very very low RPM.

 

Copper pulls heat faster than aluminum but also retains it more
So you'll want to use copper on the CPU/NB and aluminium for dissipeating the heat

But i'm not sure which one is better in passive setup though..

 

hmm, the Dell runs fanless and only has the metal plate underneat the keyboard, just as ASUS's EEEPCs, but without the fan (or the fan disconnected)

Has anyone tried to short the rpm sensing to ground see if it boots? As I've read somewhere that the Aspire One refuse to boot with the fan disconnected
But then, Acer's design have the chips on the bottom...

 

when you start looking at this, you might want to see about mixing your own thermal compound using diamond dust and silicone grease. it's 3x better than the best arctic silver or w/e they have out now.

also maybe using the ionic blowers over the traditional fan? it would take a bit more time to fabricate, and I don't know if it would ultimately use more energy, but it's a possibility.

using the diamond thermal compound and a low speed, yet efficient, fan combined with a heat pipe and copper fins should be able to dissipate the heat far better than the current system.

maybe look at getting some of those copper finned memory chip heat sinks. cut the necessary holes in the metal plate (that sits on the CPU), place the heat sinks on the given chips, and you still have a "duct" via the existing plate heat sink.

I've been thinking about modifying the current cooling system, as it's just not the best solution. it allows too much heat to build up within the system, and then the fan is constantly on. mine used to be quiet, but now the fan makes considerable noise and is constantly on. the goal is to have a system that can quickly absorb and then dissipate the heat from the chips. having an aluminum plate with a flat surface that air blows across is not conducive to achieving this goal.

a co-worker has an old sony vaio with some heat pipes and small fans. he's no longer using it as he bought a new macbook pro. I'll see what I can scavenge from it.

 

WIth the existing cooler, it is not possible to just use thermal compound to replace the thick thermal pad, so we will need a piece of metal to sandwich it between the core and the HSF

 

You need more than a simple piece of metal. Effective heatsinks require as close to a perfect mating of surfaces as possible. Typically, that mean contacting two absolutely flat surfaces, with thermal compound to fill in the microscopic crevices The thick thermal pad you abhor creates the thermal transfer medium necessary to create the contact between the processor and the heatsink.In order to get the same kind of interface with an additional piece of metal, you would need to bolt down the additional dissipater. For reference, see any heatsink setup in use: they're all held down with substantial force and the best of them use backplates with torque-down bolts instead of simple springs to ensure that the best contact is made between the metal. To reduce the space between the blocks to an absolute minimum for an effective transfer, you'd probably end up having to drill through the PCB (don't do this) or improvise your own extremely high tension spring
You would also need to apply an additional layer of thermal compound. I don't know what your experience with AS5 is, but unless it's acting strictly as a crevice-filling medium, it will actually reduce the thermal transfer into the heatsink. Excessive AS of any variety has been well-documented as a thermal insulator when it's too thick. Additional applications increase the resistance and the chance that too much has been applied. This is why watercooling setups directly the flow directly through the block instead of simply mounting into it - the additional transfer interface renders your best efforts a moot point.

You may not like thermal pads, but they're quite good at what they do, some are quite efficient given the space constraints. Unless you're planning to weld that extra piece of metal to the heatsink, I'd suggest leaving it as it is or investing your time into finding a higher quality pad.

All that having been said, if you can find a heatpipe setup that will fit within the AA1 casing with minor modifications, I'm all ears. Heatpipes have been used to move heat in laptops long before they appeared in desktop heatsinks. However, unless you're willing to dismantle a subnotebook, it'll be a real challenge to find one that will fit, since the typical laptop's internal configuration has the heatsink and the piping extending all over the interior to cover the CPU, northbridge, and graphics. Most simply won't fit in the AA1 casing.

As for the ionic blowers, I didn't realize those had even passed the concept phase, let alone prototype or production...