Some thoughts on alternative cooling

[Alexontherocks]

Dear All.... I do not know much about ss, tubes or audio in general. Just an amateur builder who is eager and happy to learn from thos who know and are willing to share.

I do however have some extensive experience in alternative methods of cooling especially in the area of PC, overclocking and benchtesting.

I would like to open a thread and submit some ideas I have (some currenly under testing) to alternatively cool large and powerful amps (like pure A pass amplifiers).

These are the main techniques I have tried with PCs

1. watercooling

2. TEC (peltier modules which need extensive and additional cooling)

3. Phase change

Computer cooling - Wikipedia, the free encyclopedia

Some of these techniques have obvious downsides. Phase change for example is extreme (multiple stages can achieve almost absolute 0 temperatures) and are bound to create interference both on the main lines, around the compressors.

However Liquid cooling, I believe, can effectively be used to help keep size, weight, cost and temeprature at bay.

Watercooling is based on the simple yet effective idea that water is used to transfer heat from the source to radiators which are designed and operate in conjuntion with fans (even undervolted) to eliminate heat.

The system usually works by using small electric 12v pumps which can be extremely effective, low noise, low power and still keep a good flow rate.





In some casewatercooling can yield asonshing results. I have personally rigged 2 audi tt radiators to passively cool an extremely overclocked CPU (putting out heaps of heat)





The heat is generate by a very small die which can put out several hundreds of watts of heat

(my e6600 with the HIS removed)

Now the essential part of the system is the waterblock.

This can bel made out in several shapes, sizes and materials. Obviousl specific designs are required for high efficiency systems which have to achiev tremendous performance in small packages.

For example this a modern waterblock:





however blocks can be much simpler....infact they can be as simple as this:



or even simpler...

right now I am experimenting with simple rectangular extrusions of pure copper wich will be interposed between an ordinary finned heatsink (like the ones we normally use in audio) and the common heatspreader of a generic array of devices




The extrsuion would behave as a simple channel for flowing water, offer almost no resistance and yet help to divert most of the heat to radiators possibily mounted on top or at the bottom of a chassis.

Imagine radiators like this:



In essence I am exploring the possibility of using watercooling as an effective way to deal with increasing power consumptions, costs and weight which for some of us can and do present serious problems.

If you have any thought on the matter please post here..

I will keep you updated with diagrams, photos of my work etc.

[Alexontherocks]

I forgot to mention that a 12v pump can be regulated thermically to provide stable tempertures under the most extreme situations. In turn temperature stability would provide longer component life, stable output figures and a more predictable device behaviour.

Obviously the amps can be made smaller in dimensions and still provide room and cooling capacity for large amplifier modules.

[Alexontherocks]

Here is what I am doing...










I believe this could be an effective design given that we would not be seeking low temperatures rather reducing the amount of heat exchanged by the ordinary finned heatsink.

[454Casull]

Can you not clamp the copper tube with the angle aluminum directly, so that you can bypass the aluminum bar? Or rather, get rid of the angle and keep the bar, but drill/tap it so that you can screw the chips into it. I see some ways you can reduce the thermal resistance there... Also, if you set up the loop properly, there will be no need for those fins.

Water cooling is definitely a solution more people need to explore, though.

[Andrew Eckhardt]

Once you go through the complexity and expense of a water system one of the offsetting features is the lack of need for convection cooling at all. Unless you are looking for failsafe protection, added convection cooling is a waste. It just makes it bulkier and more expensive.

As previously mentioned , you want the dissipation to be as close to the water as possible. Get rid of any additional interfaces if possible, especially angles or tabs that force conduction through a thin length of material.

[Alexontherocks]

Drilling on a pipelike structure is tricky and I would not do it unless an alternative is actually impossible...

However take into account the fact that low temperatures are not what we are looking for...infact we have two opposing objectives. Keep the temperatures high enough for optimal output (the multiple surfaces will actually work towards such a goal by providing a suboptimal situation for heat transfer) and obviously we seek some form of cooling which will be provided by the moving water. Also once the water temperature stabilizes the temperature itself will be regulated by the speed of the fans..

Other pics to show the rest of the project:

[Alexontherocks]

Originally Posted by Andrew Eckhardt

Once you go through the complexity and expense of a water system one of the offsetting features is the lack of need for convection cooling at all. Unless you are looking for failsafe protection, added convection cooling is a waste. It just makes it bulkier and more expensive.

As previously mentioned , you want the dissipation to be as close to the water as possible. Get rid of any additional interfaces if possible, especially angles or tabs that force conduction through a thin length of material.

I was actually thinking of a failsafe. In case waterpumps stop the water inside will act as a conductive core (a pretty good one)
So in essence some time will be pass before the underdimensioned convection heatsinks will fail and temperatures begin rising beyond tollerable levels.

I agree that less is better however dilling inside a tube structure is just not feasable.

I could look into this:
http://www.aavidthermalloy.com/sites...-hicontact.pdf

but really it would be less flexibile than just using extrusions cut to requirements.

[Alexontherocks]

Of course one could rig a flowmeter to powerdown the amp if flow goes under a specific value. Shouldn't be too hard!

[AndrewT]

can the four devices be clamped direct to the water pipe/tube by the bolted on clamp plate?

This dispenses with the angle completely.

Another clamp plate instead of the heatsink and one could clamp up twice as many devices with the leads just the thickness of the water tube apart (+3mm).

The usual temp switch on the sink to power down on overtemperature, with or without a latch.

[Alexontherocks]

Yes Andrew infact I was just working on this option... thanks for spotting this out!

This is a small rendering of what the case could look like once we reduce the dimensions to the bare minimum....