CPU cooler for amplifier

[fmak]

I'd be a little concerned about the reliability of fans. From my Having said all that, a fan run at half voltage would probably have a long life, (but I'd make sure I installed a thermal cut out just in case.)
------------------------------------------------------------------------------

Another issue is the final S/N ratio you get. There is no point aiiming for -100 dB for a power amp, and compromising this with fan noise in the room - a kind of broader band hum plus wish!

[Circlotron]

I found that if the fan is right up against the heatsink fins (big heatsink I'm talking about now) it can be quite noisy, but if you place the fan inside a square duct about 100mm or more long and make the air flow along this before it reaches the heatsink the whole thing is much quieter even though the fan is still running at the original speed. Another source of noise is vibration from the magnetic pulsations of the motor that make a droning sound. This can be minimised by mounting the fan on something soft rather than bolting it down solidly, especially if it is to a surface that can vibrate.

[kt88]

Hi,

Leaving aside for a moment the reliability/noise issues. The October 1997 issue of "Elektor" (UK) carried an interesting article by H. Bonekamp on the use of CPU coolers for applications other than that intended for them ... Interesting reading if you can manage to get hold of a copy. Sorry, I can't scan it...

[Circlotron]

One other thing I was dreaming about is a fan muffler. Visualise a 6th order bass reflex box with a fan in place of where the bass driver would normally go, on the internal panel that divides the box into two chambers. Sound absorbent material to minimise fan noises. Then you have a duct that directs the air out to your heatsink. You could have the fan going full blast without much noise at all. I'll have to make one up to see how it goes. I'm on holidays now so I just might.

[jackinnj]

Analog (among others, but I only read their PDF) has a one chip reliability solution -- sensor and controller all in one.

I like the idea of water cooling. The early Cray's used off the shelf parts and water cooling, and, of course, Eimac's big bottles used liquid cooling.

[jackinnj]

QST from May 1966, I knew I had the article somewhere: Vapor phase cooling a 4CV1500B:

[Circlotron]

What might be interesting is a thermo-siphon system like they used on car cooling systems till about the 40's where they used a notably tall radiator and low engine.The heated water rises out of the engine and is sent to the top of the radiator where it is cooled and sinks to the bottom and returns to the engine. You could have a similar setup on your amplifier with an insulated pipe going up into the roof. Have the rising pipe insulated so the water stays warm till it gets to the heat exchanger, that way it will rise better. The higher the system is, the better it should work. No pump needed. Not as good as a pumped system but simpler and in theory more reliable. Of course if you live in a hot climate and the roof cavity is cooking hot it may not work very well.

...Actually, if you used a pumped system and about 2 square feet of car radiator core area, if you supply enough airflow e.g 200 kph you could dissipate about 150kW... I wonder if Nelson Pass has thought of that?