From: "Mark Hathaway"
References: <firstname.lastname@example.org> <email@example.com> <firstname.lastname@example.org>
Subject: Re: modeling a water cooling system
X-Newsreader: Microsoft Outlook Express 6.00.2600.0000
Organization: iPrimus Customer - reports relating to abuse should be sent to email@example.com
Date: Thu, 24 Oct 2002 21:42:48 +1000
NNTP-Posting-Date: Thu, 24 Oct 2002 07:33:13 EDT
No doubt the amps are off the deep end, but they do sound good. I'm planning
on water cooling my computer rack too, and would like zero noise (studio
conditions) and will probably tap into the amp cooling system if all goes
I've built many amps, including current feedback ones with a flat harmonic
structure, flat out to 900khz, and single ended class A mosfet amps sound
really nice to my ears. Possibly a case of less is more
Anyway, thanks for the confirmation on the numbers achievable from such a
heat exchange system.
"John Larkin" wrote in
> On Wed, 23 Oct 2002 01:10:56 GMT, Satan's Little Sister
> >On Tue, 22 Oct 2002 20:22:04 +1000, Mark Hathaway babbled:
> >My personal experience with water cooling puts the typical heater core
> >at about 30W/K, 1 gal/min water, 60 cfm air. There's only minimal gain
> >with more water/air. Typical arrangement is four parallel cores around
> >the perimeter of a plastic milk crate with a single central fan above
> >(ugly, but effective). A transmission oil cooler would work even
> >better for you.
> >If you are willing to do some digging (or have a well), ground sink
> >cooling would be ideal.
> Ideal air cooling (perfect coupling between the heat source and the
> exit air) gives about 1.75 K/W per CFM, or about 0.03 K/w for 60 CFM
> available air flow. And that's almost exactly what you're seeing. That
> implies that your situation is dominated by available air flow, and
> not by any other thermal resistances. So in your example, more air
> would substantially reduce the effective thermal resistance, but
> better cooler geometry couldn't help much.
> 1 GPM of water flow has an intrinsic thermal resistance of about
> 0.0037 K/W, so there's plenty of water flow in this example.
> (but, as someone pointed out, the amplifier in question is insane)