[hpsdr] Munin chassis considerations?
Glenn Thomas
glennt at gbis.com
Tue Nov 5 13:02:52 PST 2013
I think the way to answer your question is to look at how the heat sink
disposes of thermal energy.
Presumably (I haven't seen the specific Munin heat sink configuration),
it transfers heat to air which then moves the heat elsewhere via
convection. IMHO, mounting the heat sink with the long length of the
fins vertical would be best because it will result in an increased
pressure differential between the top & bottom, leading to improved
convective air flow and thus improved cooling.
Short fin dimension vertical does not provide as much convection, partly
because the air spends less time in contact with the fins as it rises
and because the supply of cool air is partly blocked by the fins and
outer vertical convection.
Some years ago, I had the same question. My ignorant approach was to put
a big power transistor(2N3055 I think) in a heat sink and bias it to
make a lot of heat. With the heat sink fins short dimension vertical,
they became almost too hot to touch. With the heat sink fins long
dimension vertical, they were warm, but not uncomfortably so.
<RANT>
A better way? Replace the air as your convective fluid with something
else that has higher heat capacity. Air specific heat is somewhere
between 1.0 and 1.3 (kJ/kG K). Alternatives include methane at 2.2,
ammonia at 2.9, helium at 5.19 and hydrogen at 14.32 (!!). Some of these
gasses have other characteristics (toxicity, odor, explosiveness,
tendency to escape) that recommend against them.
If you're going that route anyway, perhaps a liquid cooled system is
worth consideration. The heat capacity of potential cooling fluids
(again in kJ/kG K) include castor oil at 1.8, vodka around 2.0, ethylene
glycol at 2.36, glycol at 2.5, methyl alcohol at 2.51, propylene at
2.85, milk at 3.93 and fresh water at 4.19. BTW, Vodka gets a little
better as it gets warmer. Take your pick.
You also will need a closed cooling system. Convection a'la thermo
siphon can be used to move the fluid. A pump is not necessarily required.
I leave a comparison of vapor phase cooling fluids as an exercise for
those who are interested.
I do know a worse way. That would be to put the heat sink fins, long
axis horizontal, underneath the electronics. Sort of like the 100 watt
version of the Flex-1000. That way, you not only impede convective
cooling with horizontal fins, the heated air has to flow past - or even
worse through - the circuit, thus transferring some heat back into the
circuit components.
</RANT>
Please don't hesitate to duplicate my experiment, YMMV.
Or use a large, quiet, slow-speed fan. Then horizontal/vertical probably
doesn't matter.
73 de Glenn WB6W
On 11/5/2013 11:24 AM, John wrote:
> ***** High Performance Software Defined Radio Discussion List *****
>
> I'm looking for suggestions on the best way to mount Munin in an
> enclosure. I may make my own enclosure. The fins of the heatsink
> would not be covered up by the exclosure. I don't plan to use a fan.
>
> For the best heat removal, does it make sense to mount the heatsink so
> the long length of the fins is vertical? Or should it be mounted so
> the circuit board is up-side-down (heatsink on top, short dimension of
> fins vertical)?
> Or is there another, better, way?
>
> John, WoGN
>
--
Speaking of recursion, where did the cynics go wrong?
1383685372.0
More information about the Hpsdr
mailing list