Wind Tunnel or Open Air cooling better?

[The little lost angel]

Which is a better cooling solution for electronics components?

1) Components mounted on a heatsink, open air with fan(s) blowing
across the heatsink.

2) Components mounted on a heatsink, inside a tunnel/box with fan(s)
blowing in from one end. Area of opening is the same as the area of
the fan(s)

Or is there no difference as long as everthing else (fan speed,
heatsink size, etc) remains unchanged?

Thanks!
--
L.Angel: I'm looking for web design work.
If you need basic to med complexity webpages at affordable rates, email me
Standard HTML, SHTML, MySQL + PHP or ASP, Javascript.
If you really want, FrontPage & DreamWeaver too.
But keep in mind you pay extra bandwidth for their bloated code

 

[John Larkin]

Which is a better cooling solution for electronics components?

1) Components mounted on a heatsink, open air with fan(s) blowing
across the heatsink.

2) Components mounted on a heatsink, inside a tunnel/box with fan(s)
blowing in from one end. Area of opening is the same as the area of
the fan(s)

Or is there no difference as long as everthing else (fan speed,
heatsink size, etc) remains unchanged?

Thanks!

Whichever moves the highest velocity air *through* the heatsink fins.
Since air would prefer to go *around* the fins, don't let it.

Impingement is good: fan blasting directly down onto the sink, instead
of blowing across it.

I have this (unproven) theory that air flow should be reduced to 50%
of the fan's max CFM capability by the fin's flow resistance.

Pin fins suck.

John

 

[Ian Stirling]

Whichever moves the highest velocity air *through* the heatsink fins.
Since air would prefer to go *around* the fins, don't let it.

Work out the amount of air you need to carry the heat away.
This is (very roughly) 1J/K/l.
1l/s heated by 100C will carry away 1W of heat.
This gives you an absolute lower figure for airflow.

Impingement is good: fan blasting directly down onto the sink, instead
of blowing across it.

If you blow cold air from the bits of the heatsink that are naturally coolest
to those that are hottest then you will generally get better results than
the other way round.
This is as it will encourage more heat to make it out to the edges of the
heatsink, by maximising temperature gradients.

I have this (unproven) theory that air flow should be reduced to 50%
of the fan's max CFM capability by the fin's flow resistance.

And if the temperature goes up when you turn on the fans, you'r blowing
too hard...

 

[The little lost angel]

Work out the amount of air you need to carry the heat away.
This is (very roughly) 1J/K/l.
1l/s heated by 100C will carry away 1W of heat.
This gives you an absolute lower figure for airflow.

I'm sorry, but what's l?

If you blow cold air from the bits of the heatsink that are naturally coolest
to those that are hottest then you will generally get better results than
the other way round.
This is as it will encourage more heat to make it out to the edges of the
heatsink, by maximising temperature gradients.

Does that mean, given a longish heatsink with rectangular fins running
parallel along the length, I would get better results if I blow air
frmo both ends (naturally cooler end) and let the air meet in the
center and out the top/bottom?

--
L.Angel: I'm looking for web design work.
If you need basic to med complexity webpages at affordable rates, email me
Standard HTML, SHTML, MySQL + PHP or ASP, Javascript.
If you really want, FrontPage & DreamWeaver too.
But keep in mind you pay extra bandwidth for their bloated code

 

[The little lost angel]

Whichever moves the highest velocity air *through* the heatsink fins.
Since air would prefer to go *around* the fins, don't let it.

Impingement is good: fan blasting directly down onto the sink, instead
of blowing across it.

Hmm, if mounting the fan from the side so that it blows down the
length of the fins moves more air, compared to fan blasting directly
down. Which will then have a better cooling effect?

--
L.Angel: I'm looking for web design work.
If you need basic to med complexity webpages at affordable rates, email me
Standard HTML, SHTML, MySQL + PHP or ASP, Javascript.
If you really want, FrontPage & DreamWeaver too.
But keep in mind you pay extra bandwidth for their bloated code

 

[John Popelish]

The little lost angel wrote:
(snip)

Does that mean, given a longish heatsink with rectangular fins running
parallel along the length, I would get better results if I blow air
frmo both ends (naturally cooler end) and let the air meet in the
center and out the top/bottom?

The effect is very similar if you push air in from the ends of the
fins and have it stall and ooze out at the middle, or if you push it
against the middle and have it split and go out the ends of the fins.
Both these methods have low velocity areas that do little to help
transfer heat. If you want ot get maximum transfer to a given air
flow, cover the fins, to make a parallel group of ducts, and push the
air straight across. This maintains the highest velocity and thus,
thinnest boundary layers. This falls down only if the air flow is so
low that the exiting air approaches the sink temperature and stops
transferring heat at some point along the ducts. This method does
require a bit more fan pressure than just blowing air at the flat side
of the sink. Centrifugal fans work well in this service.

 

[Roy McCammon]

Hmm, if mounting the fan from the side so that it blows down the
length of the fins moves more air, compared to fan blasting directly
down. Which will then have a better cooling effect?

if you construct a tunnel as John Popelish suggest, it has maybe twice
the heat removal for a given fan.

 

[Ross Mac]

Which is a better cooling solution for electronics components?

1) Components mounted on a heatsink, open air with fan(s) blowing
across the heatsink.

2) Components mounted on a heatsink, inside a tunnel/box with fan(s)
blowing in from one end. Area of opening is the same as the area of
the fan(s)

Or is there no difference as long as everthing else (fan speed,
heatsink size, etc) remains unchanged?

Thanks!
--
L.Angel: I'm looking for web design work.
If you need basic to med complexity webpages at affordable rates, email me
Standard HTML, SHTML, MySQL + PHP or ASP, Javascript.
If you really want, FrontPage & DreamWeaver too.
But keep in mind you pay extra bandwidth for their bloated code

Looks like John Popelish has the best system here....
I would just like to add that the anytime you have a fan, you get dirt
contamination so you may want to consider a good replacable filtration
media.....good luck...Ross

 

[The little lost angel]

transfer heat. If you want ot get maximum transfer to a given air
flow, cover the fins, to make a parallel group of ducts, and push the
air straight across. This maintains the highest velocity and thus,
thinnest boundary layers. This falls down only if the air flow is so
low that the exiting air approaches the sink temperature and stops
transferring heat at some point along the ducts. This method does
require a bit more fan pressure than just blowing air at the flat side
of the sink. Centrifugal fans work well in this service.

I can get axial fans much much cheaper (almost free), would it be the
same if I mount say 4 of them in a 2x2 arrangement? I'm not 100% sure
exactly what you mean by parallel ducts, but does it mean something
like this:

|_|_|_|

Heatsink and fins view from the long ends. Then cover them up to get
this:
______
|_|_|_|


Would it help solve the problem of the exiting air flow being slow if
I add another 2 or 4 fans at the exit end drawing air out?

Thanks again!

--
L.Angel: I'm looking for web design work.
If you need basic to med complexity webpages at affordable rates, email me
Standard HTML, SHTML, MySQL + PHP or ASP, Javascript.
If you really want, FrontPage & DreamWeaver too.
But keep in mind you pay extra bandwidth for their bloated code

 

[John Popelish]

I can get axial fans much much cheaper (almost free), would it be the
same if I mount say 4 of them in a 2x2 arrangement? I'm not 100% sure
exactly what you mean by parallel ducts, but does it mean something
like this:

|_|_|_|

Heatsink and fins view from the long ends. Then cover them up to get
this:
______
|_|_|_|
Yes.

Would it help solve the problem of the exiting air flow being slow if
I add another 2 or 4 fans at the exit end drawing air out?

The problem is one of pressure drop as the air is forced through those
long, narrow passages. Axial fans produce a lot of flow for their
size, but only at a small pressure drop. The same size centrifugal
fan produces less flow, but a higher pressure head.

So a better solution than paralleling several is stacking several in
series in a duct. to combine their pressure capabilities if one does
not produce enough head to handle the drag pressure of all those
narrow passages.

You want a fan that has a clear area (area between outer housing and
motor hub) about twice the total duct area (sum of the cross sections
of all those passages between fins) so the fan does not use up a
significant amount of its pressure just getting the air through
itself, and a tight air duct that surrounds the fan and forces all its
flow to the sink without much in the way of leaks around the fan or to
the outside.

This works about the same whether the fans push air in or suck it out
or both. The fans downstream of the heat sink will have a shorter
life because of the hot air passing through them. The advantage of
downstream fans is that their motor heat output does not warm the air
stream until after it passes through the sink.