An experiment with intake turbulence

[Timothy Daniels]

Here's an interesting experiment in ATX case cooling done
by a modder with a website called "BenchTest.com" -
http://www.benchtest.com/way2cool3.html . During his experi-
ments, he made a cutout for a fan in the front lower portion of
the metal case, centered among the existing vent holes. The
fan pulled air into the case and it was exhausted by the power
supply fan and a rear case fan.

At one point, he inadvertently placed a box within 1" of the
front intake fan, partially restricting the air intake by forcing all
the incoming air to travel parallel to the face of case before
entering the fan and the vent holes instead of entering the fan
and vent holes straight in. (See the section with the heading
"Holes in the Face Plate".) He was surprised to find that the
case temperature dropped!

He is still mystified by this. But it could well have been caused
by INCREASED AIR TURBULENCE due to ALL the air having to
make a right angle turn before entering the fan and its surrounding
vent holes.

*TimDaniels*

 

[kony]

Here's an interesting experiment in ATX case cooling done
by a modder with a website called "BenchTest.com" -
http://www.benchtest.com/way2cool3.html .

It was passingly interesting enough to comment on when you
posted it in 2004, but today is more of a sad tale of how
wrong the fellow was in almost every speculation he made.

Either he had one of his fans backwards (as I'd mentioned in
2004), or we have to consider that his case temp sensor is
only a measurement of one area that doesn't reflect a
similar change in all other areas. Merely altering flow
rates, direction in individual areas can easily cause a gap,
crack or hole to result in airflow IN past the sensor,
instead of OUT past the sensor, if not another more complex
change in flow patterns.

Recall that he was trying to balance intake and exhaust
rate, and that I'd commented in 2004 that this was NOT the
correct goal- that we WANT uneven rates to minimize dead
zones (or zones nearer to that, having lower flow rate).

Only by taking multi-point readings in the case and of all
the parts would we really we what was happening. The only
thing we can be sure of from the linked article is that he
ended up having a hard time cooling what is by today's
standards, a very cool running system. Adding the very loud
4.5" AC fan and squirrel cage were quite overkill.

 

[meow2222]

Recall that he was trying to balance intake and exhaust
rate, and that I'd commented in 2004 that this was NOT the
correct goal- that we WANT uneven rates to minimize dead
zones (or zones nearer to that, having lower flow rate).

Surely thats one goal at which one is guaranteed to succeed. If intake
and exhaust airflow are not balanced, either they soon will be or you
have a bomb!


NT

 

[Jon Danniken]

He is still mystified by this. But it could well have been caused
by INCREASED AIR TURBULENCE due to ALL the air having to
make a right angle turn before entering the fan and its surrounding
vent holes.

I smell a kook in the house.

Jon

 

[kony]

Surely thats one goal at which one is guaranteed to succeed. If intake
and exhaust airflow are not balanced, either they soon will be or you
have a bomb!

True, ultimately intake rate and exhaust rate are equal.
However, all of the intake and exhaust in a typical case is
not through fan openings, there's also lots of cracks,
crevaces, holes, etc, and some devices even depend on the
imbalance to stay cool like optical drives with slits on
their bezels.

Unless one actively tries to cause their intake and exhaust
fans to have same (resultant, considering impedance to
airflow at each point) rate, they will usually be different
rates due to size, RPM, pressure zones, etc, such that the
case is slightly positive or negatively pressurized relative
to the room.

 

[meow2222]

Prop fans introduce plenty of turbulence anyway

NT

 

[Timothy Daniels]

"kornball" continued his denials:

It was passingly interesting enough to comment on when you
posted it in 2004, but today is more of a sad tale of how
wrong the fellow was in almost every speculation he made.

The experimenter's speculations are unimportant, and his
wisdom is unimportant. But what he did certainly caused
and increase in the turbulence. What you dispute is whether
he measured the temperature properly of some part that is
of some critical or particular interest to you, not that turbulence
at the air intake didn't help cooling.

*TimDaniels*

 

[kony]

"kornball" continued his denials:


The experimenter's speculations are unimportant,

Oh?

Wasn't one of those speculations the whole reason you posted
it?

and his
wisdom is unimportant.

Untrue, if he overlooked variables in his system as I'd
described previously, the effect of a single point reading
when chassis airflow is altered.

But what he did certainly caused
and increase in the turbulence.

Appreciably? Probably not, putting a solid object in front
of an actively spinning fan is going to introduce trivial
addt'l turbulence over that already being produced by the
fan.

In fact, it probably DECREASED turbulence, because the minor
obstruction decreased the intake rate of the fan.

What you dispute is whether
he measured the temperature properly of some part that is
of some critical or particular interest to you, not that turbulence
at the air intake didn't help cooling.

See above, it's just as likely he had a corresponding
decrease in turbulence as with airflow from that scenario,
but we saw that when it was predominately turbulence instead
of airflow added (prior to when the large hole was cut out),
it didn't help nearly as much, the minor change beforehand
could be related to the minor increase of flow rate through
the several tiny holes.

It still doesn't account for the change in a specific area
where the sensor was. You have yet again prematurely tried
to conclude something, to support a prior premature
conclusion you'd made.

 

[Timothy Daniels]

Oh?
Wasn't one of those speculations the whole reason you posted
it?

Not at all. I pointed out what he did and what he observed.
I didn't at all what his speculations were. What I said was
quite clear:

"He is still mystified by this. But it could well have been caused
by INCREASED AIR TURBULENCE due to ALL the air having to
make a right angle turn before entering the fan and its surrounding
vent holes."

The experimenter himself has not idea what caused the
drop in temperature, and he doesn't even realize that what
he did caused an increase in turbulence at the air intake.

Untrue, if he overlooked variables in his system as I'd
described previously, the effect of a single point reading
when chassis airflow is altered.

No experiment, if it contradicts YOU, would be of sufficient
rigore to satisfy YOU.

Appreciably? Probably not, putting a solid object in front
of an actively spinning fan is going to introduce trivial
addt'l turbulence over that already being produced by the
fan.

He caused the incoming air to encounter the fan blades
with low axial velocity. That is to say, the air was not
traveling in the direction that the fan wanted the air to go
when the fan blades cut into the air. The air then had to
be accelerated from nearly zero axial velocity up to what
the fan blade's speed would try to induce. In an airplane
wing, that can cause a stall due to the increased turbulence
and the breakdown of laminar flow. Similarly, the increase
in volume of air entering from the edges of the front cover
caused turbulence as it encountered the air coming in the
front and had to make a right turn. If one wanted to encourage
smooth laminar flow into the case, one would have done
exactly the opposite.

In fact, it probably DECREASED turbulence, because the minor
obstruction decreased the intake rate of the fan.

That is the poorest conclusion you could make. The
change in direction of the incoming air so that it did
not enter the fan in an axial direction increased the
production of turbulence, and by conservation of
kinetic energy, that implies a reduction in bulk flow.
The increased turbulence, although it reduced the
flow rate, caused an observed drop in temperature.
Obviously, the turbulence counts more for cooling
than bulk fow rate.

See above, it's just as likely he had a corresponding
decrease in turbulence as with airflow from that scenario,
but we saw that when it was predominately turbulence instead
of airflow added (prior to when the large hole was cut out),
it didn't help nearly as much, the minor change beforehand
could be related to the minor increase of flow rate through
the several tiny holes.

Despite your convoluted grammar, the reduction in
bulk flow rate was not accompanied by a reduction
in cooling. In fact, the opposite was observed. When
the obstruction was put in front of the case, causing
the incoming air to have to make a right angle turn,
more turbulence was generated. If that decrease in
bulk flow rate dropped the case pressure it would have
caused more air flow to enter from the sides, but not
more air flow than was decreased by the front obstruction.
That is by simple conservation of energy: You don't get
*more* air flow through the case by increasing the drag!
Rather, by having more air enter from the sides and
having to make a right turn to transit the case, more
turbulence was introduced.

It still doesn't account for the change in a specific area
where the sensor was. You have yet again prematurely tried
to conclude something, to support a prior premature
conclusion you'd made.

The turbulence increased the cooling of the sensor!
What more do you want?


*TimDaniels*

 

[kony]

On Sat, 14 Oct 2006 20:37:30 -0700, "Timothy Daniels"


Tim,

It seems we all had our holes installed backwards, after
flipping them over the systems don't even need fans or
heatsinks anymore. How can we thank you enough?

You are 100% right, a Font Of Supercooling.

Mind if we call you "100% FOS" for short?

 

[Paul]

Here's an interesting experiment in ATX case cooling done
by a modder with a website called "BenchTest.com" -
http://www.benchtest.com/way2cool3.html . During his experi-
ments, he made a cutout for a fan in the front lower portion of
the metal case, centered among the existing vent holes. The
fan pulled air into the case and it was exhausted by the power
supply fan and a rear case fan.

At one point, he inadvertently placed a box within 1" of the
front intake fan, partially restricting the air intake by forcing all
the incoming air to travel parallel to the face of case before
entering the fan and the vent holes instead of entering the fan
and vent holes straight in. (See the section with the heading
"Holes in the Face Plate".) He was surprised to find that the
case temperature dropped!

He is still mystified by this. But it could well have been caused
by INCREASED AIR TURBULENCE due to ALL the air having to
make a right angle turn before entering the fan and its surrounding
vent holes.

*TimDaniels*

Air cooling is a well understood science.

http://www.electronics-cooling.com/html/2004_may_a2.html

All it takes to understand it, is the appropriate
university degree, and a lifetime of experience.
Which is why I wouldn't attempt to dabble in it
myself.

Trying to guess at what is happening inside a complex
3D space, with a compressible fluid (air), is pretty
pointless. Real thermal designers try to simplify their
test cases, to something they can hope to understand.
And then extrapolate from there, to build real designs.
No one I've ever worked with, would consider
building a CFD (Computational Fluid Dynamics) model
for something as complex as the inside of a PC. They
would have to retire, before it was completed.

Using the terminology on the above linked page, may
enable you to search and find technical articles analysing
simple structures, and make whatever kind of statement
it is, that you are trying to make. For example, this
sentence addresses what makes impingement cooling
effective:

http://www.keepmedia.com/pubs/PowerElectronicsTechnology/2001/02/01/151747?extID=10026

"Blown directly at the top of the pins, forced air
creates turbulent air between the pins, breaking the
air boundary layers around the pins and enhancing the
heat sink's high convectional thermal coefficients."

Paul

 

[Ed Medlin]

Oh?

Wasn't one of those speculations the whole reason you posted
it?


Untrue, if he overlooked variables in his system as I'd
described previously, the effect of a single point reading
when chassis airflow is altered.


Appreciably? Probably not, putting a solid object in front
of an actively spinning fan is going to introduce trivial
addt'l turbulence over that already being produced by the
fan.

In fact, it probably DECREASED turbulence, because the minor
obstruction decreased the intake rate of the fan.



See above, it's just as likely he had a corresponding
decrease in turbulence as with airflow from that scenario,
but we saw that when it was predominately turbulence instead
of airflow added (prior to when the large hole was cut out),
it didn't help nearly as much, the minor change beforehand
could be related to the minor increase of flow rate through
the several tiny holes.

It still doesn't account for the change in a specific area
where the sensor was. You have yet again prematurely tried
to conclude something, to support a prior premature
conclusion you'd made.

The fact that he had the front bezel off the case could have allowed more
passive air to flow too....... Just another likely scenerio. Blocking the
fan would only restrict airflow like you say..

Ed

 

[Timothy Daniels]

"Ed Medlin"

The fact that he had the front bezel off the case could have allowed
more passive air to flow too....... Just another likely scenerio. Blocking
the fan would only restrict airflow like you say..

With the FRONT COVER OFF, the experimenter tested with the
front air intake
1) unobstructed by the box, and
2) obstructed by the box.
He found lower temperature with the obstruction. The only explanation
could be that there was a change in the quality of the air flow, as the
restriction could only have reduced the bulk air flow. That change
in quality was increased turbulence due to ALL the air having to make
a 90 degree turn either before entering the case (at the front) or just
after entering the case (from the sides). It is important to notice that
even with REDUCED AIR FLOW, the observed effect was
REDUCED TEMPTERATURE.

*TimDaniels*

 

[Timothy Daniels]

"kornball" has a Senior Moment:

It seems we all had our holes installed backwards, after
flipping them over the systems don't even need fans or
heatsinks anymore. How can we thank you enough?

You are 100% right, a Font Of Supercooling.

Mind if we call you "100% FOS" for short?

The purpose of promoting turbulence is to make
heatsinks and fans more efficient. It has nothing
to do with the elimination of heatsinks or fans.

Furthermore, those components sensitive to heat
but which don't commonly have heatsinks, such as
hard drives and expansion cards, can get better
cooling from the addition of air turbulence.

*TimDaniels*

 

[Timothy Daniels]

"Paul" quoted:

"Blown directly at the top of the pins, forced air
creates turbulent air between the pins, breaking the
air boundary layers around the pins and enhancing the
heat sink's high convectional thermal coefficients."

"Blown" already implies turbulence. "Pins", in the plural,
already implies turbulence as the turbulence induced by
pins upstream helps cool the pins downstream. You have
chosen an example in which heatsink cooling is aided by
turbulence.

There are many other parts in a PC which need cooling,
too, but which don't get blown on by fans and which don't
have a heatsink. Turbulence aids the cooling of those parts,
just as it aids the cooling of heatsinks. In the absence of a
dedicated fan, those parts can benefit from increased cooling
by increased turbulence.

*TimDaniels*

 

[Timothy Daniels]

Prop fans introduce plenty of turbulence anyway

Isn't it interesting, then, that when air intake through the fan was
restricted, the cooling increased.

*TimDaniels*

 

[Al Dykes]

"kornball" has a Senior Moment:



The purpose of promoting turbulence is to make
heatsinks and fans more efficient. It has nothing

Define "efficient" in any situation.

Especially, define "efficient" with respect to a the operation of a
fan.

 

[Timothy Daniels]

"Al Dykes" asked:

Define "efficient" in any situation.

It means more effective with less "something".
In this situation, the cooling by air draft is increase
by adding turbulence to the flowing air. The "something"
that is less is the size and speed of the fan, and by
implication, the amount of volume of air flow per
unit of time.

Especially, define "efficient" with respect to a the operation
of a fan.

Fans cost money. Bigger fans generally cost bigger money.
Having to make a bigger hole in the case for a bigger fan
costs money and time. If you can get adequate cooling with
a small fan without having to install a bigger fan by increasing
the air turbulence, you've found "efficiency". If increased air
turbulence can get you adequate cooling without having to
put another hole and/or another fan in the case, you've found
"efficiency". If you can get adequate cooling by switching
from smooth heatsinks with parallel fins to heatsinks with
staggered "fingers" or "pins" that induce turbulence, you've
found "efficiency".

*TimDaniels*

 

[Al Dykes]

"Al Dykes" asked:


It means more effective with less "something".

In this situation, the cooling by air draft is increase
by adding turbulence to the flowing air. The "something"
that is less is the size and speed of the fan, and by
implication, the amount of volume of air flow per
unit of time.




Fans cost money. Bigger fans generally cost bigger money.
Having to make a bigger hole in the case for a bigger fan
costs money and time. If you can get adequate cooling with
a small fan without having to install a bigger fan by increasing
the air turbulence, you've found "efficiency". If increased air
turbulence can get you adequate cooling without having to
put another hole and/or another fan in the case, you've found
"efficiency". If you can get adequate cooling by switching
from smooth heatsinks with parallel fins to heatsinks with
staggered "fingers" or "pins" that induce turbulence, you've
found "efficiency".

*TimDaniels*

Clueless.

 

[kony]

Fans cost money. Bigger fans generally cost bigger money.
Having to make a bigger hole in the case for a bigger fan
costs money and time. If you can get adequate cooling with
a small fan without having to install a bigger fan by increasing
the air turbulence, you've found "efficiency". If increased air
turbulence can get you adequate cooling without having to
put another hole and/or another fan in the case, you've found
"efficiency". If you can get adequate cooling by switching
from smooth heatsinks with parallel fins to heatsinks with
staggered "fingers" or "pins" that induce turbulence, you've
found "efficiency".

BS again Tim.

Trying to design in turbulence reduces airflow requiring
MORE fans, MORE cutting openings, MORE time and cost to get
the same temps. Oh, more noise too.

Who do you think you are fooling?
Anyone can spout a random theory if they ignore the other
variables. No amount of recanting your theory changes the
fact that you are still not considering the other variables.

You MUST, necessarily, have a working reproducible model to
test. Anything less is just Tim trying to play know-it-all
inside his head. The more common term is "BS artist".