power supply overheating?

[Ed_]

Hi,

This question began with my attempts to find out why the CPU fan in my new
homebuilt PC was rotating so fast (5000+ RPM) and, as a result, causing so
much noise. I have one of the newish Intel P4 3E (Prescott/90nm)
processors - the core temperature of which seems to vary between 60 and 65
Celsius (with the Intel supplied heatsink/fan). Initially assuming that the
reason for the high fan speed was the high CPU temperature, I emailed Intel
product support who claimed that the cause was probably a failure to
maintain the internal chassis temperature at or below 38 Celcius (and that
this *chassis temperature* was what the CPU fan was responding to).

As part of this ongoing investigation, I've noticed that the case of my
power supply (an Antec TruePower 430) gets very warm - between 43 and 50
Celsius (depending on load). I can imagine that a large block of metal
emanating heat at 50 degress might make it difficult to maintain an internal
temperature of 38 Celsius! But... does this indicate a fault with the PSU?
What should the approxiate temperature be? (For instance, this review
http://www.anandtech.com/showdoc.html?i=1774&p=8 at AnandTech measures the
temperature *inside* the unit after 30 minutes at 33.7C)

As a comparison, I have another computer with a lower spec Antec TruePower
380 PSU. The temperature of its case holds rock steady at about 40C
irrespective of usage or apparent load - this is in contrast to mine which
fluctuates without really stabilising (at about 10 degrees higher).

It seems to me that I am rather stuck in a cycle between various
manufacturers/vendors claiming that there equipment is not to blame and I
definitely need some independent advice. The actual cooling system is two
120mm Antec case fans (one intake, one exhaust) in an Antec Performance One
P160 aluminium chassis. There are also the usual number of other fans in the
system - on the graphics card (ATI Radeon 9800 Pro) and North Bridge
(Gigabyte GA-8KNXP Rev 2.0 motherboard) etc.

-dan

Are your 120 case fans connected to the "fan only" connectors on the TruePower?
If so this may be causing your fans to run slower. I use the same PS on a
SuperLanboy and switched the fans to the regular non-thermally controlled
connections.

Ed

 

[PWY]

Dang it, I woulda bet ten bucks.

Have you actually felt the air being blown in at the bottom of the case ?
There may be something under the front cover blocking the inlet.

PWY

 

[Dorothy Bradbury]

Some notes...

Filters, even large ones are not suitable for PC cases
o Typical "fan filters" have a resistance of 15-25Pa when clean
---- when dirty this figure can rise to 45-60Pa
o Even a 120mm fan at 12V, shifting 55cfm (typical) is not enough
---- you will lose 50% of your airflow when clean, 75-90% when dirty
o Clean regarding small "fan filters" is a *very* short period of time

Put simply, PC fans are typical Axial fans - not Radial/Blower fans.
o Axial fans produce typically 0.1-0.25" of H2O pressure at 0cfm
---- at *0cfm* they will depress a column of water 0.1-0.25"
o Radial/Blower fans will typically produce 1-2.5" of H2O pressure at 0cfm
---- an order of magnitude larger, and overcoming filter resistance

The only way to make a PC work with filters to attack on 3 fronts:
o Extremely larger filter
---- this is unlikely to be practical on a typical PC case
o Change it often & use the lowest filtering grade
---- typically it's the white polyester media, designed to stop fan blade dusting
o High pressure fans
---- this means large fans of large DEPTH (38mm deep) & high airflow
---- this also necessitates an increase in noise generally
---- admittedly you could try a 24V Comair Rotron 172x55mm at 12V

So a few simple things to check:
o Verify the 120mm fans are running on 12V
---- as someone points out, they may be on the PSU voltage-varied fan feed
o Remove the filter
---- then verify the temperatures from the existing known baseline
o Enlarge the front intake port
---- most of the resistance in a PC comprises 2 areas
---- 1) the punched case grill -- a mere 45-55% free air
---- 2) the intake plastic cut-out -- many PCs have a tiny 3x1" cutout

Regarding 2) above...
o One single miserly 80mm exhaust fan is around 4,800mm^2
o Most PC cut-outs are barely 1,500mm^2

Excellent that you have an exhaust fan directly behind the CPU cooler,
that will reduce the tendency of the CPU cooler to recirculate its own heat.
Most coolers recirculate 40-50% of their air = more rpm/noise to compensate,
and particularly since they can't compensate, more elevated temperatures.

However, all this fails UNLESS the PC can get cool air in fast enough.

At the lowest level, 300cfm will cool 1500W of heat in electronic apps.
That figure is a EBM-Papst rule-of-thumb which must then be adjusted for
enclosure static resistance - and so *realised* cfm in your application matters.

Serial mounted fans (exhaust 120mm, intake 120mm) will not elevate the
cfm beyond that of a single fan - but it will assist the static pressure created,
and so increase the cfm realised within your enclosure. That is still limited by
the case inlet c/sectional area, and the intake & exhaust fan fingerguard grills.

Considering you have about 75cfm of exhaust air, enough to cool around 400W
of internal heat, you should see better temperatures than you are seeing. Thus in
reality your 75cfm of *spec'd* exhaust air is reduced to enclosure resistance, or
potentially fan speed control by the PSU (often overlooked).

Yes the Prescott kicks out a lot of heat, but a 120mm fan stuck behind a CPU
is going to yawn at it - IF you can get the cool air into the enclosure. You may
also want to verify you have no air inlet short-circuits - ie, the exhaust fan can
exhaust air not from the CPU area, but from other empty case fan holes above
it or omitted drive-bay blanking plates or such like.

Clearly there is an airflow static resistance problem since you get quite marked
drop in temperature with a case side removed - which will essentially *stop*
airflow to the 120mm exhaust fan being pulled over the components.

 

[w_tom]

Everyone who offered Dan advise should have no problem with
what Dorothy has just posted. This is how a repair person is
suppose to think. Facts supported by numbers. There is
nothing posted by Dorothy that is too complicated for any high
school science graduate. IOW if one repairs computers and
does not understand what she has posted, then one has no
business advising others on fan and heat solutions. Again,
Dorothy did not post anything complicated. She demonstrated
how easily computers can be cooled even with one fan.

IOW when someone foolishly recommends four or six fans for
cooling, well Dorothy's numbers demonstrate why the word
foolish applies.

 

[Timothy Daniels]

Everyone who offered Dan advise should have no problem with
what Dorothy has just posted....
Dorothy did not post anything complicated. She demonstrated
how easily computers can be cooled even with one fan.

Dorothy's quite well informed, but I wonder about her statement:

"Serial mounted fans (exhaust 120mm, intake 120mm) will
not elevate the cfm beyond that of a single fan - but it will
assist the static pressure created, and so increase the cfm
realised within your enclosure."

That sounds self-contradictory. Serially mounted fans won't increase
the CFM beyond that of a single fan - but they will? Hmmmm....

*TimDaniels*

 

[kony]

IOW when someone foolishly recommends four or six fans for
cooling, well Dorothy's numbers demonstrate why the word
foolish applies.

Not necessarily true.

If considering her post, then the part about filtering alone is evidence
that given an enclosure that cannot accept one or more monster-sized fans,
it may become a necessity to have 3 or more fans, and with a modern
high-end system, potentially MUCH more than 3. Although more and more
cases accept 120mm fans these days, the majority do not, typically 80mm or
a few more fortunate to have 92mm.

On the other hand, a differt very good reason to have multiple fans is the
noise level per same airflow rate. Very low RPM fans can be nearly if not
completely inaudible, but it'll obviously require more fans to move same
amount of air. This arrangement may be the most optimal since the
multiple fans build in a partial redundancy, that if a system were
engineered beyond environmetal and component cooling needs, it would likey
continue to function until failed fan can be replaced, if one were to
fail, which is also going ot happen at longer intervals due to the lower
RPM.

 

[w_tom]

Dorothy has posted accurately. Due to some airflow
resistance, the CFM of one fan may be trivially decreased.
Better fan manufacturers even have charts for this. But flow
through each fan still remains about same CFM. One fan or one
million fans in series makes no difference. They all still
move the same CFM. That CFM only slightly decreased by
airflow resistance.

Having said this, if airflow resistance is created by a
closed box (infinite resistance), then yes a 30 CFM fan is now
outputting 0 CFM. But the primary reasons for putting an
incoming and outgoing fan in series 1) is backup so that if
one fan fails, the other still maintains airflow, and 2)
creates turbulence so that interior hot spots (ie created by a
ribbon cable) are less likely to exist. IOW Dorothy has
posted accurately.

Do two fans in series double the airflow? Double the CFM?
Of course not. Obviously not. And yet some would even 'feel'
that is so.

 

[w_tom]

If intent is to install six fans of very low noise and
therefore low CFM; then yes, that is a good reason for many
fans. But that is not the cited example of 'more fans'.
Again, solution is about first doing numbers which is what
Dorothy demonstrated.

In my example, more fans for lower noise is not the case.
The recommendation for many fans all of standard CFM is too
often recommended because they 'feel' that one fan is not
enough. They have the classic Home Improvement solution -
"More Power" - as a solution to CPU problems. The point
remains, fans should be selected for reasons as Dorothy has
provided. Notice the character of her post. Extensive use of
numbers. Those who recommend more fans as a solution to CPU
crashes typically do so without citing any numbers. They
don't need no fancy numbers. They just know. They never
learned any numbers. They just know. Experience without
numbers makes them their own worst enemy.

That being the point. Many will recommend more fans only
because a CPU crashes or because they think a CPU is hot.
They don't do as Dorothy has just demonstrated - first provide
and study the numbers.

 

[Creeping Stone]

Do two fans in series double the airflow? Double the CFM?
Of course not. Obviously not. And yet some would even 'feel'
that is so.

I have put two fans on top of each other, during practical experimentation
and especialy when both are run at lower voltage, they CAN move
significantly more air than either does alone. It all depends on the
details of the implementation, as does the accuracy of the theoretical
formula some are taught to rely on too eagerly.

Theorise how you will, but testing and observation routinely deviate from
the normalised math.

 

[David Besack]

Dorothy has posted accurately. Due to some airflow

resistance, the CFM of one fan may be trivially decreased.
Better fan manufacturers even have charts for this. But flow
through each fan still remains about same CFM. One fan or one
million fans in series makes no difference. They all still
move the same CFM. That CFM only slightly decreased by
airflow resistance.

Only slightly? I hope so. Right now I've got a 2-fan system (both
120mm). The intake fan has an aluminum mesh dust filter. I opted for
the aluminum fitler in hopes that the flow resistance would be
negligabele. Your post gives me a little more confidence in that area.

Of course, even if the 2-fan system is very even in terms of
CFM-in/CFM-out, I sitll have a 2-fan PSU exhausting air, and an exhaust
fan located beneath my GPU. Which is why I'm planning on adding a 3rd
80mm fan... somewhere. Hopefully I will be supplying more than enough
air into the case, and let the exhausts do whatever they can.

Having said this, if airflow resistance is created by a
closed box (infinite resistance), then yes a 30 CFM fan is now
outputting 0 CFM. But the primary reasons for putting an
incoming and outgoing fan in series 1) is backup so that if
one fan fails, the other still maintains airflow, and 2)
creates turbulence so that interior hot spots (ie created by a
ribbon cable) are less likely to exist. IOW Dorothy has
posted accurately.

Interesting. I would think that one intake and one exhaust fan, located
on either side of the case, would be MORE likely to create hotspots,
because they would create a tunnel effect. As in, the air flows along a
specific path and right out of the case again. What would create
turbulence is 1) internal fans moving air "sideways" or 2) objects in
the direct path of the airflow, forcing it to move around the case more.

Of course, this is all conjecture until I tie little ribbons to the
inside of my case and "see" my airflow

 

[Nomad]

Just a couple of things regarding cooling I thought I would throw in.
In my case I found that going for the round drive data cables rather
than the flat ones helped. I also threw in one of those PCI slot fans
that are designed similar to a furnace fan. The PCI slot fan is not
thermally controlled but it is quiet enough for my purposes. I am also
running a dual fan power supply, the Intel supplied CPU fan on my P4
3.0, a fan on my sound card and one case fan. Current temps after
running three hours with five apps going, proc. zone 34C, zone 1 30C,
zone 2 31C. All of this is with 2GB of DDR 400 PC3200.
The meek shall NOT inherit the earth.

 

[Dorothy Bradbury]

Many fans can help in some applications:
o Pursuit of silence over single fan solutions
---- 1-fan-PC -- ATX PSU fan is typically 40cfm & 38-45dB(A)
---- 2-fan-PC -- ATX PSU fan of 25cfm & exhaust fan of 25cfm = 50cfm & 23dB(A)
o Pursuit of redundancy thro serial fans
---- if a server uses 2 fans, serial redundancy doubles that number

Redundancy thro serial fans can increase the risk of downtime:
o Directly -- more fans = more probability of failure in a given period
o Indirectly -- major cause of outage in a data-centre is PC maintenance
---- that is not just to your PC, but to surrounding PCs unknowingly to them

So serial redundancy is best viewed as boosting availability & allowing you to
choose when someone services the machine re downtime - versus the PC choosing.

Serial fans never boost cfm?
o In some instances they can do - but it's best to assume not
o It comes down to blade foil design, fan separation & post-foil-profiles
---- intake - blades expect largely axial & laminar airflow into their blades
---- exhaust - blades eject largely vortex airflow - definately not axial

So when you put two fans in series, the intake airflow into the 2nd fan is not axial
& laminar, but vortex non-axial - so the 2nd-fan intake foil design is wrong, and as
a result the blade experiences more stall in aerodynamic turns. So you can get 1)
odd noises 2) a drop in total cfm or at best an improvement from more pressure.

Thus for serial fan use it comes down to testing & local enclosure design decisions.
Some fans like Delta use a post-foil airflow straightener - particularly their very high
airflow 80mm & 120mm designs, where airflow noise is less of a problem since the
fan itself is making 40-56-69dB(A) in an industrial application setting. You can also
need an enclosure around serial fans for the design to work even where it can do.

So it's suck-n-see, test before-n-after for serial fan usage.

Fewer fans can always work?
o We'll find out soon with BTX - which proposes a few fan solution
---- main driver is as always cost, as well as noise in relation to that cost
---- costs nothing to make a fan go faster, costs a lot of margin to fit more fans
o Thermal outputs envisaged with BTX present a problem
---- CPU -- 89-103W now, potentially 125-150W as the eventual limit
---- CPU-VRM -- often ignored, but no longer - that itself needs cooling
---- RAM-VRM -- socket limit has partly been due to dissipation here
---- Graphics -- the bug-bear of BTX quite possibly
o BTX poses a 1-fan for CPU/VRM/Graphics solution
---- blow-thro design a la Apple G5, but with 1 fan
---- larger c/sectional fan from the present screaming 60mm CPU fan coolers
o Unfortunately LT this isn't a solution, perhaps even ST
---- BTX poses a largest case size of 3.98", so a 92mm fan is possible
---- even so, graphics cards could exceed 100W, CPU 125W, RAM 30W
---- add into that the VRMs totalling around 15W
---- plus capacitors of 105oC routinely running close to it, Mil is 125oC & $$$
---- bear in mind even an Extended life Panasonic 125oC is just 3000hrs at 125oC
---- total -- 250W of co-located in-line heat sources to cool

The reality is even with BTX, multiple fans are still likely to be used:
o Usual case exhaust fan - twin 80mm or whatever
---- BTX 3.0" height case is nauseating as it projects 60mm or 70mm fans
---- those are horrible sizes for either good flow or low noise levels really
o Clone PC makers aren't giving up tower designs any time soon
---- particularly as hi-end graphics cards DO run well beyond 75W even now

With HDs set to drop to 2.5", that sounds like a solution to that area, but in reality
3.5" 15,300rpm drives run hot now - and they already use 2.5" platters inside re rpm.

So the reality is - it will come down to your internal component choice (heat output),
and the choices made by the case designer (airflow design, intake c/sectional area).

Main restriction on PCs to airflow is the fan grill c/sectional area and the intake
port,
I'm mindful that even now we still have PCs with quite tiny cut-outs in plastic fascia.

Filters are something you really can't do on a PC unless you redesign the enclosure.
They were good in days of 15-50W internal dissipation re 486, 20MB-HD, 8MB.
The military tend to use indirect cooling or coolplate cooling - the former is where
you have an internal sealed but fan-circulated air-path thro a heat exchanger, so no
external contaminants can affect your internal components re water/dust/bio etc.

 

[Timothy Daniels]

"w_tom" blurted out loud:

...Due to some airflow resistance, the CFM of one fan may
be trivially decreased. Better fan manufacturers even have
charts for this. But flow through each fan still remains about
same CFM. One fan or one million fans in series makes no
difference. They all still move the same CFM. That CFM
only slightly decreased by airflow resistance....

Numerous people have posted that airflow increased
when they removed or cut away the grill or the perforated
blade guard that covered their fans. Doing so obviously
reduced the resistance to airflow and thus the back pressur,
allowing more air to flow. Adding a fan in series *also*
reduces the back pressure and results in an increase of
air flow. The effect is very analogous to reducing a
resistance in an electrical circuit while keeping the voltage
across the resistance the same - the current increases as a
result, just as airflow increases with the fan. And electrically,
one can do what two fans in series do by applying twice the
voltage across the resistance - which doubles the current.
If fans were ideal fixed pressure differential devices, putting
two in series would do the same thing - it would double the
air flow.

Do two fans in series double the airflow? Double the CFM?
Of course not. Obviously not.

Obviously not. And no one claimed that two fan in series
*would* double the airflow over that of one fan. And
that is because a fan is a combination of a fixed CFM device
and a fixed pressure differential device. In other words,
it's a real world device, not an ideal device, and two fans
in series aid each other whether they are in series or in
parallel, but not to the extent of doubling airflow. What
they *do* do in increasing air velocity is to reduce the angle
of incidence of the air as it meets the leading edge of the
fan blades, reduce the turbulence produced by the blades
as they pass through the air, thus reducing the noise. Would
the amount of noise be cut in half? Of course not. But I
would expect a reduction of maybe a couple dB.

*TimDaniels*

 

[w_tom]

That aluminum screen is a major obstruction. Look at the
cross sectional inlet area that fan needs. Now you have put a
screen that has cut that cross sectional area by maybe as much
as one half. So now you have done the simplistic equivalent
of half a closed ended box. Therefore you have reduced CFM.
Obstructing the fan was beyond scope of this discussion.

Why are you installing the aluminum screen? When the
standard computer user has dust problems, it is usually from
moving too much air through the case (or the room is a
construction site). Solve dust problems by only moving
sufficient air - not using so many fans.

Only slight resistance in airflow means the incoming hole and
outgoing hole are same as the cross sectional area demanded by
fan. Usually this is understood by default.

In the meantime, why are you adding so much more airflow?
If I remember the ballpark number, the second fan only adds,
at best, maybe 5 degree decrease in case temperature.
Irrelevant. If that temperature decrease is important, then
you have a serious design problem elsewhere in the machine.

Where is this hot spot creating so much worry? Without
understanding what is and is not a significant hot spot
problem, then a proper airflow design cannot be developed.

 

[w_tom]

Of course removing obstructions increases fan output to
rated CFM. See the other post about cross sectional area for
details.

Have you assumed that a fan with 50% obstruction is the fan
being posting about? Obviously not. A fan 50% obstructed is
installed wrong - which means all data is now distorted just
like a fan trying to blow into a sealed box. Will that 120
VAC power supply work properly with half the incoming AC
current obstructed? Of course not.

Airflow is the equivalent of a current source - but not an
ideal current source. Airflow is equivalent to current - not
voltage. An ideal fan will maintain CFM just like an ideal
current source. A 'less than idea' fan is more like a real
world current source. It mostly maintains airflow (current)
until obstructions (resistance) become too great. Then CFM
(or current) falls off appreciably. You are unfortunately
trying to make the fan equivalent to a voltage source.

Two fans will double fan CFM if the fans are excessively
obstructed. Do we therefore install two fans in series to
double the airflow? Of course not. We eliminate the
obstruction as the original poster demonstrated with his
cardboard side panel and larger outlet hole. But the
excessively obstructed fan was beyond the scope of numbers
provided for fans and beyond the scope of what Dorothy
posted. Excessive and unacceptable obstruction invalidates
all numbers, data, and concepts. Apparently this was not
understood by some posters here.

 

[Timothy Daniels]

In the meantime, why are you adding so much more airflow?
If I remember the ballpark number, the second fan only adds,
at best, maybe 5 degree decrease in case temperature.
Irrelevant.

No, relevant. A rule of thumb in chemistry is that a reaction
doubles in speed for every 10 degrees F. increase in temperature.
Since many of the thermal mechanisms at work in chemistry
are also those for disk drive wear, I'd say a 5 degree drop in
temperature is worthwhile.

*TimDaniels*

 

[Timothy Daniels]

Of course removing obstructions increases fan output to
rated CFM. See the other post about cross sectional area for
details.

Have you assumed that a fan with 50% obstruction is the fan
being posting about? Obviously not. A fan 50% obstructed is
installed wrong...

You're waffling. And you're setting up a straw man, again.

First, the grills and perforated screens covering fans do not
cover 50% of the cross-sectional area and they are not
"installed wrong" as every manufacturer does that. So arguing
against a 50% obstruction - which no one has mentioned - is
a straw man.

Second, you've said that CFM remains constant regardless of
decrease or increase of resistance to air flow. Now you're
backing off and waffling. Just what IS your position?

Airflow is the equivalent of a current source - but not an
ideal current source. Airflow is equivalent to current - not
voltage. An ideal fan will maintain CFM just like an ideal
current source. A 'less than idea' fan is more like a real
world current source. It mostly maintains airflow (current)
until obstructions (resistance) become too great. Then CFM
(or current) falls off appreciably. You are unfortunately
trying to make the fan equivalent to a voltage source.

And just where did I equivacate a fan to a voltage source
or to its analogue, a constant pressure differential device?
You're arguing against a straw man, again, as I made no
such claim. In fact, I pointed out that a fan is a combination
of a constant flow device and a constant pressure device,
IOW not an ideal example of either but rather, a real world
device.

Two fans will double fan CFM if the fans are excessively
obstructed.

Waffling, again. Define "excessively".

In truth, two fans in series will increase air flow, but
not double it, two fans in parallel will increase air flow,
but not double it.

*TimDaniels*

 

[Dan Brill]

Hi Matt,

Functionality shouldn't be ugly.

Perhaps not but some of those cases really are

http://www.intel.com/cd/channel/res...t_inst_info/dsk_tested_source_lists/53211.htm

Most of them look fairly hard to get hold of in the UK as well and I can't
even find the Coolermaster model on their own site.

Looks like you could just take your side panel to a glass shop to have
them cut a hole in the plexiglass. And buy a grill and/or an 80mm fan
to put over the hole. Or take it to somebody who does case
modifications. Or, if you are so inclined, practice a little with a
plexiglass cutter (cheap) and some scrap plexiglass and do it yourself.

Well, it is an aluminium side panel, not windowed but, no doubt, there is
some way to get it modified (though it would be easier if Antec sold a
replacement side panel which already achieved this).

Then you could keep the oh-so-attractive P160.

I've been reading the remainder of this thread and considering issues such
as the potential obstuction created by the dust filter etc. Personally,
there is no difference in temperature in my case whether the dust filter is
present or not (which isn't to suggest that there is no difference in the
quantity of air being moved because I'm sure there is). I've also tried
attaching the case fan to a 'normal' power supply outlet rather than the
'special'/variable fan-only ones provided by the PSU. Again, it makes no
difference to the speed of the fan which was already rotating at around its
maximum 1700RPM anyway.

My current thinking is still to buy an Antec Plus1080-AMG, which is on
Intel's "Thermally Advantaged Tested Chassis List" - at least then neither
company's technical support can use "you aren't using a thermally advantaged
chassis" as a reason to ignore my questions or generally obfuscate the whole
situation. I've got another use for the P160 to house an older machine.

I'm still not sure about the Plus1080-AMG as a case though. I'd feel better
if instead of being on Intel's list or advertising itself as 'thermally
advantaged', it actually conformed to Intel's Chassis Design Guide (version
1.0 or 1.1) which recommends a Chassis Air Guide assembly. At least it has a
side-vent which, I believe, has a good chance of curing the problem.

-dan

 

[w_tom]

Now lets put example numbers to your speculations from
chemistry. If the CPU temperature increases, then the CPU
might only last 100 years instead of 200 years. That 5
degrees is totally irrelevant except when other unresolved
hardware problems make it relevant. Again, I cite the Dorothy
post as example. Use numbers such as from datasheets before
making conclusions. Otherwise we get this junk science
conclusion about chemical reactions inside semiconductors.
Manufacturer provides a temperature limit that the CPU works
at just fine. As long as semiconductor stays within those
limits, then all is well. And yes, don't bother citing
electromigration. The "CPU might last 100 years instead..."
sentence includes that concept as well.

If 5 degrees makes a big difference to a system, then the
system has serious hardware problems elsewhere. If 5 degrees
make a big difference to the human, then human needs to first
learn basic semiconductor concepts or read numbers from
manufacturer datasheets. Those 5 degrees created by doubling
the fans is totally irrelevant. But then many also want to
fix computers using the Home Improvement concept of "More
Power". Double the fans (and noise) for an amazing 5 degree
temperature reduction!

 

[w_tom]

Tim - 50% obstruction was provided as an example and not as
what is happening in every case. Clearly all you want to do
is argue. Cut the bullshit with 'waffling' and 'strawmen'.
When you are ready to discuss technical, then it will be
obvious that you first read a manufacturers datasheet. You
are posting numbers using junk science concepts - such as
irrelevant chemical reactions inside a semiconductor.

In the meantime, I never said "CFM remains constant
regardless of decrease or increase of resistance to air
flow." This is the twisted reasoning so common by lawyers and
those who argue subjectively. If I had posted that, then I
would have also said a fan blowing into a closed box would
move 100% of rated CFM. Obviously I posted completely the
opposite - remember? If only you did not read selectively.
Then what I had posted would not have been intentionally
misrepresented. Then 'waffling' and 'strawman' would be
replaced with good technical numbers. Currently your posts,
devoid of numbers, are classic of junk science reasoning
combined with selective reading. Please feel free to cite
manufacturers data sheets and numbers before replying. Use
technical facts rather then a lawyer's subjective terms such
as 'waffling' and 'strawmen'. Demonstrate a technical grasp
before replying. Don't continue posting using 'junk science'
reasoning.