Original ATX specs called for power supply to blow air into the
chassis. At least one of your citations says same. Don't know which
ATX revisions changed the defacto standard. It was never a mistake
even though 'a mistake' is the popular myth promoted among and by
computer assemblers.
One of your citations even bluntly agrees with what I have posted -
but only in subjective terms:
In general, exhausting air from the system chassis enclosure
via a power supply fan at the rear panel is the preferred, most
common, and most widely applicable system-level airflow
solution. However, some system/chassis designers may
choose to use other configurations to meet specific system
cooling requirements.
The point: chassis cooling is defined by intake air temperature and
CFMs. Direction is not relevant.
My posts come from learning the engineering, doing the design
numbers, and making things work. I did not learn from above summary
citations written only for technicians and layman. None of those
citations are sufficient for engineeing knowledge. For example, where
are the equations? No equations because those citations are the
equivalent of an executive summary - for a boss so that he can 'feel'
he knows.. Damning in every citation are no crtical numbers such as
CFMs. Instead, discussion is subjective - what English or
communication majors do to somehow know a fact. 'Subjective
reasoning' is why some make assumptions and then declare an assumption
as fact. None of those citations was sufficient to understand thermal
engineering.
Most embarrasing is the third citation that says:
When a power supply fails, it often sends random jolts
through the rest of the system, killing (and sometimes
burning) your valuable computer components.
If the author had basic electrical knowledge, then he would have
never written that. He has turned popular myth into a technical
fact. Shame on him. Anyone with basic electrical knowledge knows
that a properly designed power supply must never send "random jolts
through the rest of the system". A fact so universal that everything
written by that author is questionable. Ken - my point again. Some of
those citations are so subjective as to even post myth as fact.
How did we know Bush was lying about Saddam's WMDs? The numbers
from engineering source kept contradicting what the president
claimed. But how many ignored the numbers - instead believed
subjective reasoning? An example of what we are supposed to have
learned from history.
I would be embarrassed to have learned technology from these
subjective and sometimes technically inaccurate citations. But then I
also recommended learning from application notes from fan
manufacturers. IOW the numbers; not subjective conclusons.
Ken posted:
There was no mention of any other venting of the PSU enclosure.
But that is not relevant since (if I recall) he (the OP) was
operating with an open chassis.
Point three was :
Third, a computer can boot and the power supply is defective.
Where was anything misleading. One typical failure is for a power
supply to be defective, boot the computer OK, and have the computer
become intermittent or suddenly stops working months later. What is
misleading about that reality? The point: misleading is to assume a
power supply is OK only because the computer boots. A booting computer
says little about the integrity of a power supply or the rest of that
power supply 'system'.
Point four was about what he felt. The OP called 130 degrees too
hot. Once we got numbers from him, then we discover it was only
warm. Too hot means semiconductors are so hot as to leave skin.
This contradicts popular myths where a trivial 130 degrees F get
assumed to be hot. larry moe 'n curly also defined 50 degrees C in
proper perspective:
Did it cool down, or were you exaggerating in your original
post? Because 50C isn't that hot.
He is right. That power supply box temperature was not hot. Too
hot is too often when one assumes rather than first learn the
perspective of numbers. Again a reference to subjective knowledge
verses knowledge that is accurate and definitive.
Point five is about removing the second fan that provides near zero
advantages (air filters are not in anyway relevant or mentioned).
You should discover only trivial chassis temperature difference
assuming the power supply fan is working properly.
That second cooling fan does near zero temperture improvement which
would be obvious if your citations provided numbers and equations.
The second fan provides trivial temperature improvement. Third and
more fans provide exponentially less improvement. Again, perspective
from numbers not found (as best I saw) in any of your citations. This
paragraph contradicts popular myth because conclusions are based both
in engineering experience and in the numbers from equations and
manufacturer specs.
This author learned from engineering; not from subjective sources.
Removing that second fan should result in a trivial temperature
increase. Of course, if temperature increase is significantly large,
then we have an engineering fact. A technical problem that must be
discovered and corrected.
Notice the difference between engineering and subjective reasoning.
In engineering we know what the numbers should be theoretically AND
must obtain same numbers experimentally. Without both facts, then
only speculation or subjective reasoning exists. And again, I
complain about those citations that provide no numbers.
Why do some use filters? Too many fans create dust problems.
Instead of fixing the problem - excessive airflow - some want to cure
symptom with filters.
Meanwhile I cannot see if that fan is blowing in or out. I asked
that question repeatedly in questions that define pro and con of both
directions.
Why do you assume because a fan mount exists, then a fan is
required? That is not reasoning? That is akin to wild speculation.
They are selling computer chassis to computer assemblers. I too would
put a fan mount there even though I know it has no purpose. Put it
there only because it would sell to the naive who subjectively know
chassis needs more than one fan. Just because a fan mount exists does
not mean the fan mount is required. In fact, to increase sales to
these people, I would put fan mounts all over the chassis and sell
them 20 fans. They would know only from subjective sources? Good.
They would know its a better computer because they spent more money on
my fans? Good. Atfer all, more fans always means more cooling -
even though the numbers say otherwise? Again, conclusions when
subjective reasoning has no numbers and equations.
If they want to think subjectively, then I want to reap more
profits.
Is subjective reasoning sufficient for me? Obvioius?. Subjective
sources are sometimes a symptom of facts based in wild speculation or
junk science. Again, my post comes from engineering training - not
from subjective publications.
I can appreciate why you may have assumed conclusions in those five
points. But the points themselves are accurate. Assumptions taken
from those points (due to missing technical knowledge) would explain
your difficulty with them.
But again, I would be embarrassed to have learned technology from
those subjective citations. I would give them to the boss so that he
'feels' he understands - and goes away.
While that is totally irrelevant to advising anyone working
with reasonably current systems, you don't specifically identify
the "ATX spec" you are referring to. There are a great many.
There are a number of specs for special environments that
allow for implementing a cooling scheme that could use the
PSU fan as you describe, but only the very original ATX
specification ever required such a thing, and that was quickly
found to be a mistake and corrected.
http://www.ibm.com/developerworks/power/library/pa-spec9.html
http://www.formfactors.org/developer/specs/PSU_DG_rev_1_1.pdf
http://www.pcstats.com/articleview.cfm?articleID=1720
It may be sites such as this that gave you that idea, but note that
their references on the subject are the first two I posted above.
http://users.erols.com/chare/atx.htm
http://www.pcguide.com/ref/power/sup/form_ATX.htm
That was my position as well, but does not appear in the
last post of yours, which was the one I quoted and was
replying to.
...
It may be that English is not your first language, so I only
point this out to assist you, "Political Correctness" has no
bearing to discussions of inanimate objects (except Stoned
Slackers, perhaps).
Don't worry, I am well aware of the factors that effect
component cooling, read my other posts in this thread.
The problem the OP stated involved a "Hot" PSU,
which he described as having two fans; both of which
were blowing air out of the PSU enclosure. There was
no mention of any other venting of the PSU enclosure.
There would be no need for any additional openings in
the enclosure (impacting on its RF shielding), if the fans
were following the current standards and not competing
with each other. Taking all that into account, it makes
more since that the PSU was made to the current spec,
but that fan on the bottom of the PSU is blowing in the
wrong direction.
As to your third point in your last post, the first sentence
is logical enough, but the rest I disagree with and view as
potentially misleading.
Your Forth point is misleading as well. The OP was
talking about the temp he felt on the PSU enclosure, not
feeling an overheated IC. I have no idea how hot the ICs
would have to be to make a PSU enclosure, with two fans,
reach the criteria you described; "second or third degree
burns", but I would expect to smell smoke long before that.
Telling people that a hot PSU enclosure would be providing
"perectly ideal temperatures" unless touching it produces
burns, is reckless, at least.
The fifth point is probably the most counter-productive
(and in part counter to basic physics). Most cases are
designed to incorporate the airflow I described in my other
post. Some even put filters over the intake vents at the
bottom front of the case, is it your position that the filters
are there to filter the outgoing air? Do you think it odd or
just a quirk that most case fan mounting possitions are just
below the PSU on the back wall of the PC case? Can't
you see that if the 120mm fan is blowing out it will be pulling
the air in from any open vents in the PC case, but mainly
from the intake vents designed into the case? Why do you
think any fan placement to cool the hdd; has the fans blowing
the cool intake air between the drives and into the rest of
the PC case? (Your description of the "proper" airflow
would require that they be blowing the air from inside the
PC case across the hdd and out the vent.)
[Small sidebar: Fans have a high pressure side and a low
pressure side. To push air through something you want to
use the high pressure side, the side the air comes out of the
fan. The thing you are trying to push through needs to be
very close or even coupled to the fan, as unless constrained,
the pressure dissipates rapidly. You can use the low pressure
side to pull air from a greater distance than you can push air
with the high pressure side. The speed/acceleration of the air
being pulled is effected by the path it travels as well as the
negative pressure/suction provided by the fan. So, you need
to put your fans close to, or connected to things with restrictive
spaces like heatsinks, or hdd racks, that you want to cool. For
creating airflow through the open space of a box, you can use
the low pressure side and have the fan placed away from the
components being cooled by the air flow.]
Does that provide sufficient reason, for you?
