CFM calculator to get rid of HEAT (TDP or WATTS)
Skybuck Flying
I'm curious if it's possible to build a powerfull PC/iCore/GPU/Cuda PC
without getting into heat problems ! ;)
For example CPU has about 30 TDP (which is close to watts, otherwise find
watts listing) and GPU could have as much as 200 to 300 TDP (watts) (top
model for maximum bandwidth).
Additional watts for memory( 20 watt ? just guessing ;)) /motherboard (10
watt ?)/4 harddisk (60 watt*4=240)
Let's say 800 watts.
How much CFM would be needed to get rid of all that heat ?!?
A CFM heat (PC) calculator would be most handy... maybe somebody can
make/program it.
What would be good formula's to use ?!? ;)
Bye,
Skybuck =D
Skybuck Flying
Wikipedia shows conversion chart:
http://en.wikipedia.org/wiki/Conversion_of_units#Power_or_heat_flow_rate
1 atmosphere-cubic foot per minute = 47.820 074 682 24 W
Which is probably close to 48 Watts.
But this is atm-cfm and not cfm hmm... not sure if that is same thing... but
for now I'll go with it ;)
So (800 * 60) / 48 = 1000 CFM needed ! ;) =D
Let's see if antec case is sufficient lol:
http://www.antec.com/pdf/manuals/1200_EN%20manual.pdf
It's about 240 CFM...
So there is no way that the antec 1200 case would be able to cool such a
system sufficiently.
Yes perhaps the graphics card would output a little bit of air as well...
but it's waaayyyy too little.
And this manual probably assumes at maximum operating fan speed... so let's
divide 240 by 2 is 120 CFM... and let's divide it one more time because of
dust is only 60 CFM.
Yeah...
Well it's pretty clear where this is going:
OVERHEAT DAMAGE ! ;) =D
Bye,
Skybuck ;)
"Skybuck Flying" <Blood...@hotmail.com> wrote in message
news:e706d$4dd3961c$54192c06$14...@cache4.tilbu1.nb.home.nl...
Skybuck Flying
More interesting stuff:
http://extension.usu.edu/files/publications/factsheet/Ag_poultry_Vent_03.pdf
Bye,
Skybuck.
Skybuck Flying
play a roll as does temperature of air:
http://en.wikipedia.org/wiki/Standard_cubic_feet_per_minute
Bye,
Skybuck.
Skybuck Flying
http://en.wikipedia.org/wiki/Heat_transfer
Getting down to basics now lol... getting down to the core of it ! ;) =D
Bye,
Skybuck.
Skybuck Flying
Skybuck Flying
http://www.aavidthermalloy.com/technical/papers/64bit.shtml
It mentions CFD analysis.
Why calculate one little simple fokking formula if you can simulate the
whole system ! LOL.
What hopefully CFD analysis software ! ;) =D
CFD:
Csomething Fluid Dynamics LOL.
Bye,
Skyflow.
Skybuck Flying
http://www.solidworks.com/sw/products/cfd-flow-analysis-software.htm
And the pirates have it too !
Har Har Har ! ;) =D
Doei,
Skypirate ! ;) =D
Skybuck Flying
adventage, since I am on a PIII 450 mhz ;) and best of all it's free, alas
only for MAC OS and Linux:
Could still be interesting though.
Bye,
Skybuck.
Skybuck Flying
I heared about these motherfokkers too:
http://www.ansys.com/Products/Simulation+Technology/Fluid+Dynamics/ANSYS+CFD
They've been around for a long time... maybe old maybe not old maybe good
maybe not good...
So far it's a bit vague...
Butt the pirates have it too ! ;) =D
Bye,
Skybuttpirate ! LOL.
Skybuck Flying
has the internals... hmmmm.....
http://sketchup.google.com/3dwarehouse/details?mid=a8b7ee219d82949bcce101f6f733eba0
Bye,
Skycad.
Skybuck Flying
Joel Koltner
> 1 atmosphere-cubic foot per minute = 47.820 074 682 24 W
> Which is probably close to 48 Watts.
1+1 is close to 3, for large values of 1.
No offense, but your analysis after this point is completely off the mark; you
can't just willy-nilly ignore units and make conversions the way you're doing
them. You also need to consider the allowable temperature rise, the thermal
resistance of your components, how the airflow is routed (you're not going to
get 1000CFM into every nook and cranny), if it's laminar or turbulent, etc. --
"How many CFM do I need to cool an 800W computer system?" doesn't have enough
information so as to allow for reasonably accurate answers.
That being said, 1000CFM is a *huge* amount of air movement and, routed
proplerly, will definitely be adequate to cool an otherwise well-designed
system.
---Joel
.p.jm.@see_my_sig_for_address.com
Excuse me - 1,000 CFM is enough to heat and cool a 1,500 sq
ft + home. It's what a 2 & 1/2 ton AC unit moves.
Yeh, I think it would be enough for a computer :-).
______________________
Air flow - get BTUH from CFM and T.D. ( rise )
CFM = 1000 , T.D. = 10 ,
Calculated Sensible Heat BTUH = 10,734 ( aka 3,147 watts )
______________________
Air flow - get BTUH from CFM and T.D. ( rise )
CFM = 1000 , T.D. = 20 ,
Calculated Sensible Heat BTUH = 21,467 ( aka 6,295 watts )
Hell of a gaming box you got there, that pulls 3,000 watts and can
literally heat a medium sized house in the middle of winter :-)
--
Click here every day to feed an animal that needs you today !!!
www.theanimalrescuesite.com/
Paul ( pjm @ pobox . com ) - remove spaces to email me
'Some days, it's just not worth chewing through the restraints.'
'With sufficient thrust, pigs fly just fine.'
HVAC/R program for Palm PDA's
Free demo online at www.pmilligan.net/palm/
Free 'People finder' program now at www.pmilligan.net/finder.htm
Whenever a Liberal utters the term "Common Sense approach"....grab your
wallet, your ass, and your guns because the somnabitch is about to do
something damned nasty to all three of them.
Pray for Obama : Psalms 109:8
Robert Myers
> Yeah,
>
> I heared about these motherfokkers too:
>
>
> They've been around for a long time... maybe old maybe not old maybe good
> maybe not good...
>
> So far it's a bit vague...
>
> Butt the pirates have it too ! ;) =D
>
> Bye,
> Skybuttpirate ! LOL.
It's only fair, Skybuck, that you probably don't read most of my
posts, because I don't read most of yours.
If you are trying to advance the state of the art in heat transfer as
applied to microelectronics, my suggestion for any of the codes you
have mentioned is that you find a dumpster for them so as not to waste
your time. Mitch Alsup might not agree; I'm not sure what he's up to.
The fundamental difficulty is that heat transfer such as would occur
in a microprocessor heat sink depends intimately on details of
microscopic turbulent flow that are poorly understood for reasons I
have been ranting about for years. The theoretical problems get worse
as you push to higher Reynolds number, which is where you are headed
if you are trying to push air faster.
If you are ever in a window seat on a modern airliner, look out the
window at the row of vortex generators on the upper surface of the
wing. Those tiny little pieces of metal have a dramatic influence on
the aerodynamics of that big old wing, and that should give you an
idea of how hard this problem is. Predicting heat transfer is harder
than predicting pressure forces, so the problem is even harder than
those teensy little pieces of metal would indicate. The problem is
absolutely fundamental, and there is no way out of it. For an
airliner, the largest scale is the wingspan. The smallest scale of
relevance may not be visible to the naked eye.
There probably are significant opportunities for improving heat
transfer in microelectronics. My suspicion is that most of the
advances are made by bending tin and making measurements, rather than
by running CFD codes. I'm not a tin bender, so I could be wrong.
Conceivably, someone with a deep understanding of fluid mechanics,
heat transfer in solids, and computation could make progress that
would make life a lot less random for the tin benders. If you don't
have those skills, you are probably better off putting your money into
a high-stakes lottery, because that's how much chance you have of
making progress by fumbling around in the dark.
Robert.
.p.jm.@see_my_sig_for_address.com
<rbmye...@gmail.com> wrote:
>On May 18, 8:35 am, "Skybuck Flying" <BloodySh...@hotmail.com> wrote:
>> Yeah,
>>
>> I heared about these motherfokkers too:
>>
>> http://www.ansys.com/Products/Simulation+Technology/Fluid+Dynamics/AN...
>>
>> They've been around for a long time... maybe old maybe not old maybe good
>> maybe not good...
>>
>> So far it's a bit vague...
>>
>> Butt the pirates have it too ! ;) =D
>>
>> Bye,
>> Skybuttpirate ! LOL.
>
>It's only fair, Skybuck, that you probably don't read most of my
>posts, because I don't read most of yours.
>
>If you are trying to advance the state of the art in heat transfer as
>applied to microelectronics, my suggestion for any of the codes you
>have mentioned is that you find a dumpster for them so as not to waste
>your time. Mitch Alsup might not agree; I'm not sure what he's up to.
>
>The fundamental difficulty is that heat transfer such as would occur
>in a microprocessor heat sink depends intimately on details of
>microscopic turbulent flow that are poorly understood for reasons I
>have been ranting about for years. The theoretical problems get worse
>as you push to higher Reynolds number, which is where you are headed
>if you are trying to push air faster.
So you think 1,000 CFM, requiring a 2.5 - 3 HP motor, might be
excessive for a desktop computer ? :-)
Robert Myers
>
> So you think 1,000 CFM, requiring a 2.5 - 3 HP motor, might be
> excessive for a desktop computer ? :-)
>
Heat transfer is not my thing, but you know that if someone is talking
about using out of the box CFD to decide if something is even
plausible, without seeing if anyone has done anything even remotely
similar, and without knowing anything about fluid mechanics, heat
transfer, or CFD, they're off on a wild goose chase.
Yes, I know, I could have figured that out from the poster. Somehow,
I just could not stop myself.
Robert.
Skybuck Flying
tiny little processor ;)
Just to see if the case can transport the heat out of it.
I'm a bit worried that it might not run on my PIII 450 mhz to be usuable...
like too slow GUI... or maybe it will run... anyway that's why I posted
multiple CFD softwares, maybe one of them will work on 450 mhz ;)
Would be nice to be able to verify if my case can actually transport enough
heat away for any new computer design.
However if I would go with a very powerfull graphics card then the PC would
probably require a new supply which is a bit too expensive for my taste at
this point.
I have not tracked down the source of death and nothing has been done so
far... perhaps my "land lore" or whatever it's called... they guy/company
who collects the rent might contact me to see if something can be done about
it... but perhaps not...
Perhaps the appartment and everything else is hard-wire for 2 wires only...
of which one is nul-wire...
The weird thing is the kitchen has 3 wire plugs but perhaps those are
fake... the bathroom seems to have 2 wires to washing machine which is
weird...
Perhaps I just live in a shitty appartment if that's the case well then to
bad... but that will have consequences for my future PC's... I would then be
a bit too scared to buy expensive equipment because it's out-dated so
fast...
So then I would have to go with cheap equipment which kinda sucks as well.
So currently I see no way out of this, thus I am lingering waiting for a
solution ;)
I'm also investigating some software technologies to see if I can program
that... which will influence my decision how I shall repair or replace my
computer ;)
Perhaps I will even construct two computers, one for desktop, and perhaps
one for server, and then a router.
The problem is the old-cpu and old-graphics cards are so hot I don't dare to
re-use them for a server, so buying a super computer was probably a big
mistake... but I wanted to try it ;)
Bye,
Skybuck.
Robert Myers
> I would use it to test the airflow of the case itself, not necessarily the
> tiny little processor ;)
>
> Just to see if the case can transport the heat out of it.
>
into the air, not getting the air in and out of the case fast enough.
As I understand things, the economically-practical limit for TDP for
mass-market PC's is 130 watts, and for IBM products, 200 watts. The
problem is not airflow in the cabinet, but heat-exchanger (heat sink)
design.
The kinds of numbers you are looking for should be done on the back of
an envelope by a competent engineer. If CFD were used at all, it
would be to tweak the placement of inlet and exhaust and to look for
possible problems due to obstruction of the flow by components inside
the case. Don't try to use CFD as a substitute for the back-of-the-
envelope step. That kind of engineering is how NASA and its
contractors manage to mix up English and metric units.
If you want to design a clever new heat exchanger, you either need to
know a lot of fluid mechanics or be ready to do a lot of fiddling.
Robert.
Androcles
"Robert Myers" <rbmye...@gmail.com> wrote in message
news:77aab8a1-007d-4959...@e8g2000vbz.googlegroups.com...
Robert.
==========================================
Heat is proportional to current, current is proportional to
switching rate, semiconductors are semi-insulators, aka resistors.
Making the processor small helps speed but concentrates the
energy into a smaller volume and raises temperature at a hot
spot. The heat sink is there to distribute temperature away
from the hot spot.
Nature solved the problem by multiprocessing, you only want
more speed for multitasking.
What is better, one processor doing all the work or a thousand
processors sharing it?
A manufactory turns out 2 cars a minute by multiprocessing on an
assembly line, if you want 4 cars a minute build two factories.
In computing, gaming is where the speed is needed most, you
have 10,000,000 pixels on a large screen to write from a
3D database. Instead of distributing heat, distribute processors
on the back of the screen, a pyramid of management.
Lowest level - One processor controls 100 pixels.
Next level - one middle management processor controls 100 low.
Third level - one upper management processor controls 100 middle.
Fourth level - one boss controls all upper management.
The design challenge you face is getting the data from the database
onto the screen, not getting the air in and out of the case fast enough.
The design challenge you face is to write a hierarchy of management
software for a multiprocessor organization.
nik Simpson
>
> That being said, 1000CFM is a *huge* amount of air movement and, routed
> proplerly, will definitely be adequate to cool an otherwise
> well-designed system.
>
the thing float in mid-air and save on floorspace ;-)
--
Nik Simpson
Jacob
>Hmmm... could be really simple...
>
>Wikipedia shows conversion chart:
>
>
>http://en.wikipedia.org/wiki/Conversion_of_units#Power_or_heat_flow_rate
>
>1 atmosphere-cubic foot per minute = 47.820 074 682 24 W
>
>Which is probably close to 48 Watts.
>
>But this is atm-cfm and not cfm hmm... not sure if that is same thing... but
>for now I'll go with it ;)
It is not the same thing.
Atmosphere is a unit of pressure. Pressure times volume flow rate is
power, which can be expressed in watts. That's why it is in the
section on units of "Power or heat flow rate."
Skybuck Flying
Wikipedia has a nice article on it... it might be of some interest:
http://en.wikipedia.org/wiki/Heat_sink
Bye,
Skybuck.
Skybuck Flying
"Jacob" <as...@hsu.edu> wrote in message
news:hbbbt6pe92om7bvst...@4ax.com...
Perhaps but there is some pressure going on in the case I bet ! ;)
So I don't trust those "pure" CFM ratings and theory ! ;) =D
Also you guys gotta compensate for intel which lies about TDP !
Add at least 20 to 30% extra ! ;) =D
It's to late now anyway... I ordered 1 motherboard and 1 graphics card...
all cheap stuff... 140 euro's... wouldn't call that cheap but ok.
asrock's 939 + gt 520.
Though maybe I will try CFM in the future once my dream pc back online ! ;)
=D LOL.
Bye,
Skybuck =D
Jacob
>Perhaps but there is some pressure going on in the case I bet ! ;)
What does that mean: "pressure going on in the case?"
Nobody > (Revisited)
Yep, rather good. I was especially interested in the "pin type" metallic
airflow heatsink, so I dug a little deeper and did some more calculations.
For what you have, you need about a 3"x4" rectangle with about a
base-plate of about 1/4". The pins should be about 1-1 1/2" high with
about a 3/16" diameter at base tapering down to about 1/64" at the top,
spaced at about 1/4" apart in a square or hex pattern.
The installation is a little tricky, and some improvisation may be needed.
Temporarily place your new heatsink on the top of your head, with the
pins (not the base) touching your scalp. Secure it (temporarily) with
ducttape or one of the silicone-based double-sticky adhesives.
Now, walk out in the nearest "heavy-traffic" street. Wait for a bus to
approach, and if it's not stopping, run out in front of it and bow
toward it, as if aiming the heatsink at the front of the bus.
I think that will cure your TDP heat issues.
--
"Shit this is it, all the pieces do fit.
We're like that crazy old man jumping
out of the alleyway with a baseball bat,
saying, "Remember me motherfucker?"
Jim “Dandy” Mangrum
nm...@cam.ac.uk
Skybuck Flying
> On Fri, 20 May 2011 14:40:57 +0200, Skybuck Flying wrote:
>
>>Perhaps but there is some pressure going on in the case I bet ! ;)
>
> What does that mean: "pressure going on in the case?"
Pressure against all kind of components.
Pressure against the air itself, back fan spinning more slowly then front
fan.
Or CPU fan counter acting.
Or obstacles in path.
Bye,
Skybuck.