T
TheJoker
The fact is, you do not know the first thing about PCs.
You think that having a 800W power supply means that your computer is
using 800 Watts.
That's not how it works, idiot.
The fact is, you do not know the first thing about PCs.
You think that having a 800W power supply means that your computer is
using 800 Watts.
That's not how it works, idiot.
S
Skybuck Flying
TheJoker said:
I know how it works idiot, it's needed for "peaks".
Real usage could however still be 600 watts or so... or 400 watts, still a
lot of power and thus heat !
Bye,
Skybuck.
R
Robert Myers
The design challenge you face is getting the heat from the processorI 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.
A
Androcles
The design challenge you face is getting the heat from the processorI 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.
==========================================
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.
N
nik Simpson
Yeah, but if you put some wings on the side of the chassis you can have
the thing float in mid-air and save on floorspace ;-)
J
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."
F
FatBytestard
He won't get it.
I
I AM THAT I AM
SkyBuck's Shitty, Shitty, Bang Bawx?
S
Skybuck Flying
Fair enough...
Wikipedia has a nice article on it... it might be of some interest:
http://en.wikipedia.org/wiki/Heat_sink
Bye,
Skybuck.
Wikipedia has a nice article on it... it might be of some interest:
http://en.wikipedia.org/wiki/Heat_sink
Bye,
Skybuck.
S
Skybuck Flying
Jacob said:
Perhaps but there is some pressure going on in the case I bet !
So I don't trust those "pure" CFM ratings and theory !
=DAlso you guys gotta compensate for intel which lies about TDP !
Add at least 20 to 30% extra !
=DIt'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
P
Paul
Skybuck said:Also you guys gotta compensate for intel which lies about TDP !
Add at least 20 to 30% extra !
Not true. I have a Core2 2.6GHz processor with dual cores, rated
TDP is 65 watts, *measured* power with a clamp-on ammeter is
36W running Primt95 multithreaded load test. So in fact,
TDP can be conservative.
I have a 3GHz Core2 as well, and it also measures less than TDP.
On older generations of Intel products, measured power is closer
to TDP (thermal_design_power) value. I've also measured that
with my clamp-on ammeter, and that's how I know.
Paul
J
Jacob
Perhaps but there is some pressure going on in the case I bet !
What does that mean: "pressure going on in the case?"
T
The Great Attractor
It is actually between his ears.
His skull is so thick that being so full of shit is causing his blood
pressure to rise.
"Pressure going on in the case".
Bwuahahahahaha!
N
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
N
nmm1
Brain overheating?
Regards,
Nick Maclaren.
S
Skybuck Flying
Jacob said:
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.
S
StickThatInYourPipeAndSmokeIt
It does not get any more stupid than you are.