PSU Cooling
Doughboy
What I'd like to know is, how come my audio amp, which the manual
tells me is rated at 50W for each of the front channels, and 50W for
the centre and rear channels, and has a power consumption of 550W,
doesn't need a fan to operate, whereas my PC PSU, which is rated at
350W, does?
Sure, my amp gives off a fair bit of heat, but I'd rather take heat
than the noise my PC's PSU fan makes.
Doughboy
Arthur Jernberg
your receiver and be lulled to sleep by it's drone??
"Doughboy" <an...@invalid.com> wrote in message
news:si1m1vgr4o9ajfqfk...@4ax.com...
CoyoteBoy
larger boxes giving more room for convective cooling and venting, whereas
PSUs are packed into tiny boxes with bad airflow for free convection.
JB
Doughboy
If this is the only reason, then maybe I could repackage the PC's PSU
components into a larger box, and do away with the fan? Perhaps a box
on the *outside* of the PC case, with vent holes on 5 sides (only side
not vented being the bottom)?
Doughboy
Arthur Jernberg
PSU. It needs to be contained within some type to faraday shield, I.E. Metal
case to eliminate unwanted transmission of interfering signals. Present
designs, inside a metal case and then placed also inside another case tend
to accomplish that fairly well. Having it external may present other
problems that you may not be really happy with. It would be possible to
spread the supply out and re-design it to allow more efficient cooling but
definite concern as to the RFI generated needs to be considered. Not only
for your own installation but also for neighbors.
CoyoteBoy
vertical with vents on all sides - free convection in the best possible
way - providing you had a hole-y case top and bottom minimum.
I have considered linking a micro heat exchanger to my CPU to run tap water
through before now, maybe the same system could be used to cool the PSU.
Better be waterproof though! My mate is doing research into laser sintered
micro heat exchangers at uni - I'm gonna get him to make me one..... :)
JB
--
test...@tislabs.com <NAI Spam Filter address
Doughboy <an...@invalid.com> wrote in message
news:d05o1vkent7jb4u19...@4ax.com...
Doughboy
<stu...@comcast.net> wrote:
>Major problem would be the required shielding necessary for the re-designed
>PSU. It needs to be contained within some type to faraday shield, I.E. Metal
>case to eliminate unwanted transmission of interfering signals. Present
>designs, inside a metal case and then placed also inside another case tend
>to accomplish that fairly well. Having it external may present other
>problems that you may not be really happy with. It would be possible to
>spread the supply out and re-design it to allow more efficient cooling but
>definite concern as to the RFI generated needs to be considered. Not only
>for your own installation but also for neighbors.
I don't see that this would be a major problem to get round. The back
panel of the PSU when in the case is exposed to the outside anyway, so
I would have thought that if RFI was a problem, it would still escape
from there. And my amp's PSU just sits in the case, whose top panel is
filled with lots of vent holes, so plenty of routes for RFI to escape
from, although I've never had any problems from it. Nor have my
neighbours, as far as I know.
Having said that, if it is necessary to contain the PSU case in a
further metal box, that's easy enough.
I'm not even sure that it would be necessary to 'spread the supply out
and re-design it'. Surely just putting it in a well-ventilated box,
external to the PC's case should be sufficient. If not, I might as
well design and build a new PSU from scratch.
Apparently, PC PSU's are 70% efficient (which is why they generate
heat). Is this due to using cheap components (in which case, how easy
/ cost effective would it be to make it more efficient by changing
components), or due to the design?
Doughboy
Doughboy
<coyot...@hotmail.com> wrote:
>better still, mount it to the back of the case with the base of the PSU
>vertical with vents on all sides - free convection in the best possible
>way - providing you had a hole-y case top and bottom minimum.
If I did this, what would work best, having the hot components at the
top, or the bottom, of the PSU?
Personally, I think mounting it to the side of the case would be
better, as there's too many cables, etc on the back, and the PSU would
obstruct the airflow out the back of the case.
Doughboy
CoyoteBoy
occurs from a flat plate mounted vertically - this allows the air to rise
freely either side and cool it. Providing they are mounted side on, having
them at the bottom as the entry air would be slightly colder for them (the
critical ones) - this would raise the average temp inside the PSU but the
other components will cope IMO and the enclosed body of warm air escaping at
the top will provide better bouyancy to suck more cold air in, I'm not sure
on that, thats my prediction! I can draw a diagram to explain what I mean
better if you want.
All the PSU's I've ever destroyed have had the heat generating components
mounted on flate plate heatsinks, mounted vertically as thermodynamics
suggests is best. The one time I've seen this not to be the case was on my
graphics card where the processor is mounted underneath the card - the
thermodynamically worst position for it! Not the WAS in that sentence, it
ended up so hot even with a spare fan on it that its little life was
shortened drastically - and it was a known problem with the CL Blaster
Banshee.
So overall, put hot components at the bottom with their heatsinks mounted so
that the main body or the main body of fins is vertical allowing free
flowing air through it to remove heat.
If it gets more complicated than this I'll have to check my textbooks on the
subject!
JB
--
test...@tislabs.com <NAI Spam Filter address
Doughboy <an...@invalid.com> wrote in message
news:bkes1vsh6heuhufav...@4ax.com...
---
Outgoing mail is certified Virus Free.
Checked by AVG anti-virus system (http://www.grisoft.com).
Version: 6.0.441 / Virus Database: 247 - Release Date: 09/01/03
Doughboy
<coyot...@hotmail.com> wrote:
>OK, by the laws of thermodynamics the best possible convection cooling
>occurs from a flat plate mounted vertically - this allows the air to rise
>freely either side and cool it. Providing they are mounted side on, having
>them at the bottom as the entry air would be slightly colder for them (the
>critical ones) - this would raise the average temp inside the PSU but the
>other components will cope IMO and the enclosed body of warm air escaping at
>the top will provide better bouyancy to suck more cold air in, I'm not sure
>on that, thats my prediction! I can draw a diagram to explain what I mean
>better if you want.
That makes complete sense. I knew that the bottom would be colder than
the top (Heat rises. Wooh, I learnt something in physics!), but wasn't
sure where it would be best to locate the hot components.
>All the PSU's I've ever destroyed have had the heat generating components
>mounted on flate plate heatsinks, mounted vertically as thermodynamics
>suggests is best. The one time I've seen this not to be the case was on my
>graphics card where the processor is mounted underneath the card - the
>thermodynamically worst position for it! Not the WAS in that sentence, it
>ended up so hot even with a spare fan on it that its little life was
>shortened drastically - and it was a known problem with the CL Blaster
>Banshee.
I always thought the way graphics cards are made was a bit dumb and
that it would make more sense to have the hot components, and fan, on
the topside where the heat could rise to the extraction fan in the
PSU, rather than blowing hot air onto whatever cards you have under
the graphics card.
I'm not sure how to go about mounting the heatsinks vertically. In
it's normal orientation the heatsinks are like this:
_________________________
| |
extraction fan | | | | air intake vents
this side | | | | this side
| | | |
|______________|____|___|
So if I turned the case on it's side, they'd be like this:
Large hole for heat to exit up here
___ ___
| |
| |
| |
|------- |
| |
|------- |
| |
| |
|_ _ _ _ __|
vent holes for cool air to enter
down here
Whereas you seem to be saying it should be like this:
___ ___
| |
| |
| |
| |
| | | |
| | | |
| | | |
| | | |
|_ _ _ _ __|
But seeing as the heatsinks are mounted on semiconductors, which are
on a circuit board, if I were to put the existing board so that the
heatsinks were orientated as above, the circuit board would block the
air intake holes below it.
The only way I can think of to get round this, is to take all the
components of the existing board, and put the semiconductors, and
heatsinks, on a small board by themselves, which would block some of
the air intake vents, but allow air to rise either side of the board.
There would still be no air inbetween the heatsinks though, so perhaps
it would be better to have each semiconductor and heatsink on it's own
board, which would allow a gap for air to flow between them. Have you
got any ideas about this?
>So overall, put hot components at the bottom with their heatsinks mounted so
>that the main body or the main body of fins is vertical allowing free
>flowing air through it to remove heat.
>
>If it gets more complicated than this I'll have to check my textbooks on the
>subject!
Sorry if I just made it more complicated ;)
Doughboy