heat and more heat with heatsinks? what do you think?

  • Thread starter Thread starter Crackles McFarly
  • Start date Start date
C

Crackles McFarly

Just hold on a sec, I'll get to my point in a second..patience please
;-)

I recall in school building a bench power supply for an exam grade. We
were told about heatsinks but not told how large or even if a fan
would help. A lot of us use a small heatsink which was fine for the
volt regulators which were super hot. One guy added a very crappy tiny
cheap fan to blow over his heatsink and it made a huge difference.
Many COPIED his idea, I did not. I decided to go bigger and eventually
made the interior and back of the PSU into a heatsink. I used cheap
materials and correct compound. MY version of passive cooling was
better than the whole class and none of the idiots knew why; actually
I didn't either until 2 weeks or so later.

I had an idea about all this CPU/over comp heat problem. What if we
simply made a case that had a large fin heatsink, say 15-20 pounds.
The Cpu and other components would connect to it and have correct
compounds. I do not know the math on this but the surface area in
combination with the mass of the heatsink, compared to what we're
doing now, would have to be enormous in comparison.

I would think, based on my past experiences with dissipation of heat
in non-computers, that the percentage improvement in this design would
be on the order of 100%+ in even the most modest design.

It wouldn't be a popular design mostly because of the WEIGHT, making
the total case weigh more than 80 pounds! You wouldn't need fans for
it either, adding them would be pointless and silly with this size of
heatsink.

Materials would be cheap since the size would be so massive.

Has anything like this been attempted with a home PC?????


I know I've never owned a TV that used active heatsinks and they last
for over a decade in warm environments. I opened one up about 5 years
ago, it was a dead tv, and recall the large heatsink and the weight
was nearly 5 pounds for one spot.


Yes I know I am not asking or bringing up a NEW idea but I would like
to know if it's been tried on a PC, what were the results and if
anyone at all would want such a heavy case?


thanks, and sorry if this seemed a bit off-topic but I thought it was
interesting.

truly yours,
crackles
 
In alt.comp.hardware.pc-homebuilt Crackles McFarly
Just hold on a sec, I'll get to my point in a second..patience please
;-)

I recall in school building a bench power supply for an exam grade. We
were told about heatsinks but not told how large or even if a fan
would help. A lot of us use a small heatsink which was fine for the
volt regulators which were super hot. One guy added a very crappy tiny
cheap fan to blow over his heatsink and it made a huge difference.
Many COPIED his idea, I did not. I decided to go bigger and eventually
made the interior and back of the PSU into a heatsink. I used cheap
materials and correct compound. MY version of passive cooling was
better than the whole class and none of the idiots knew why; actually
I didn't either until 2 weeks or so later.

I had an idea about all this CPU/over comp heat problem. What if we
simply made a case that had a large fin heatsink, say 15-20 pounds.
The Cpu and other components would connect to it and have correct
compounds. I do not know the math on this but the surface area in
combination with the mass of the heatsink, compared to what we're
doing now, would have to be enormous in comparison.

I would think, based on my past experiences with dissipation of heat
in non-computers, that the percentage improvement in this design would
be on the order of 100%+ in even the most modest design.

It wouldn't be a popular design mostly because of the WEIGHT, making
the total case weigh more than 80 pounds! You wouldn't need fans for
it either, adding them would be pointless and silly with this size of
heatsink.

Materials would be cheap since the size would be so massive.

Has anything like this been attempted with a home PC?????
Click to expand...
Not with a *home* PC; but passive cooling is more common than you might
think.

There are passive heat-sinks available:
http://www.pcper.com/article.php?aid=286
Here's a guy that *really* cheated; and ran a passive water-cooling
system out to his garage. ;-}
http://www.overclockers.com/tips1093/
Another inside-the-case passive cooler:
http://www.bigbruin.com/reviews05/review.php?item=thermalrighthr01&file=1
Here's somebody building PCs somewhat like you describe, where the case
is the cooler. A bit pricey, but silent.
 
Crackles said:
Just hold on a sec, I'll get to my point in a second..patience please
;-)

I recall in school building a bench power supply for an exam grade. We
were told about heatsinks but not told how large or even if a fan
would help. A lot of us use a small heatsink which was fine for the
volt regulators which were super hot. One guy added a very crappy tiny
cheap fan to blow over his heatsink and it made a huge difference.
Many COPIED his idea, I did not. I decided to go bigger and eventually
made the interior and back of the PSU into a heatsink. I used cheap
materials and correct compound. MY version of passive cooling was
better than the whole class and none of the idiots knew why; actually
I didn't either until 2 weeks or so later.

I had an idea about all this CPU/over comp heat problem. What if we
simply made a case that had a large fin heatsink, say 15-20 pounds.
The Cpu and other components would connect to it and have correct
compounds. I do not know the math on this but the surface area in
combination with the mass of the heatsink, compared to what we're
doing now, would have to be enormous in comparison.

I would think, based on my past experiences with dissipation of heat
in non-computers, that the percentage improvement in this design would
be on the order of 100%+ in even the most modest design.

It wouldn't be a popular design mostly because of the WEIGHT, making
the total case weigh more than 80 pounds! You wouldn't need fans for
it either, adding them would be pointless and silly with this size of
heatsink.

Materials would be cheap since the size would be so massive.

Has anything like this been attempted with a home PC?????


I know I've never owned a TV that used active heatsinks and they last
for over a decade in warm environments. I opened one up about 5 years
ago, it was a dead tv, and recall the large heatsink and the weight
was nearly 5 pounds for one spot.


Yes I know I am not asking or bringing up a NEW idea but I would like
to know if it's been tried on a PC, what were the results and if
anyone at all would want such a heavy case?


thanks, and sorry if this seemed a bit off-topic but I thought it was
interesting.

truly yours,
crackles
Click to expand...

TNN500AF uses heatpipes to transfer major heat sources, to the outer heatsink
fins.

http://www.xbitlabs.com/articles/coolers/display/zalman-tnn500af_3.html
http://lib.store.yahoo.net/lib/xoxide/tnn500af3.jpg

They have a smaller one, for microATX motherboards.

http://www.zalman.co.kr/eng/product/view.asp?idx=176&code=020

These are expensive, and that is the main reason nobody you know,
owns one.

Paul
 
I had an idea about all this CPU/over comp heat problem. What if we
simply made a case that had a large fin heatsink, say 15-20 pounds.
The Cpu and other components would connect to it and have correct
compounds. I do not know the math on this but the surface area in
combination with the mass of the heatsink, compared to what we're
doing now, would have to be enormous in comparison.

I would think, based on my past experiences with dissipation of heat
in non-computers, that the percentage improvement in this design would
be on the order of 100%+ in even the most modest design.

It wouldn't be a popular design mostly because of the WEIGHT, making
the total case weigh more than 80 pounds! You wouldn't need fans for
it either, adding them would be pointless and silly with this size of
heatsink.

Materials would be cheap since the size would be so massive.

Has anything like this been attempted with a home PC?????
Click to expand...

There are a number of such products available, they work and
all that. The only limitation iof that they are prohibitively expensive
for a mass market. For example, Zalman's excellent fanless
workstation cases cost as much as the complete new computer
that may go inside:

http://www.pcmall.com/pcmall/shop/detail.asp?dpno=702076&ref=lmcd

DK
 
Just hold on a sec, I'll get to my point in a second..patience please
;-)

I recall in school building a bench power supply for an exam grade. We
were told about heatsinks but not told how large or even if a fan
would help. A lot of us use a small heatsink which was fine for the
volt regulators which were super hot. One guy added a very crappy tiny
cheap fan to blow over his heatsink and it made a huge difference.
Many COPIED his idea, I did not. I decided to go bigger and eventually
made the interior and back of the PSU into a heatsink. I used cheap
materials and correct compound. MY version of passive cooling was
better than the whole class and none of the idiots knew why; actually
I didn't either until 2 weeks or so later.

I had an idea about all this CPU/over comp heat problem. What if we
simply made a case that had a large fin heatsink, say 15-20 pounds.
The Cpu and other components would connect to it and have correct
compounds. I do not know the math on this but the surface area in
combination with the mass of the heatsink, compared to what we're
doing now, would have to be enormous in comparison.

I would think, based on my past experiences with dissipation of heat
in non-computers, that the percentage improvement in this design would
be on the order of 100%+ in even the most modest design.

It wouldn't be a popular design mostly because of the WEIGHT, making
the total case weigh more than 80 pounds! You wouldn't need fans for
it either, adding them would be pointless and silly with this size of
heatsink.

Materials would be cheap since the size would be so massive.

Has anything like this been attempted with a home PC?????

I know I've never owned a TV that used active heatsinks and they last
for over a decade in warm environments. I opened one up about 5 years
ago, it was a dead tv, and recall the large heatsink and the weight
was nearly 5 pounds for one spot.

Yes I know I am not asking or bringing up a NEW idea but I would like
to know if it's been tried on a PC, what were the results and if
anyone at all would want such a heavy case?

thanks, and sorry if this seemed a bit off-topic but I thought it was
interesting.

truly yours,
crackles
Click to expand...

a question i really want to ask you, is what type of school (high
school? university?) , and what subject, does one build that sort of
thing?
I don't know of any high school or university.. I'm interested,
because it'd a good background to have.


On to an attempted answer.. I think it's a great idea.. For cooling a
case. But I am not convinced that it'd cool a processor.
Nevertheless, cooling processors passively is not such a problem these
days. There are ULV processors, that take little passive heatsinks.
And there are large passive heatsinks for processors. *

The thing that I am not sure has been cooled CHEAPLY is a computer
case. Zalman had an expensive one, though it has dropped alot in
price, but it covers cooling many components, not just te case. So
cooling the case passively should be very cheap. But maybe just
leaving the lid off is enough?!


I did look some years ago. The closest i found were watercooled cases
on sale. But on contacting the company, they said they had a small
fan.
But that check was back in around 2003 when CPU Heatsinks weren't yet
passive. Maybe things have improved since

Your idea of a large heatsink on a PSU was actually done, in around
2003 or 2004.. Prob the first fanless regular size/watt PSU
commercially available by SilentMaxx. It was beautiful.
A bit like this
http://www.rockhounding.net/projects/quiet-computing/pics/fanless-power-2.jpeg
http://img175.imageshack.us/img175/8265/fanlesspower2ru4.jpg

That looks like what you describe.. Though it was never meant to cool
a cpu in any way.

A poster called Ken Maltby is into building heatsinks or heatpipes..


*
I haven't tried the small on ULV, they prob have issues. The large
ones can have issues,
- case not wide enough or not enough space between psu and cpu
socket. (quietpc sell cases and heatsinks and say their cases fit
them. you could check with them)
- heatsink not fitting on mbrd 'cos capacitors in the way!
http://www.pureoverclock.com/article643-2.html (some heatsinks may
have a mbrd compatibility list to get around this)
- I did read a review, can't find it now..that speculated that some
passive heatsinks may not have cooled the voltage regulators, even
with the optional fan, due to fan placement, maybe being on the side
instead of on top. So, how much moreso without a fan. If that
speculation was correct. But now it's prob ok 'cos many MBRDs passive
cool voltage regulators as well as North Bridge and SB, and maybe all
other MBRD parts.

http://www.hardwaresecrets.com/article/339
"ASUS M2N32-SLI"
"What really catches the eye when looking to this motherboard is its
passive cooling"
 
There are a number of such products available, they work and
all that. The only limitation iof that they are prohibitively expensive
for a mass market. For example, Zalman's excellent fanless
workstation cases cost as much as the complete new computer
that may go inside:

http://www.pcmall.com/pcmall/shop/detail.asp?dpno=702076&ref=lmcd
Click to expand...

$1400 !?
I would have never thought this idea would be that expensive?

Must be the R&D expense because it couldn't be the parts?
 
a question i really want to ask you, is what type of school (high
school? university?) , and what subject, does one build that sort of
thing?
I don't know of any high school or university.. I'm interested,
because it'd a good background to have.
Click to expand...

I was a higher level electrical engineering course for designing power
supplies. Not static voltages but ones that varied from near-zero to
13+ volts. The idea was to keep the voltage regulators from drifting
and making the voltage slowly drop.

Was tested on a large power resister to make the result come quicker.
The grade was based on the drifting and also how much time it took for
a drift to be seen.

No drifting was the goal but very few could do that for more than 10
minutes...The old heat got the best of most of us.

The worst was a drift of 5 volts in under 3 minutes while the best
went nearly an HOUR and even then only dropped less than a volt.

It was an interesting little experiment for me at least.
 
In alt.comp.hardware.pc-homebuilt Crackles McFarly
$1400 !?
I would have never thought this idea would be that expensive?

Must be the R&D expense because it couldn't be the parts?
Click to expand...

Actually, it's probably mostly the special parts.
It costs MONEY to make and machine special cases and such.
Just making a pretty standard PC case, with only minor mods costs
several hundred dollars more than the ones you buy made overseas that
are stamped out by the hundreds of thousands, usually for less than a
hundred bucks, complete with power-supply, speaker, and basic wiring.

Economics.
 
In alt.comp.hardware.pc-homebuilt Crackles McFarly
I was a higher level electrical engineering course for designing power
supplies. Not static voltages but ones that varied from near-zero to
13+ volts. The idea was to keep the voltage regulators from drifting
and making the voltage slowly drop.

Was tested on a large power resister to make the result come quicker.
The grade was based on the drifting and also how much time it took for
a drift to be seen.

No drifting was the goal but very few could do that for more than 10
minutes...The old heat got the best of most of us.

The worst was a drift of 5 volts in under 3 minutes while the best
went nearly an HOUR and even then only dropped less than a volt.

It was an interesting little experiment for me at least.
Click to expand...
Hmmm ... Must have been pretty simple supplies, or with a set specific
diagram. Any power-supply that *I* ever designed that drifted more than
a tenth of a volt under no load or full load for several *days*, would
have been scrapped and redesigned.

Mostly I expected far better than that. I wanted my calibrated setting
to match my output if possible; and far more-so I didn't want the damned
thing *drifting* with changes of input-source, use, or output-load. If
I couldn't rely on my power-supply to remain stable, then how could I
trust my circuit to continue working?

(Of course, a complementary task was to make a circuit WORK with a wide
range of input volages; allowing the use of cheaper power-supply
circuits. But to test *that*, called for a reliable test-bench supply.)

It wasn't until I was pretty much out of the business of designing audio
circuits that I finally managed to afford a GREAT HP power-supply with
even better specs than those I designed myself. Great, huh? You
finally get decent equipment when you no longer need it. However, I
still have one of the supplies on my bench that I built myself. For a
*tiny* supply, I haven't yet found anything to match it in
specifications or ease-of-use. Both voltage and current-regulated in a
tiny 3" by 4" box. I tore out the guts of a truly *cheap* commercial
supply I got that was pretty worthless, kept the knobs and case, and
replaced pretty much everything else. A bit of a chore making that
tight a control of the output without the tight feedback circuit wanting
to oscillate; but it does work wonderfully. "Hard as a brick."

One of the tricks of the device was to make the pre-regulation of the
voltage-standard inside independant of the circuit it was regulating;
yet using the same standard. (If the voltage to the standard varies
under line or load ...)

I've seen that happen on all too many otherwise good supplies. Under
full rated load, the supply to the reference dropped, which dropped the
output, even though the supply itself could have otherwise handled the
extra current.
 
In alt.comp.hardware.pc-homebuilt Crackles McFarly

Hmmm ... Must have been pretty simple supplies, or with a set specific
diagram. Any power-supply that *I* ever designed that drifted more than
a tenth of a volt under no load or full load for several *days*, would
have been scrapped and redesigned.

Mostly I expected far better than that. I wanted my calibrated setting
to match my output if possible; and far more-so I didn't want the damned
thing *drifting* with changes of input-source, use, or output-load. If
I couldn't rely on my power-supply to remain stable, then how could I
trust my circuit to continue working?

(Of course, a complementary task was to make a circuit WORK with a wide
range of input volages; allowing the use of cheaper power-supply
circuits. But to test *that*, called for a reliable test-bench supply.)

It wasn't until I was pretty much out of the business of designing audio
circuits that I finally managed to afford a GREAT HP power-supply with
even better specs than those I designed myself. Great, huh? You
finally get decent equipment when you no longer need it. However, I
still have one of the supplies on my bench that I built myself. For a
*tiny* supply, I haven't yet found anything to match it in
specifications or ease-of-use. Both voltage and current-regulated in a
tiny 3" by 4" box. I tore out the guts of a truly *cheap* commercial
supply I got that was pretty worthless, kept the knobs and case, and
replaced pretty much everything else. A bit of a chore making that
tight a control of the output without the tight feedback circuit wanting
to oscillate; but it does work wonderfully. "Hard as a brick."

One of the tricks of the device was to make the pre-regulation of the
voltage-standard inside independant of the circuit it was regulating;
yet using the same standard. (If the voltage to the standard varies
under line or load ...)

I've seen that happen on all too many otherwise good supplies. Under
full rated load, the supply to the reference dropped, which dropped the
output, even though the supply itself could have otherwise handled the
extra current.
Click to expand...


If only I had designed mine with a superconductor..
;-)
 
In alt.comp.hardware.pc-homebuilt Crackles McFarly
If only I had designed mine with a superconductor..
;-)
Click to expand...

Would have been nice to have.
Even better would be a transistor which didn't lose gain or change
bias-point as it changed temperature. Especially on audio circuits.

You'd LOSE gain as the temperature of the junction rose; at the same
time the bias-point dropped so less voltage would cause more current.
Compensating offsets, you might think ... if you were optimistic or an
idiot. Thermal runaway being *so* easy to run into; while damned
difficult to avoid or fix.

So *many* days, weeks, and months spent designing, testing, and working
on biasing arrangements for power amplifiers; especially "CLASS A" amps.

(Everything from special matched diodes, biasing circuits built into
power-transistors, combinations of thermistors and resistors, "active"
circuits, etc.. Then Heaven help you if one or more of the devices
weren't mounted properly ....)

In comparison, power supplies were easy.
 
Back
Top