Cooling Questions

It's not nessarily a 'bad' design, just a different one.

Dell does it that way on purpose. That way, by using a large finned
heat-sink on the CPU along with a duct to dirrect the airfow to/from the fan
on the rear of the case, they have enough room on the back of the case to
use a larger fan.

A larger fan can spin at a slower RPM yet give as much airflow as a faster
small fan directly on the heat-sink.

A slower fan translates into less fan noise, making the system in theory,
quiter, which in importatnt to many customers.


"127.0.0.1" wrote in message
ink.net...

"Peter Cavan" wrote in message
...
Hi
The thing is, there isn't a CPU fan on the CPU itself. There is just a
HS
on
the CPU. There is a big fan on the back of the case which is vented over
the
CPU's HS. I think the fan is 12cm, so I don't think you could add a
bigger
one. Could I add a fan to the existing HS? I'm not sure if it would fit
with
the vent on top.
Thanks
PC

thats a poor design.
i would purchase a quality HS + Fan and remove that case fan. if there are
other case fans to move the air, it should be sufficient. or put a smaller
fan to replace that obstrusive case fan. there are other options on
circulating the case. but the most important part of your problem is
getting
a proper HS + fan installed.

"127.0.0.1" wrote in
ink.net:


"Peter Cavan" wrote in message
...
Hi
The thing is, there isn't a CPU fan on the CPU itself. There is just
a HS
on
the CPU. There is a big fan on the back of the case which is vented
over
the
CPU's HS.
snip
-----
thats a poor design.
i would purchase a quality HS + Fan and remove that case fan. if there
are other case fans to move the air, it should be sufficient. or put a
smaller fan to replace that obstrusive case fan. there are other
options on circulating the case. but the most important part of your
problem is getting a proper HS + fan installed.

/\/\/\/\
Actually .. the ducted fan is state of the art.

A fan bolted to the top of a heatsink only recirculates the hot air that
collects around it (the bigger the fan you put there .. the faster it
recirculates the same hot air).

The ducted fan sucks the heat from the CPU and moves it outside the
case.

In addition ... by "not touching" the CPU heatsink most of the noise
associated with CPU cooling is eliminated.

If you want to see state of the art thermal design recommendations for
PC cases (that look just like what is in your Dell) then look at the
thermal PDF file at the ATX formfactor organization site:

http://www.formfactors.org/developer...roatxspecs.htm

IMHO as always ...
campbell

"Timothy Drouillard" wrote in message
...
It's not nessarily a 'bad' design, just a different one.

Dell does it that way on purpose. That way, by using a large finned
heat-sink on the CPU along with a duct to dirrect the airfow to/from the
fan
on the rear of the case, they have enough room on the back of the case to
use a larger fan.

A larger fan can spin at a slower RPM yet give as much airflow as a faster
small fan directly on the heat-sink.

A slower fan translates into less fan noise, making the system in theory,
quiter, which in importatnt to many customers.


a choice of quieter with overheating vs. a bit noisier with efficient
cooling....
they are still years behind macintosh designs (cube). i forget what the
original processor is. but having no fan on a cpu HS is not a good design.
i'd rather see the dell executives cut their bonuses to save on cost than
skimping out on proper cpu cooling.

Overheating??

Possibly for some people but I don't think so for most of the Dimension 8200
owners.

I bought my Dimension 8200 in June of 2002
2.53Ghz (fastest on the market at the time)
1gig PC800 ram
Geforce TI4600
36gig WD Raptor
120gig Seagate SATA
250gig WD
Audigy 2 Platinum
DVD
Plextor 52x premium CD-RW
3Com NIC

It's been running pretty much 24/7 since I bought it, with absolutley no
overheating problems at all.
No fans added, subtracted, or modified.
Pretty darn quite also.

"127.0.0.1" wrote in message
ink.net...

a choice of quieter with overheating vs. a bit noisier with efficient
cooling....
they are still years behind macintosh designs (cube). i forget what the
original processor is. but having no fan on a cpu HS is not a good design.
i'd rather see the dell executives cut their bonuses to save on cost than
skimping out on proper cpu cooling.

On Sat, 30 Aug 2003 21:45:14 +0100, "Peter Cavan"
wrote:

Hi
The thing is, there isn't a CPU fan on the CPU itself. There is just a HS on
the CPU. There is a big fan on the back of the case which is vented over the
CPU's HS. I think the fan is 12cm, so I don't think you could add a bigger
one. Could I add a fan to the existing HS? I'm not sure if it would fit with
the vent on top.
Thanks
PC


You'd probably have to fabricate some sort of attachment method to
attach a fan to the existing heatsink. It may be possible but likely
worth the $ to just buy a different heatsink instead. With either a
fan attached or a different heatsink you'd most likely benefit more
from removing the duct than leaving it on.


Dave

Peter Cavan wrote:
It's 60-70 degrees Celsius. I've done too much stuff to the machine to list
all the changes. There is no easy way that I can see to add a fan to the
front or side.

Contrary to the knee jerk reaction of those who think anything different than
what they're used to is a 'bad design', the Dell design, in general, is quite
good but knowing what you've changed is precisely what's needed because, odds
are, that's where something is sufficiently different, vs what the system was
originally designed for, to be causing higher temperatures.

I can certainly see why Dell doesn't want to warrantee cover it if there are a
ton of changes but did they tell you what the temperature should be? Is the
temperature reading you have from the CPU thermal diode or an under socket
thermistor? Do you have a case temperature reading?

When did it start being 'hot'? After one, or more, of your additions/mods or did
it work fine for a while and then suddenly get 'hot', or did it gradually get
warmer over time? Do you have it shoved into a closed desk bay so it can't
circulate air or is it in the open? Did you move, alter, or remove any internal
ducting/air guides?

Is that the original CPU?

I missed your original post so could you repost the model number so I can check
the Dell site for the mechanical layout docs?

I was considering a new case that would easily accomodate better cooling,
this would involve simply moving everything into a new case. Does anybody
know if there would be any problems with moving a Dell system into a
non-Dell case? I was thinking, for example, about things like PSUs, power
buttons, and front panel USB, headphone etc.
Thanks
Peter Cavan

"TC" wrote in message
...
"127.0.0.1" wrote in
ink.net:

thats a poor design.
i would purchase a quality HS + Fan and remove that case fan. if there
are other case fans to move the air, it should be sufficient. or put a
smaller fan to replace that obstrusive case fan. there are other
options on circulating the case. but the most important part of your
problem is getting a proper HS + fan installed.

/\/\/\/\
Actually .. the ducted fan is state of the art.

A fan bolted to the top of a heatsink only recirculates the hot air that
collects around it (the bigger the fan you put there .. the faster it
recirculates the same hot air).

The ducted fan sucks the heat from the CPU and moves it outside the
case.

In addition ... by "not touching" the CPU heatsink most of the noise
associated with CPU cooling is eliminated.

If you want to see state of the art thermal design recommendations for
PC cases (that look just like what is in your Dell) then look at the
thermal PDF file at the ATX formfactor organization site:

http://www.formfactors.org/developer...roatxspecs.htm

IMHO as always ...
campbell


I agree 100%

In addition a ducted fan system eliminates the "dead zone" effect caused by
top mounted fans.

The center of the heatsink ... which is the hottest due to where the CPU die
is ... doesn't get blown on directly because the top-mounted fan's blades
obviously don't go all the way to the middle. In fact, the air there just
kind of swirls around in a "vortex" of sorts. Heat only finally escapes when
it moves to the outer edges of the heatsink where it is blown on directly.

A duct cooling system pulls air across the heatsink sideways, hitting all of
the fins, even the hottest ones in the middle ... resulting in much more
even cooling of the heatsink fins.

And on the noise front .... duct style cooling systems can utilize larger
fans, which can move more air at lower RPM. Resulting in even greater
reduction in noise generated.

It really makes you wonder if this guy thinks that Dell is "cutting costs"
by engineering new cooling solutions for each new system they come out with.

A duct cooling system involves a custom heatsink, and a custom duct housing
which has to be changed for every new system model. I would think that just
plonking a decent CPU "HS/fan" type cooler on there would be much easier,
and cheaper. Why then re-invent the wheel with each new model?

Obviously ... because it WORKS BETTER!

Get a clue 127!

And TC, thank you for saying what I had been itching to since the first
ignorant post I saw from this guy.


Drumguy

On Sat, 30 Aug 2003 21:49:12 GMT, TC wrote:


/\/\/\/\
Actually .. the ducted fan is state of the art.

Nope, Gateway, Compaq and others have been using passive coolers and
ducts since the Pentium 1 days. The CPU runs hotter as a result,
then and today. Ductwork is merely cheaper than a high-quality
heatsink, which would keep the CPU cooler AND be as quiet. The other
benefit is that these systems are being shipped, sometimes great
distances under less-than-ideal conditions, so reduced heatsink weight
as on the Dells is reducing the chance of RMA due to socket or CPU
damage.

Proper fan selection is crucial to achieve low noise on an active
heatsink, but so it was also when Dell built their ducted system.


Dave

"kony" wrote in message
...
On Sat, 30 Aug 2003 21:49:12 GMT, TC wrote:


/\/\/\/\
Actually .. the ducted fan is state of the art.

Nope, Gateway, Compaq and others have been using passive coolers and
ducts since the Pentium 1 days. The CPU runs hotter as a result,
then and today. Ductwork is merely cheaper than a high-quality
heatsink, which would keep the CPU cooler AND be as quiet. The other
benefit is that these systems are being shipped, sometimes great
distances under less-than-ideal conditions, so reduced heatsink weight
as on the Dells is reducing the chance of RMA due to socket or CPU
damage.

Proper fan selection is crucial to achieve low noise on an active
heatsink, but so it was also when Dell built their ducted system.


Dave


And why not go with a cheaper alternative if it works?

Sure, the Dell duct cooler may run hotter ... but Dell's main interest is
RMA's, and if their coolers didn't keep systems stable they would do
something different.

The truth is, duct cooling systems ARE better ... they allow Dell to get
away with what would seem crazy in terms of the number of fans and
individual fan speeds.

Imagine a duct system with a whisper quiet 120mm fan pulling as much air as
3 or 4 loud 60mm fans using a well made oversized copper passive heatsink.
It would blow away any active cooler you can name. Not to mention pulling
all that hot air outside of the case ... which would keep the internal
ambient temperatures down.

Just because Dell doesn't implement it to the best of it's potential doesn't
mean that it's inferior. In fact, it's what allows Dell to achieve stable
performance with such (on it's face) inadequate cooling.


Drumguy

On Sun, 31 Aug 2003 02:03:35 -0500, "drumguy1384"
wrote:

"kony" wrote in message
.. .
On Sat, 30 Aug 2003 21:49:12 GMT, TC wrote:


/\/\/\/\
Actually .. the ducted fan is state of the art.

Nope, Gateway, Compaq and others have been using passive coolers and
ducts since the Pentium 1 days. The CPU runs hotter as a result,
then and today. Ductwork is merely cheaper than a high-quality
heatsink, which would keep the CPU cooler AND be as quiet. The other
benefit is that these systems are being shipped, sometimes great
distances under less-than-ideal conditions, so reduced heatsink weight
as on the Dells is reducing the chance of RMA due to socket or CPU
damage.

Proper fan selection is crucial to achieve low noise on an active
heatsink, but so it was also when Dell built their ducted system.


And why not go with a cheaper alternative if it works?

It's cheaper for Dell, not necessarily for us. They would likely hit
the same price-points either way.

Sure, the Dell duct cooler may run hotter ... but Dell's main interest is
RMA's, and if their coolers didn't keep systems stable they would do
something different.

I suggest a reread of the OP, which stable.

Someday when boxes are static, sealed systems like a VCR, it could be
a good standard to follow, but PCs are user-configurable devices, even
if it does void the warranty. A chassis that's not very forgiving of
airflow changes can't accomodate user upgrades and/or may not last as
long. Why would Dell care about that... after all, it'll last till
the warranty is up so long as nothing is changed (which voids the
warranty), they assume.


The truth is, duct cooling systems ARE better ... they allow Dell to get
away with what would seem crazy in terms of the number of fans and
individual fan speeds.

Fans have been around a long time, it's not crazy or hard in any way
to put an appropriate amount of cooling in a system... no more, no
less. Better is relative... it's better for Dell, but not
necessarily for their customers. Dell as well as anyone else has the
resources to build quiet PCs with fans on the heatsink. What they did
instead IS an acceptable method of cooling, but still inferior when it
comes to noise... Yes, it's LOUDER. A ducted fan requires higher RPM
to move the same amount of air, and there is more turbulent noise
escaping the chassis. A lower-RPM rear fan with no duct can be
quieter while moving same amount of air though the system, while the
heatsink fan can have even less noise escaping the chassis than the
rear fan, and keep the CPU cooler than Dell's method.

Imagine a duct system with a whisper quiet 120mm fan pulling as much air as
3 or 4 loud 60mm fans using a well made oversized copper passive heatsink.

Imagine same 120mm fan without the duct, so it has a better airflow to
noise ratio.

120mm fans are great for moving a lot of air quietly, but it can't be
forgotten that axial fans are low-pressure air-movers, having a large
fan ducted down to create vacuume to pull air though a heatsink isn't
optimal airflow, it would be better to use a diagonal fan (which BTW
is not a mounting position but airflow characteristic), but even then
to maintain high enough airflow through the duct to have temps
parallel to active 'sinks, the fan would have to spin fast enough (or
an alternate type fan producing more noise at same RPM would be
needed) that it's not what most people would consider quiet-silent.


It would blow away any active cooler you can name.

It won't, because fans that create a lot of vacuume are noisey.
Without more vacuume than a traditional fan can provide at low noise
levels, there will always be less airflow though the heatsink than a
fan [of same or lesser (operating) noise while] atop a 'sink.

Users everywhere with these Dell sytems have higher CPU temps than
people using _good_ active coolers. The idealistic notions of the
ducted cooling system are contradicted by fact. The fact is that a
fan mounted on a heatsink can be quiet and still move more air through
the 'sink than the duct will, which results in lower temp. The
difference is great enough that air recirculated though the 'sink
doesn't reduce the effectiveness to the lower level of a ducted
cooling system. Removing the duct also allows the rear fan to move
more air. Greater airflow to noise ratio results, in addition to a
lower CPU temp.

The downside to active heatsink cooling is not performance, it's cost,
and possibly strain on the socket (or other mounting with a P4) during
transportation. Dell can buy 120mm fans and ducts in bulk for low
prices, unlike the situation an end-user will see, where a simple duct
is generic, not custom-designed for the system, and still costs WAY
too much for being a relatively simple piece of plastic. A relatively
elaborate, effectively designed duct might help, but would then might
be more expensive, less immune to dust buildup, and still it won't
change the anemic vacuume of the fan.

all that hot air outside of the case ... which would keep the internal
ambient temperatures down.

An oversimplified theory which goes against the evidence that our OP
and other Dell owners have hotter running CPUs. It's a good idea to
get the hot air out without it recirculating, but doing without the
CPU 'sink fan is not helping to accomplish that. If Dell had chosen
to they couldn've put a silent fan on the 'sink and still used a
duct... I'm not aways against ducts, but there are usually unnecessary
and often not a good substitute for active cooling.

Just because Dell doesn't implement it to the best of it's potential doesn't
mean that it's inferior. In fact, it's what allows Dell to achieve stable
performance with such (on it's face) inadequate cooling.

It's inferior from a cooling standpoint, almost equal from a noise
standpoint. A well-designed duct might be used, but we can't be so
sure Dell hasn't weighed all the alternatives and that their duct
isn't the best implementation from their perspective. It would have
to be the OEM's perspective at issue, since end users or smaller
builders have more variable cases and motherboards to consider (or to
put it another way, higher cost due to low volume, per each duct
design), so a custom duct could only be made at great expense or
hand-made, taking a fair amount of time (and to some time is money).

A duct in itself isn't necessarily a bad addition to a system, but it
does have different benefits than an active cooler. Lower temps,
lower noise, aren't two of the benefits unique to a duct-cooling
strategy... it's just that so many crap coolers with junk high-RPM
fans are out there, that duct-cooling is good in comparision to the
lowest-cost alternatives.


Dave