<!-- BBCode Start --><A HREF="http://www.directron.com/noise.html " TARGET="_blank">http://www.directron.com/noise.html </A><!-- BBCode End -->-- "The relative noise levels of 28 different 80mm case fans are measured under the same conditions."
<BR>
<BR>Amazing how poorly this is done, despite so-called good intentions. The test data conflicts with virtually all the fan listening I've been doing over the past several months. The methodology is, well .. here is the post I tossed up at the directron forum & rating feedback:
<BR>
<BR>-------------------
<BR>Serious problems with "Noise shootout of 80mm case fans" --
<BR>
<BR>1) thermally controlled fans in the mix. The ambient test temp of 70F guarantees these fans were running at minimal speed. The Enermax thermally controlled fan were probably getting 5V or less and providing less than half its rated airflow. Non-thermally controlled fans all ran at 12V.
<BR>
<BR>2) No airflow info! Without this, the noise info is only partly useful.
<BR>
<BR>3) No indication of whether the mic was placed in the airflow path: It is stated that "A case fan was kept vertical to the microphone to avoid the impact of the wind on the microphone. The arrangement of set up is shown below." But no such setup picture is provided.
<BR>
<BR>if the airflow comes anywhere close to the mic, wind noise DEFNITELY affects the noise measurement. This means that all the higher airflow fans are further penalized, because they will obviously have more airflow-induced mic wind noise than the low airflow ones. The mic always exacerbates the effect of the airflow noise.
<BR>
<BR>Now you could argue that this airflow noise is part and parcel of fan noise, but everyone knows that voltage controllers (such as the Zalman Fanmate 1 and similar devices that can be home-built with minimal soldering skills) can be used to slow any fan to a desired speed.
<BR>
<BR>For all these reasons, I think there is more misinformation in the page than there is info.
<BR>
<BR>I have done extensive listening and testing of many of the fans on this list, and I have at least a couple of samples of each model. I disagree strongly with the rankings.
<BR>
<BR>There is absolute no question that if you have a Pabst 8412NGL without ticking in the bearing (a common flaw), it is by far the quietest fan on the list when the fans are run at the same voltage (12V or less). Also, the 9 dB quieter level of the Silencer fan over a Panaflow FBA08A12L1A is sheer nonsense. The Panaflo is quieter. By a small margin, maybe just 1-2 dB, but it is definitely quieter, with several samples of each. No question.
<BR>
<BR>In many ways, the test data on this page is worse than no info at all.
[addsig]
Fan Misinformation
Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee
There are 2 things to consider in choosing fans:
o dB(A) quoted is in free air, no static resistance
o dB(A) quoted is different when fitted to a substrate (case)
Thus dB(A) is just one criterion, the other is:
o Thrust noise - when a fan sucks or blows against a resistance, the smaller bearing area at the thrust faces can result in noise (& wear)
---- ball-bearing suffer less from this
---- fluid, hydro, whatever-we-call-it-this-week suffer a bit less
o Vibration noise - PAPST-80mm vibrate more than NMB Panaflo PAPST-92mm
The other huge criterion is to do with where dB(A) fails:
o dB(A) is the weighting supposedly modelling human noise perception
o Unfortunately it is merely passable as opposed to great
---- it was meant for a very wide range of noise re health & safety
---- it was not meant for the intra-fan size comparisons it get used for
There are new (indeed several) ISO standards to cover this.
Torque variation in fixing the higher airflow fans to a substrate will
result in a 3dB(A) variation in noise - it really is quite critical. Yet the
even greater variation will be in the quality of the noise heard as the
substrate (case) itself has harmonics & resonant frequencies of its
own which can be excited into the equation by the fan's operation.
--
Dorothy Bradbury
http://homepage.ntlworld.com/dorothy.br ... anaflo.htm (Direct Prices)
o dB(A) quoted is in free air, no static resistance
o dB(A) quoted is different when fitted to a substrate (case)
Thus dB(A) is just one criterion, the other is:
o Thrust noise - when a fan sucks or blows against a resistance, the smaller bearing area at the thrust faces can result in noise (& wear)
---- ball-bearing suffer less from this
---- fluid, hydro, whatever-we-call-it-this-week suffer a bit less
o Vibration noise - PAPST-80mm vibrate more than NMB Panaflo PAPST-92mm
The other huge criterion is to do with where dB(A) fails:
o dB(A) is the weighting supposedly modelling human noise perception
o Unfortunately it is merely passable as opposed to great
---- it was meant for a very wide range of noise re health & safety
---- it was not meant for the intra-fan size comparisons it get used for
There are new (indeed several) ISO standards to cover this.
Torque variation in fixing the higher airflow fans to a substrate will
result in a 3dB(A) variation in noise - it really is quite critical. Yet the
even greater variation will be in the quality of the noise heard as the
substrate (case) itself has harmonics & resonant frequencies of its
own which can be excited into the equation by the fan's operation.
--
Dorothy Bradbury
http://homepage.ntlworld.com/dorothy.br ... anaflo.htm (Direct Prices)
-
dukla2000
- *Lifetime Patron*
- Posts: 1465
- Joined: Sun Mar 09, 2003 12:27 pm
- Location: Reading.England.EU
Exactly - I have this brilliant 80mm fan - 0mA, 0dB(A), unfortunately 0 cfm because I removed the hubMikeC wrote:2) No airflow info! Without this, the noise info is only partly useful.
. It is a shame the main conclusion is a quietest fan - perhaps I should send them mine for a recommendation (after I patent it).Something that has been bugging me: are (our bread and butter 12V axial) fans designed to blow or to suck? i.e. is there any difference in performance (primarily noise but potentially other) if the static pressure is on the intake side of the fan compared to on the exhaust side of the fan?jafb2000 wrote:Thrust noise - when a fan sucks or blows against a resistance
DC Axial fans be used to either blow or suck.
o For sucking, all fans are the same
o For blowing onto a hot-source (heatsink/component/area) then
the Panaflo are actually the best fan you can get for that usage
---- they exhaust air in a parallel path (tunnel vortex)
---- other fans exhaust air in a tangential path
For 1U cases the difference becomes more pronounced:
o If Panaflo used, you would not locate the hotest components close to
the sides of the fan exhaust because it is relatively dead air space
---- instead hotest components go in the direct air stream they produce
o If Panaflo used, you achieve a highly directional momentum to airflow,
so it helps to have vents similarly orientated to maintain this airflow
---- this is impossible for 1U unfortunately, but easy on 2U cases
However, no 1U cases will use 40x40x20mm Panaflo re cost - they
are ~10ukp in Qty 1,000 and some 1U cases use 6-7 - adds up fast.
In 2U applications, where there are few hot sources but a requirement
for directional thro airflow Panaflo provide better cooling than other fans.
o Cooling a particular hot-object = Panaflo best
o Cooling a general enclosure = Fans all the same
o Sucking air out of an enclosure = Fans all the same
Which is best for a computer?
o Fill all exhaust holes first
o Then try to use a fan to get cool-air-to or hot-air-from the CPU-cooler
o Then intake fans if required to cool specific items (eg, hard drives
o Then intake fans if available
---- remember that they act serially with exhaust fans
---- so you don't get additive cfm, but an improvement on existing cfm
---- serial fans increase the net static pressure capabaility beyond one
Re noise:
o Panaflo give less noise blowing on an object that other fans
---- however vary the fan-to-object distance re reflected pressure waves
---- these can result in unnecessary noise which is easily corrected
o In cramped enclosures, extracting air is best
---- because here, the blowing air into & thro them with objects close
to the exhaust of a fan will result in excessive airflow noise
---- by comparison, sucking air out of that enclosure will be less noisy
Hence you will find the cramped PSUs like 1U & 2U use air-exhaust.
When they use a dual intake fan, if the intake is high airflow/speed,
you will generally hear significant airflow noise. However since there
are 2 fans acting serially this noise is less than if just 1 intake fan used.
Yes, positive pressure (intake-only) enclosures can reduce dust, but in
reality they rarely do - because the dead-space in cases causes dust to
drop out of air-suspension. Filters destroy cfm numbers from 50-100%
with 100% quickly re clogging - axial fans have low static pressure
capability unless they are 90mm deep and several hundred cfm. Most
filters will only stop fan-dusting despite their resistance, so the case
will still with the finer dust just the same after a few months.
Stick the filter on the vacuum cleaner, not your cooling unit
Finally, few cases have air-intake holes as big as air exhaust.
--
Dorothy Bradbury
http://homepage.ntlworld.com/dorothy.br ... anaflo.htm (Direct Prices)
o For sucking, all fans are the same
o For blowing onto a hot-source (heatsink/component/area) then
the Panaflo are actually the best fan you can get for that usage
---- they exhaust air in a parallel path (tunnel vortex)
---- other fans exhaust air in a tangential path
For 1U cases the difference becomes more pronounced:
o If Panaflo used, you would not locate the hotest components close to
the sides of the fan exhaust because it is relatively dead air space
---- instead hotest components go in the direct air stream they produce
o If Panaflo used, you achieve a highly directional momentum to airflow,
so it helps to have vents similarly orientated to maintain this airflow
---- this is impossible for 1U unfortunately, but easy on 2U cases
However, no 1U cases will use 40x40x20mm Panaflo re cost - they
are ~10ukp in Qty 1,000 and some 1U cases use 6-7 - adds up fast.
In 2U applications, where there are few hot sources but a requirement
for directional thro airflow Panaflo provide better cooling than other fans.
o Cooling a particular hot-object = Panaflo best
o Cooling a general enclosure = Fans all the same
o Sucking air out of an enclosure = Fans all the same
Which is best for a computer?
o Fill all exhaust holes first
o Then try to use a fan to get cool-air-to or hot-air-from the CPU-cooler
o Then intake fans if required to cool specific items (eg, hard drives
o Then intake fans if available
---- remember that they act serially with exhaust fans
---- so you don't get additive cfm, but an improvement on existing cfm
---- serial fans increase the net static pressure capabaility beyond one
Re noise:
o Panaflo give less noise blowing on an object that other fans
---- however vary the fan-to-object distance re reflected pressure waves
---- these can result in unnecessary noise which is easily corrected
o In cramped enclosures, extracting air is best
---- because here, the blowing air into & thro them with objects close
to the exhaust of a fan will result in excessive airflow noise
---- by comparison, sucking air out of that enclosure will be less noisy
Hence you will find the cramped PSUs like 1U & 2U use air-exhaust.
When they use a dual intake fan, if the intake is high airflow/speed,
you will generally hear significant airflow noise. However since there
are 2 fans acting serially this noise is less than if just 1 intake fan used.
Yes, positive pressure (intake-only) enclosures can reduce dust, but in
reality they rarely do - because the dead-space in cases causes dust to
drop out of air-suspension. Filters destroy cfm numbers from 50-100%
with 100% quickly re clogging - axial fans have low static pressure
capability unless they are 90mm deep and several hundred cfm. Most
filters will only stop fan-dusting despite their resistance, so the case
will still with the finer dust just the same after a few months.
Stick the filter on the vacuum cleaner, not your cooling unit
Finally, few cases have air-intake holes as big as air exhaust.
--
Dorothy Bradbury
http://homepage.ntlworld.com/dorothy.br ... anaflo.htm (Direct Prices)
Action, reaction
I've read that Directron tutorial, and I completely agree that it is flawed. And while several people have pointed it out, many more seem to accept it as some enlightened message from the computer gods and take heed. I'm SURE that Directron has sold many more of their Silencer fans because of it, but I'm not entirely convinced that it was intentionally misleading.
And so I propose this: Send Directron a study done by silentPCreview. Something that compares several models of fans -- probably ones that Directron sells -- and gives honest answers and comparisons. In fact, you might get them to dish out some free samples for this; I know that they pay people who write tutorials, so just take payment for this article in product samples. Then, add on to the comparison a little description about what constitutes good fans -- not just raw dB at ambient temperature but also cfm, and other factors.... even a spectrum annalysis of the noise that the fans produce! And then, of course, add a link to the website. That way, you correct the problem of misinformation, create a smarter public, build good relations with a company that is at least trying to carry quiet products, potentially gain some free fans, and generate hits on the website!
P.S. I'm currently double-majoring in marketing and psychology.... can you tell?
And so I propose this: Send Directron a study done by silentPCreview. Something that compares several models of fans -- probably ones that Directron sells -- and gives honest answers and comparisons. In fact, you might get them to dish out some free samples for this; I know that they pay people who write tutorials, so just take payment for this article in product samples. Then, add on to the comparison a little description about what constitutes good fans -- not just raw dB at ambient temperature but also cfm, and other factors.... even a spectrum annalysis of the noise that the fans produce! And then, of course, add a link to the website. That way, you correct the problem of misinformation, create a smarter public, build good relations with a company that is at least trying to carry quiet products, potentially gain some free fans, and generate hits on the website!
P.S. I'm currently double-majoring in marketing and psychology.... can you tell?
The problem with spectrum analysis of fan noise, amongst any
attempt to define fan noise, is does it match how we perceive noise?
The answer is, no.
Human Perception of noise is a problem for comparison.
There are new standards for measurement of noise which will have
direct application to fan noise. Fitment is an issue (overtighten and
you can put 3-4dB(A) on a fan noise, fit further from a heatsink and
you can get lower noise for no change in temperature.
Basically quality matters, and you get what you pay for.
Any who have compared an NMB dual-ball-bearing fan to the
average generic no-name fans will know what I mean. Any who
have run fans at very high or very low temperatures will know
that only the top fans last - others have lube then bearing failure.
Using the Canadian winter temps or even car/industrial temp range on
generic fans can see many "Gee Whiz" CPU-coolers failing in 5 months.
Something to note for those fitting PCs in cars re CPU cooler quality,
either go fanless (Via C3) or ensure a good quality fan is fitted.
Case & PSU fans are the most critical re quality - because they
are outside the scope of any soundproofing inside the case:-
Noise / Case-Soundproofing / Case-Fans / Ears
Which is why many find case soundproofing doesn't do much.
It does, stick your head inside the case with everything off and
notice all room noise is simply absent whilst inside. Eerie when
you have a 48U rack-cabinet flight-case fully soundproofed.
Best 120mm = PAPST 4312MV or NMB B10 - 21dB(A) & 27dB(A)
Best 80mm = NMB B10 or Panaflo L1A or the rare slotted Panaflo
Best 40mm = Panaflo 40mm without any competitor, bonkers price
B19 = tacho fitted re NMB part numbers I assume everyone knows.
Best 50mm is actually the Delta 50x50x10mm used on QuietPC coolers.
Beats the Panaflo just a bit. The 60mm Panaflo isn't remarkable, due to
the huge fluid bearing forcing a big hub & so considerably higher rpm.
If you are spending 17-20ukp on a 120mm fan, it is pointless
to buy anything but the PAPST 4312MV re quality/longevity/noise.
--
Dorothy Bradbury
attempt to define fan noise, is does it match how we perceive noise?
The answer is, no.
Human Perception of noise is a problem for comparison.
There are new standards for measurement of noise which will have
direct application to fan noise. Fitment is an issue (overtighten and
you can put 3-4dB(A) on a fan noise, fit further from a heatsink and
you can get lower noise for no change in temperature.
Basically quality matters, and you get what you pay for.
Any who have compared an NMB dual-ball-bearing fan to the
average generic no-name fans will know what I mean. Any who
have run fans at very high or very low temperatures will know
that only the top fans last - others have lube then bearing failure.
Using the Canadian winter temps or even car/industrial temp range on
generic fans can see many "Gee Whiz" CPU-coolers failing in 5 months.
Something to note for those fitting PCs in cars re CPU cooler quality,
either go fanless (Via C3) or ensure a good quality fan is fitted.
Case & PSU fans are the most critical re quality - because they
are outside the scope of any soundproofing inside the case:-
Noise / Case-Soundproofing / Case-Fans / Ears
Which is why many find case soundproofing doesn't do much.
It does, stick your head inside the case with everything off and
notice all room noise is simply absent whilst inside. Eerie when
you have a 48U rack-cabinet flight-case fully soundproofed.
Best 120mm = PAPST 4312MV or NMB B10 - 21dB(A) & 27dB(A)
Best 80mm = NMB B10 or Panaflo L1A or the rare slotted Panaflo
Best 40mm = Panaflo 40mm without any competitor, bonkers price
B19 = tacho fitted re NMB part numbers I assume everyone knows.
Best 50mm is actually the Delta 50x50x10mm used on QuietPC coolers.
Beats the Panaflo just a bit. The 60mm Panaflo isn't remarkable, due to
the huge fluid bearing forcing a big hub & so considerably higher rpm.
If you are spending 17-20ukp on a 120mm fan, it is pointless
to buy anything but the PAPST 4312MV re quality/longevity/noise.
--
Dorothy Bradbury
-
MikeC
- Site Admin
- Posts: 12285
- Joined: Sun Aug 11, 2002 3:26 pm
- Location: Vancouver, BC, Canada
- Contact:
Re: Action, reaction
That's a single major as they're one and the same, by and large, aren't they?bengoerz wrote:P.S. I'm currently double-majoring in marketing and psychology.... can you tell?
-
marc999
- *Lifetime Patron*
- Posts: 230
- Joined: Sat Oct 26, 2002 4:05 pm
- Location: Calgary, Alberta, Canada
Jafb2000 wrote:
Also, you mentioned the "rare slotted Panaflo". Do you have a link or anything that has some info on this fan because I've never heard of it either and now I'm curious.
By the way, thanks for your posts. They are very informative.
Mike C,
Good job on analyzing the Directron test. I saw that a few weeks ago and found myself wondering who was right: Directron or Mike C. I should have known it was Mike C !!
I haven't heard of the Papst 4312MV so I decided to check it out. I guess it's a thermally controlled fan. I don't really understand how these work. Do they get the temperature from the motherboard or do they have their own temperature sensor? I assume it most have its own thermistor, it intakes 12V since it can't manipulate the PSU voltage it is receiving. Anyway, so how does one "lock-in" the Papst 4312MV to its bottom end 40CFM/22dB(A)/1150 RPM ?Best 120mm = PAPST 4312MV or NMB B10 - 21dB(A) & 27dB(A)
Best 80mm = NMB B10 or Panaflo L1A or the rare slotted Panaflo
Also, you mentioned the "rare slotted Panaflo". Do you have a link or anything that has some info on this fan because I've never heard of it either and now I'm curious.
By the way, thanks for your posts. They are very informative.
Mike C,
Good job on analyzing the Directron test. I saw that a few weeks ago and found myself wondering who was right: Directron or Mike C. I should have known it was Mike C !!
Looks like my post never made it.
o 4312MV is temperature controlled by a thermistor
---- where you put the thermistor is up to you
o If you truly only wanted it to run at the lowest level
---- I'd simply buy the 4312L and run it at 7V or 8.7V
---- 8.7V is via the 12V/3.3V voltage differential
I don't care for the 4412 line, it's poor versus the NMB 120mm.
Amusingly you can run a NMB B50 (high cfm) fan and simply
not really notice the noise since it's harmonics are so limited.
Marked contrast to many high flow fans & lousy bearings.
Rare slotted Panaflo...
o Panaflo with slots in the housing
o So reducing turbulence between rotor tip & housing
o Also reducing dead space in at the sides of the fan
Now, it's a nice idea, but to really use it requires very different design
of a computer or any enclosure for that matter. Without that you are
simply using it like any other Axial fan. Basically it gets closer to the
submarine style unshrouded prop - with some acoustic benefits.
Price wise, I'd stick with the standard fans - themselves excellent.
Thanks for the kind comments.
--
Dorothy Bradbury
www.stores.ebay.co.uk/panaflofan (Ebay)
http://homepage.ntlworld.com/dorothy.br ... anaflo.htm (Direct Prices)
o 4312MV is temperature controlled by a thermistor
---- where you put the thermistor is up to you
o If you truly only wanted it to run at the lowest level
---- I'd simply buy the 4312L and run it at 7V or 8.7V
---- 8.7V is via the 12V/3.3V voltage differential
I don't care for the 4412 line, it's poor versus the NMB 120mm.
Amusingly you can run a NMB B50 (high cfm) fan and simply
not really notice the noise since it's harmonics are so limited.
Marked contrast to many high flow fans & lousy bearings.
Rare slotted Panaflo...
o Panaflo with slots in the housing
o So reducing turbulence between rotor tip & housing
o Also reducing dead space in at the sides of the fan
Now, it's a nice idea, but to really use it requires very different design
of a computer or any enclosure for that matter. Without that you are
simply using it like any other Axial fan. Basically it gets closer to the
submarine style unshrouded prop - with some acoustic benefits.
Price wise, I'd stick with the standard fans - themselves excellent.
Thanks for the kind comments.
--
Dorothy Bradbury
www.stores.ebay.co.uk/panaflofan (Ebay)
http://homepage.ntlworld.com/dorothy.br ... anaflo.htm (Direct Prices)
-
Bluefront
- *Lifetime Patron*
- Posts: 5316
- Joined: Sat Jan 18, 2003 2:19 pm
- Location: St Louis (county) Missouri USA
jafb2000(Dorthy Bradbury)......your conclusions about filters are somewhat mis-leading. If properly designed, they do keep the case much cleaner. The standard computer filter is a flat piece of some material placed right next to the fan blades. No wonder the cfm is affected adversly. If you use a good filter material with sufficient surface area from which the fan can draw air, the cfm will be much less affected.
The problem here is few cases can accomodate such a filter setup. You have to DIY if you want good filtering with good airflow....
The problem here is few cases can accomodate such a filter setup. You have to DIY if you want good filtering with good airflow....
Dorothy Bradbury wrote...
> Filters destroy cfm numbers from 50-100% with 100% quickly
> re clogging - axial fans have low static pressure capability
> unless they are 90mm deep and several hundred cfm. Most
> filters will only stop fan-dusting despite their resistance,
> so the case will still with the finer dust just the same after
> a few months.
Bluefront wrote...
> jafb2000(Dorthy Bradbury)......your conclusions about filters
> are somewhat mis-leading. If properly designed, they do keep
> the case much cleaner. The standard computer filter is a flat
> piece of some material placed right next to the fan blades.
> No wonder the cfm is affected adversly. If you use a good filter
> material with sufficient surface area from which the fan can draw
> air, the cfm will be much less affected.
> The problem here is few cases can accomodate such a filter setup.
> You have to DIY if you want good filtering with good airflow....
There is nothing miss-leading about my post in relation to context.
PC filters are very poor:
o Size is limited to fan form-factor
---- ie, 40-120mm fans have 40-120mm filter units
o Filter profile is generally low reducing media capacity
---- filter media is compressed between 2 structures
---- permanent fan-fixed backing, media & front clip-on grill
---- Media of 5mm is compressed to typically 2mm
o Filter material is often plastic
---- plastic grills & backing are large losing 35% of free area
---- even before filter resistance is involved, the grill has some
---- the rarer all-metal grills lose less area to backing & grill
o Filter media is generally of 2 types
---- OEM - open-cell foam, higher particulate capacity
---- retail - fibrous glass-fibre matting, lower particulate capacity
If lucky, the filter media is blue so when it is white it needs replacing.
USA may use different colour coding, but the same principle applies.
The general guideline for filter cfm impact due to Pa resistance is:
o 50% reduction in cfm with new filter
o rising to 70% within a very short period of time
o 90% reduction in cfm will occur without obvious user knowledge
--- exception being higher monitored temperatures inside a PC
--- however areas the fan was directly cooling can be more affected
Fan filters are not suitable for 40-120mm DC Axial fans:
o Their static pressure capability is too low - 0.2-2.2mm water
--- to put that in perspective humans suck drinks up 250mm straws
o If filters are used, 25mm fans can not be used
--- generally 38mm depth fans are used which increases noise
o If filters are used, a roughly 200% upgrading in cfm is required
--- a 21cfm fan may be 21dB(A), but a 42cfm fan can be 36-38dB(A)
So maintenance of airflow involves a penalty in fan noise, or
alternatively, a penalty in a larger number of fans to compensate.
In industry, correct, a different approach is used:
o Blower fans offer far higher static pressure
--- these offer 10-40mm static pressure in small packages
--- but do so at higher noise levels per cfm than Axial fans
o Much deeper Axial fans are used
--- typically 58, 61 & 89mm depth
--- this is true of USA Air Conditioning/Furnance systems
--- high static pressure = deeper fan blade depth to compensate
o Filter design is very different
--- filters rely on surface area to minimise the pressure drop
------ for example, 24x24x2" is a common industrial size
--- the media is corrugated to offer high area & rigidity
------ less backing wire is required - usually thinner c/section
------ less air resistance from a larger filter area
For custom PC use the following works reasonably well:
o An inlet array of 9 21dB(A) 25cfm fans gives 225cfm
--- it is assumed a similar array is on the outlet (*)
--- thus the fans act serially to maintain the intake cfm
o An 18x18x1" industrial filter for "clean blades" is added
--- this needs to be set away from the fan inlet array by 1-2"
----- this ensures 18x18" surface area offers to fan inlets
----- rather than 9 circles of 80mm surface area instead
o The net result is roughly a halving of cfm to 100cfm
--- by manometer pressure test, or milar volume bag
--- including tower-like enclosure resistance
300cfm will cool 1500W, 100cfm will cool 500W.
So basically it pushes a PC into requiring a lot of fans.
(*) An obvious problem here is that the cfm of the exhaust is
higher than the inlet due to inlet filter resistance. This is
no longer a positive pressure design, negative pressure in the
enclosure is likely to exist pulling dust in thro every hole.
A positive pressure enclosure is required when filtration is
desired, to ensure that the primary air inlet IS via the filter.
One could dispense with the exhaust fans and use 38mm intake
fans, which would probably push the per-fan dB(A) up to 23dB(A).
Rack cabinets offer an easier way out:
o Seal the case
o Fit 2-9 550cfm Comair Rotron 254x89mm Axial fans
o Turn the door as a large filter panel
This is done for Telco & Industrial areas where dust is a problem.
There are custom designs which create far better solutions of course.
In industry there are usually 2 main grades of filtration:
o Keep fan blades & heatsinks generally clean of dust aggregation
o Actually filter to a specific particulate size (to 99-99.99% etc)
The former does still allow a degree of very fine dust in the machine,
so for asthmatics or allergy suffers it is likely insufficient.
The humble "PC filter" unfortunately doesn't work well:
o PC fan cfm drops very rapidly with any pressure resistance at all
o Media has low dust capacity (clogs quickly) & filters poorly
---- it does prevent blade/heatsink dust aggregation
o Media should be replaced versus cleaned
---- cleaning is very ineffective at removing particles
Industry re-locates the filtration to the A/C units:
o Electrostatic air-filters, Large panel air-filters
Reason being ease of replacement, efficiency and re-location of noise.
Increasingly a reason is A/C is required to offset thermally overloaded
racks - which often require "Ambient" to never exceed 30oC or less.
Wooden floors are useful at minimising the main PC dust build-up,
however they prevent a hard surface which will not absorb PC noise.
--
Dorothy Bradbury
www.stores.ebay.co.uk/panaflofan (Ebay)
http://homepage.ntlworld.com/dorothy.br ... anaflo.htm (Direct Prices)
> Filters destroy cfm numbers from 50-100% with 100% quickly
> re clogging - axial fans have low static pressure capability
> unless they are 90mm deep and several hundred cfm. Most
> filters will only stop fan-dusting despite their resistance,
> so the case will still with the finer dust just the same after
> a few months.
Bluefront wrote...
> jafb2000(Dorthy Bradbury)......your conclusions about filters
> are somewhat mis-leading. If properly designed, they do keep
> the case much cleaner. The standard computer filter is a flat
> piece of some material placed right next to the fan blades.
> No wonder the cfm is affected adversly. If you use a good filter
> material with sufficient surface area from which the fan can draw
> air, the cfm will be much less affected.
> The problem here is few cases can accomodate such a filter setup.
> You have to DIY if you want good filtering with good airflow....
There is nothing miss-leading about my post in relation to context.
PC filters are very poor:
o Size is limited to fan form-factor
---- ie, 40-120mm fans have 40-120mm filter units
o Filter profile is generally low reducing media capacity
---- filter media is compressed between 2 structures
---- permanent fan-fixed backing, media & front clip-on grill
---- Media of 5mm is compressed to typically 2mm
o Filter material is often plastic
---- plastic grills & backing are large losing 35% of free area
---- even before filter resistance is involved, the grill has some
---- the rarer all-metal grills lose less area to backing & grill
o Filter media is generally of 2 types
---- OEM - open-cell foam, higher particulate capacity
---- retail - fibrous glass-fibre matting, lower particulate capacity
If lucky, the filter media is blue so when it is white it needs replacing.
USA may use different colour coding, but the same principle applies.
The general guideline for filter cfm impact due to Pa resistance is:
o 50% reduction in cfm with new filter
o rising to 70% within a very short period of time
o 90% reduction in cfm will occur without obvious user knowledge
--- exception being higher monitored temperatures inside a PC
--- however areas the fan was directly cooling can be more affected
Fan filters are not suitable for 40-120mm DC Axial fans:
o Their static pressure capability is too low - 0.2-2.2mm water
--- to put that in perspective humans suck drinks up 250mm straws
o If filters are used, 25mm fans can not be used
--- generally 38mm depth fans are used which increases noise
o If filters are used, a roughly 200% upgrading in cfm is required
--- a 21cfm fan may be 21dB(A), but a 42cfm fan can be 36-38dB(A)
So maintenance of airflow involves a penalty in fan noise, or
alternatively, a penalty in a larger number of fans to compensate.
In industry, correct, a different approach is used:
o Blower fans offer far higher static pressure
--- these offer 10-40mm static pressure in small packages
--- but do so at higher noise levels per cfm than Axial fans
o Much deeper Axial fans are used
--- typically 58, 61 & 89mm depth
--- this is true of USA Air Conditioning/Furnance systems
--- high static pressure = deeper fan blade depth to compensate
o Filter design is very different
--- filters rely on surface area to minimise the pressure drop
------ for example, 24x24x2" is a common industrial size
--- the media is corrugated to offer high area & rigidity
------ less backing wire is required - usually thinner c/section
------ less air resistance from a larger filter area
For custom PC use the following works reasonably well:
o An inlet array of 9 21dB(A) 25cfm fans gives 225cfm
--- it is assumed a similar array is on the outlet (*)
--- thus the fans act serially to maintain the intake cfm
o An 18x18x1" industrial filter for "clean blades" is added
--- this needs to be set away from the fan inlet array by 1-2"
----- this ensures 18x18" surface area offers to fan inlets
----- rather than 9 circles of 80mm surface area instead
o The net result is roughly a halving of cfm to 100cfm
--- by manometer pressure test, or milar volume bag
--- including tower-like enclosure resistance
300cfm will cool 1500W, 100cfm will cool 500W.
So basically it pushes a PC into requiring a lot of fans.
(*) An obvious problem here is that the cfm of the exhaust is
higher than the inlet due to inlet filter resistance. This is
no longer a positive pressure design, negative pressure in the
enclosure is likely to exist pulling dust in thro every hole.
A positive pressure enclosure is required when filtration is
desired, to ensure that the primary air inlet IS via the filter.
One could dispense with the exhaust fans and use 38mm intake
fans, which would probably push the per-fan dB(A) up to 23dB(A).
Rack cabinets offer an easier way out:
o Seal the case
o Fit 2-9 550cfm Comair Rotron 254x89mm Axial fans
o Turn the door as a large filter panel
This is done for Telco & Industrial areas where dust is a problem.
There are custom designs which create far better solutions of course.
In industry there are usually 2 main grades of filtration:
o Keep fan blades & heatsinks generally clean of dust aggregation
o Actually filter to a specific particulate size (to 99-99.99% etc)
The former does still allow a degree of very fine dust in the machine,
so for asthmatics or allergy suffers it is likely insufficient.
The humble "PC filter" unfortunately doesn't work well:
o PC fan cfm drops very rapidly with any pressure resistance at all
o Media has low dust capacity (clogs quickly) & filters poorly
---- it does prevent blade/heatsink dust aggregation
o Media should be replaced versus cleaned
---- cleaning is very ineffective at removing particles
Industry re-locates the filtration to the A/C units:
o Electrostatic air-filters, Large panel air-filters
Reason being ease of replacement, efficiency and re-location of noise.
Increasingly a reason is A/C is required to offset thermally overloaded
racks - which often require "Ambient" to never exceed 30oC or less.
Wooden floors are useful at minimising the main PC dust build-up,
however they prevent a hard surface which will not absorb PC noise.
--
Dorothy Bradbury
www.stores.ebay.co.uk/panaflofan (Ebay)
http://homepage.ntlworld.com/dorothy.br ... anaflo.htm (Direct Prices)
-
ez2remember
- Friend of SPCR
- Posts: 809
- Joined: Thu Nov 28, 2002 5:07 pm
- Location: London, UK
Agreed. A filter that does its job of filtering dust out will block of most of the air. Also think about how small dust particles are, that means the weaves would have to have tiny holes to do any filtering and blocking a lot of the airflow.
It's more of a cleaning job, cleaning the filters every single day to get them to perform at it's best. I vacum my PC once every few month and is much easier.
It's more of a cleaning job, cleaning the filters every single day to get them to perform at it's best. I vacum my PC once every few month and is much easier.
-
Bluefront
- *Lifetime Patron*
- Posts: 5316
- Joined: Sat Jan 18, 2003 2:19 pm
- Location: St Louis (county) Missouri USA
What's so hard about making this filter.....almost no temp hit because of the surface area of the filter material. Used a cut-up piece of furnace filter.
Furnace filter would be quite good.
A simple litmus test for a filter:
o Get a cardboard box & tape it up fully
o Fit your fan(s) at one end set to exhaust
o Fit your filter at the other end
Note the airflow drop by media on/off, so you get a good
real-use idea of how many fans and /how big/ a filter to use.
--
Dorothy Bradbury
A simple litmus test for a filter:
o Get a cardboard box & tape it up fully
o Fit your fan(s) at one end set to exhaust
o Fit your filter at the other end
Note the airflow drop by media on/off, so you get a good
real-use idea of how many fans and /how big/ a filter to use.
--
Dorothy Bradbury
-
Bluefront
- *Lifetime Patron*
- Posts: 5316
- Joined: Sat Jan 18, 2003 2:19 pm
- Location: St Louis (county) Missouri USA
All the computers I build for myself have intake air filtration. The design varies by the case. I also buy cases with the filter setup in mind....some cases are a mess in this regard.
I am a firm believer in testing for temperatures, and fan speeds, with and without the filters in place. I am using a variety of measurement hardware, but usually rely on a DigitalDoc5 with eight temp probes. I also usually use temp controlled fans.
If your temps go up, and the fan speeds increase when you attach a filter...you've got a problem and it's usually surface area of the filter. You need all the surface area you can get. And if you have a fast, powerful fan, you need even more area. I am willing to take a slight temp hit in order to have a good filter, but I usually find a temp increase of less than 2c, and/or rpm rise of maybe 10% or less with my setups.....
Check the size of this filter in my latest setup(Xterra in-cabin filter)
I am a firm believer in testing for temperatures, and fan speeds, with and without the filters in place. I am using a variety of measurement hardware, but usually rely on a DigitalDoc5 with eight temp probes. I also usually use temp controlled fans.
If your temps go up, and the fan speeds increase when you attach a filter...you've got a problem and it's usually surface area of the filter. You need all the surface area you can get. And if you have a fast, powerful fan, you need even more area. I am willing to take a slight temp hit in order to have a good filter, but I usually find a temp increase of less than 2c, and/or rpm rise of maybe 10% or less with my setups.....
Check the size of this filter in my latest setup(Xterra in-cabin filter)
-
Bluefront
- *Lifetime Patron*
- Posts: 5316
- Joined: Sat Jan 18, 2003 2:19 pm
- Location: St Louis (county) Missouri USA
Well I've been using PCToys crystal max temp controlled fans. Found a bunch on sale one day at CompUSA...about $5 ea. They start out at low voltage and go up as the heat increases.
The best thing about this fan is the sensor....it's at the end of a long wire about 12" long. So you can attach the sensor say to the CPU, and have your case fan controlled directly by the CPU. I've never seen any other fan with a sensor wire this long. Most fans with a sensor have a sensor wire too short to even splice...
The best thing about this fan is the sensor....it's at the end of a long wire about 12" long. So you can attach the sensor say to the CPU, and have your case fan controlled directly by the CPU. I've never seen any other fan with a sensor wire this long. Most fans with a sensor have a sensor wire too short to even splice...