SPCR's Fan Round-Up #2: 120mm Fans
@MikeC: I suggest that SPCR should also test the fans (maybe just the best ones) when mounted on a HS or on a case - not everyone is going to cut the case grill(s).
Also, since the cfm is not directly measured (air speed is actually measured), variations between the fans in the 'open area' influence the result.
I noticed that the Nexus fans are more 'open' than others, or at least the transition from the 'circle' to the 'square' is smoother.
Also, since the cfm is not directly measured (air speed is actually measured), variations between the fans in the 'open area' influence the result.
I noticed that the Nexus fans are more 'open' than others, or at least the transition from the 'circle' to the 'square' is smoother.
-
- Posts: 1283
- Joined: Thu Aug 07, 2003 2:43 pm
- Location: Charlotte, NC, USA
There are many many different grill designs out there, from extremely restrictive to mesh, so they'd have to test them all. Open air results give the best baseline; any better cooling and you'll have to compensate by running a faster fan and/or cutting out the grill.Tzupy wrote:@MikeC: I suggest that SPCR should also test the fans (maybe just the best ones) when mounted on a HS or on a case - not everyone is going to cut the case grill(s).
A lot of work went into that, so thanks. However, I’m starting to question one crucial part of SPCR’s testing methodology:
If this is an inaccuracy, and it’s a uniform inaccuracy across all fans tested (+27% or whatever), then we don’t get an accurate cfm number, but still a number that’s relative to calculating each fans respective performance. If this inaccuracy isn’t uniform, there’s no common denominator to face fans off with. Does SPCR still maintain faith in this method of cfm testing? You guys know far more about the physics involved than I do - I’d have to do a lot of reading up, and then wouldn’t know half as much. My uneducated opinion is starting to prefer the Madshrimps airflow/cooling test. I think there’s one too many inaccuracies and variables in SPCR’s airflow evaluation. Not least, trying to find the fans ‘peak air speed’ (sweet spot?) with the anemometer...
Nexus D12SL-12, rated by Nexus at 36.87 cfm, tested by SPCR at 47 cfm = +27%[Editor's Note: This particular CFM measurement technique was adopted after several others were tried. The main reason for our choice is that the results are usually repeatable, and they are the closest to the CFM specifications provided for fans from manufacturers whose technical documentation appears reasonably trustworthy. All the other methods gave us considerably lower CFM numbers.]
If this is an inaccuracy, and it’s a uniform inaccuracy across all fans tested (+27% or whatever), then we don’t get an accurate cfm number, but still a number that’s relative to calculating each fans respective performance. If this inaccuracy isn’t uniform, there’s no common denominator to face fans off with. Does SPCR still maintain faith in this method of cfm testing? You guys know far more about the physics involved than I do - I’d have to do a lot of reading up, and then wouldn’t know half as much. My uneducated opinion is starting to prefer the Madshrimps airflow/cooling test. I think there’s one too many inaccuracies and variables in SPCR’s airflow evaluation. Not least, trying to find the fans ‘peak air speed’ (sweet spot?) with the anemometer...
...then multiplying that by the fans blade area. Surely some fans sweet spots are ‘sweeter’ or more disproportional to it’s overall output, than others?Even with these precautions, the position of the instrument relative to the fan can also affect the measurements, so measurements are taken slowly enough that time can be taken to find the peak air speed, which is then used for subsequent calculations.
-
- Site Admin
- Posts: 12285
- Joined: Sun Aug 11, 2002 3:26 pm
- Location: Vancouver, BC, Canada
- Contact:
Ah but the mfg's cfm numbers are not arrived at the same way. We generally get higher cfm readings, but not across the board, and at least part of the reason for any non-uniformity is that the mfgs don't all use the same measurement method... and maybe some of them exaggerate. Regarding the calculation required to get volumetric flow from airspeed -- there's no way to avoid such calculations; no anemometer we know of can measure volumetric flow directly.Sooty wrote:If this is an inaccuracy, and it’s a uniform inaccuracy across all fans tested (+27% or whatever), then we don’t get an accurate cfm number, but still a number that’s relative to calculating each fans respective performance. If this inaccuracy isn’t uniform, there’s no common denominator to face fans off with.
The problem with Madshrimp's method is that different heatsinks are optimized for different airflow levels. It's not just a matter of --- "oh this fan leads to a higher temp, it must blow less air". We have no idea how much less. With a high density fin pattern, very little difference would show up at low airflow; with a low density fin pattern, beyond a certain airflow, there'd be little difference in temperature. It's just a different variable the heatsink introduces, one that gives less direct information about airflow compared to an anemometer.
For the people who think why ARX is included ....
http://www.xbitlabs.com/articles/other/ ... su2_3.html
These In-Win-PSU's use them....
hehe, so now you know those won't be near-quiet
http://www.xbitlabs.com/articles/other/ ... su2_3.html
These In-Win-PSU's use them....
hehe, so now you know those won't be near-quiet
It always makes me wonder why some manufacturer pushing harder on fluid or other exotic bearings. It sounds like some of these are quieter and just as reliable as the old ball and sleeve types.(basically superceding the old tech) I'm sure price is the biggest consideration, but if there was a large scale focus on better bearings, it would just be better for the entire world of fans. Not to mention prices and costs tend to drop that way. A silentpcreview manufacturing company is ripe for the taking!jaganath wrote:It appears to be some kind of hydrodynamic bearing:What type of bearing is the Noctua "SSO Bearing"?
http://www.noctua.at/main.php?show=sso_bearing&lng=en
(small aside)
It's funny how some aspects of computer design have been stuck in 2nd gear forever now, while some march tirelessly. Input interfaces like keyboards/mice also seem to be pretty stuck, while cpus/video cards are pushed like no tomorrow. It feels like storage is heading toward multiple breakthroughs however, I'm suspecting it'll end up similar to displays. There's a large number of display choices now with different advantages/disadvantages. Future storage looks headed on a similar path once we move on from the Magnetic world. It's going to be fun and have a huge amount of growing pains, just like the transition away from CRT's.
great read as usual. now for part two be sure to get some fans from those ppl :
http://www.noiseblocker.de/en/produktmeta_luefter.php
and
http://www.glacialtech.com/product/standard_PC.htm
greetings !
http://www.noiseblocker.de/en/produktmeta_luefter.php
and
http://www.glacialtech.com/product/standard_PC.htm
greetings !
-
- *Lifetime Patron*
- Posts: 5316
- Joined: Sat Jan 18, 2003 2:19 pm
- Location: St Louis (county) Missouri USA
IMHO.....it's very obvious the problems you face when trying to rate/judge fans against each other. There are so many different brands, blade designs, bearing types, rpm and cfm ratings, etc, that the relatively few fans you can test will never satisfy everyone. And your testing methods will never satisfy everyone.
For me....I'm looking for a single fan that will blow enough air to keep my computer cool under load, that will run very quietly at idle conditions, that is not much affected by drawing through a filter, that has bearings that will last a long time, that is not affected by mounting position, that is not affected by PWM, that will start and keep running at a low voltage, that I don't have to import from Europe, etc.
The only way to find such a fan, if one even exists, is to read about every fan design out there, to read every review I can find, to ask questions on many different forums.......and finally be willing to buy a large number of different fans, and to try them all under the right conditions (for me). A daunting task to say the least.
I just received the 1600rpm S-Flex Scythe fan........maybe this one will be the right one. But I doubt it.
For me....I'm looking for a single fan that will blow enough air to keep my computer cool under load, that will run very quietly at idle conditions, that is not much affected by drawing through a filter, that has bearings that will last a long time, that is not affected by mounting position, that is not affected by PWM, that will start and keep running at a low voltage, that I don't have to import from Europe, etc.
The only way to find such a fan, if one even exists, is to read about every fan design out there, to read every review I can find, to ask questions on many different forums.......and finally be willing to buy a large number of different fans, and to try them all under the right conditions (for me). A daunting task to say the least.
I just received the 1600rpm S-Flex Scythe fan........maybe this one will be the right one. But I doubt it.
-
- *Lifetime Patron*
- Posts: 495
- Joined: Thu Apr 15, 2004 3:53 am
- Location: In Front of PC
- Contact:
I agree, it seems the Noctua fans do deliver in open air, but when paired with a heatsink which has tighter fins, its performance will drop, as the fan can't deliver a lot of air pressure.Tzupy wrote:@MikeC: I suggest that SPCR should also test the fans (maybe just the best ones) when mounted on a HS or on a case .
see: http://www.cooling-masters.com/articles-39-6.html
and
this only comes up when tested on a heatsink/case where pressure is needed to get air moved through the fins/case grill.
in my own test the Noctua 1200 paired with the GlobalWin NCB, which in turn was pretty much on par with the Nexus http://www.madshrimps.be/gotoartik.php?articID=516
would recommend to get your hands on the Sharkoon Silent Storm fans, they also made a good showing.
but you guys taking a different approach then us, only leads to more information on the tested fan, which is good in the end
Interesting graph, it looks like the Noiseblocker SX2 Pro is very good at coping with pressure.
I used it in my previous system, as back case fan in SLK3000B, turned down to ~6V at low load. It kept having high frequency bearing noise.
The 120 mm Nexii I have now in P150 have very little bearing noise.
I used it in my previous system, as back case fan in SLK3000B, turned down to ~6V at low load. It kept having high frequency bearing noise.
The 120 mm Nexii I have now in P150 have very little bearing noise.
-
- Posts: 2049
- Joined: Thu Dec 15, 2005 11:06 am
- Location: Klamath Falls, OR
JMKE, that's a heck of an interesting graph. The lower pressure and CFM of the lower RPM fans is (yawn) to be expected. But for the 1550RPM fan to soundly beat the 1600RPM fan in both pressure and CFM is... interesting. One suspects either a misprint on the "1550" or mendacity.
Or perhaps the 1550RPM fan has 150 fan blades instead of the usual 7?
Or perhaps the 1550RPM fan has 150 fan blades instead of the usual 7?
Here is my quick visual comparison between the SX2 Pro and the Nexus (I don't yet have a camera):
1) the motor hub on the Nexus is 40 mm diameter and 44 mm on the SX2 Pro (a bit more blade area for the Nexus).
2) the Nexus is B & W, while SX2 Pro is all black.
3) the Nexus has closed corners, while the SX2 Pro has open corners.
4) the Nexus blades are more curved and thinner than the SX2 Pro blades.
5) the 3 struts without wires are on the Nexus thinner than on the SX2 Pro.
6) the transition area between the 'circle' and the 'square' is larger on the Nexus.
1) the motor hub on the Nexus is 40 mm diameter and 44 mm on the SX2 Pro (a bit more blade area for the Nexus).
2) the Nexus is B & W, while SX2 Pro is all black.
3) the Nexus has closed corners, while the SX2 Pro has open corners.
4) the Nexus blades are more curved and thinner than the SX2 Pro blades.
5) the 3 struts without wires are on the Nexus thinner than on the SX2 Pro.
6) the transition area between the 'circle' and the 'square' is larger on the Nexus.
Is it just me or is that chart irrelevant? The Papst 4412/FGL has 400 rounds per minute more than the Noctua S12-1200 and the Noiseblocker SX2 Pro has 350 rounds per minute more. If we compare the Noctua S12-1200 to the Noctua S12-800 we can see that the Noctua scales well with higher RPM, the 1200 version performs better. So, furthermore, we can conclude that if the Noctua was at 1600-1550 RPM as well the performance would have been even better. For a completely fair comparison at least make sure that all the fans are at the same rounds per minute.
Yeah, that doesn't suprise me. Heard long ago from one of leaders of the time that the less space between the blades the better for high restrictive uses e.g. radiators/heatsinks, and that seems to be the case.jmke wrote:I agree, it seems the Noctua fans do deliver in open air, but when paired with a heatsink which has tighter fins, its performance will drop, as the fan can't deliver a lot of air pressure.
Anyway, funny how the article compares the the 120mm sticky thread I compiled from the forums years ago and every once in a while update. Finding that "perfect" fan has sure come a long way from scuttlebut to more scientific methods -- but seems to remain just as dynamic and elusive.
DrCR
_____________
p.s. Amazing to see so many new names around. Awesome to see the SPCR community continuing to grow.
Confirmed. Noctua fans have problems with heatsinks/grills. I got some noctua fans when I saw the great results in the tests and I've been testing them for a week in my systems.
My systems are similar, all of them in an antec sonata (classic, grill fan not removed), with phantom 500 PSU, ninja heatsink. There are only two fans in the system in a push/pull configuration, one in the case the other in the heatsink.
Before I installed them I was using papst 4412 F2GLL fans. So I can compare with these ones only. I usually undervolt the fans with speedfan to a point where I can't hear them anymore for idle (I do tests at night in a very silent environment in an acoustic isolated room). With papst this point is found at 35%, and with noctua at 25%. I don't know about rpms or equivalent voltages for the fans at those speedfan values.
Well, at these values the temperatures with the papst where much lower (about 6º-8º). Even more, with the noctua fans my phamtom 500 fans turned on, which is something I rarely have seen before, as the push-pull fan configuration I have avoids usually hot air going towards the PSU.
As I was very amazed by the bad results I turned up the rpms of the noctuas. I tried with 35%, 45%, 55% and the phantom fan always ended turning on. So these fans are not giving me good airflow. This was all for idle. For load the results were similar. I couldn¡t get enough airflow with the noctuas.
Conclusion: If you are willing to use noctua fans in your case, don't forget to remove any grills, and use with care when using a heatsink. I think that the noctuas don't give enough air pressure, and I also think that they disperse the air they flow more than a conventional fan. (I mean, with papst fans it seems that the air travels in a straight line perperndicular to the fan, however with noctua air travels all ways).
My systems are similar, all of them in an antec sonata (classic, grill fan not removed), with phantom 500 PSU, ninja heatsink. There are only two fans in the system in a push/pull configuration, one in the case the other in the heatsink.
Before I installed them I was using papst 4412 F2GLL fans. So I can compare with these ones only. I usually undervolt the fans with speedfan to a point where I can't hear them anymore for idle (I do tests at night in a very silent environment in an acoustic isolated room). With papst this point is found at 35%, and with noctua at 25%. I don't know about rpms or equivalent voltages for the fans at those speedfan values.
Well, at these values the temperatures with the papst where much lower (about 6º-8º). Even more, with the noctua fans my phamtom 500 fans turned on, which is something I rarely have seen before, as the push-pull fan configuration I have avoids usually hot air going towards the PSU.
As I was very amazed by the bad results I turned up the rpms of the noctuas. I tried with 35%, 45%, 55% and the phantom fan always ended turning on. So these fans are not giving me good airflow. This was all for idle. For load the results were similar. I couldn¡t get enough airflow with the noctuas.
Conclusion: If you are willing to use noctua fans in your case, don't forget to remove any grills, and use with care when using a heatsink. I think that the noctuas don't give enough air pressure, and I also think that they disperse the air they flow more than a conventional fan. (I mean, with papst fans it seems that the air travels in a straight line perperndicular to the fan, however with noctua air travels all ways).
-
- Moderator
- Posts: 7681
- Joined: Mon Dec 09, 2002 7:11 pm
- Location: Maynard, MA, Eaarth
- Contact:
-
- Site Admin
- Posts: 12285
- Joined: Sun Aug 11, 2002 3:26 pm
- Location: Vancouver, BC, Canada
- Contact:
kike_1974 --
Although you imply that the Noctuas are running more noisily because of the % speed used in Speedfan (which, as you say, is not that useful) you make no direct mention of acoustics: Do the Noctua fans actually make your system noisier?
It should be noted that...
...when the Phantom is used in a conventional ATX case like the Sonata, the back case fan tends to draw air through the PSU. This is, however, an extraordinarily high impedance situation, imo, as the Phantom has very small vent openings and is tightly packed inside. Most quiet optimized designs will not have such high impedance even pulling from the front vent... although the distance does matter. I would guess that with reduced pressure, the Noctuas pulled less air through the PSU, and were probably drawing largely from the "Antec" pattern holes on the sides of the case.
Also, an unmodified Sonata is a high impedance case by today's SPCR standards. Even when we originally reviewed it, we did not recommend the case highly; there are too many points of airflow impedance.
Although you imply that the Noctuas are running more noisily because of the % speed used in Speedfan (which, as you say, is not that useful) you make no direct mention of acoustics: Do the Noctua fans actually make your system noisier?
It should be noted that...
...when the Phantom is used in a conventional ATX case like the Sonata, the back case fan tends to draw air through the PSU. This is, however, an extraordinarily high impedance situation, imo, as the Phantom has very small vent openings and is tightly packed inside. Most quiet optimized designs will not have such high impedance even pulling from the front vent... although the distance does matter. I would guess that with reduced pressure, the Noctuas pulled less air through the PSU, and were probably drawing largely from the "Antec" pattern holes on the sides of the case.
Also, an unmodified Sonata is a high impedance case by today's SPCR standards. Even when we originally reviewed it, we did not recommend the case highly; there are too many points of airflow impedance.
Ah, forgot to note that in the sticky thread. Thanks for the reminder.jaganath wrote:Perhaps a postscript needs to be added to the review to reflect the Noctua's poor pressure capability? Something like "we recommend these fans as intake or case fans, but their poor pressure delivery means they should not be used in proximity to heatsinks or other sources of impedance".
-
- Posts: 2
- Joined: Wed Apr 09, 2003 2:56 pm
- Location: life cancerville, PA
- Contact: