My informal 120mm fan test...

Sunfox · Post by **Sunfox** » Sun Jan 14, 2007 6:50 pm

So, I'm building a new PC and I've bought a variety of fans to try out. But before I begin building I've decided to try them all out using a spare power supply so I can keep the swapping to a minimum. Below are my subjective impressions - I have no equipment to test this with, so I'm just going by what I hear and what I feel.

SilenX Iextrema 120mm (1200rpm, 38mm thick) - 11db my butt... this fan moves a LOT of air, but also has an annoying ticking AND buzzing sound. It's also remarkably flimsy - holding it I'm easily able to bend it so the blades hit the sides. I was hoping to use it for my P180's lower chamber, but that's not going to happen. (Tried 1 unit)

Antec Tricool 120mm (tested clear and black) - in Low (1200rpm)mode this fan is completely silent... but that's probably because it moves next to no air. In medium (1600rpm) mode it moves signifiantly more air, but there's also a fair bit of blade noise. High? Don't make me laugh! (Tried 4 units)

Cooler Master A12025-12CB-5BN-L1 (1220rpm) - Fairly quiet, moves a good amount of air, but has a bit of an annoying droning quality to the sound. Outclassed by some of the other fans here. (Tried 2 units)

Nexus RealSilent (1000rpm) - Moves less air than the Cooler Master, but has a nicer, more pleasant sound. No annoying ticks or buzzes like some other models. (Tried 2 units)

Aerocool Turbine (950rpm) - Surprisingly similar in performance to the Nexus. I'd say it gives a LITTLE bit more air pressure than the Nexus, but at a neglegable noise level increase. Overall tone is lower in pitch than the Nexus. (Tried 2 units)

Noctua NF-S12 (800rpm) - Wow, heavy. Moves a good amount of air for the RPM (less air than the Aerocool), but has a bit of a ticking sound when placed to my ear. With ULNA (500rpm) - similar performance to the Antec at low - no noise, no air. (Tried 2 units)

Noctua NF-S12 (1200rpm) - This moves a LOT of air. None of the annoying ticking sound of the 800rpm version, but there's a good amount of air noise. With ULNA (600rpm) - seems to perform about the same as the 800rpm with ULNA. Overall I would suggest the 1200rpm version over the 800rpm version. (Tried 2 units)

My impressions - I came out favoring the Aerocool Turbine and Noctua 1200rpm, but for different purposes. When silence is important, such as on the front of the P180 case, I plan on using the Aerocool (with the Nexus a close second). For the back of the case or the lower chamber where a little more noise can be tolerated in exchange for more airflow, I plan on trying out the Noctua 1200rpm.

Of course this may all change once I get them in the case. But for now I at least know that the Antec, Silenx and Cooler Master aren't likely to be better than the others.

kaange · Post by **kaange** » Sun Jan 14, 2007 7:45 pm

From my experience with the 2 NF-S12 1200rpm fans that I bought, the ticking you experienced with the 800rpm version may be a QA issue as one of my 1200rpm fans has the slight tick while the other does not. At 12V, they still have a noticable amount of airflow noise.

Strangely, I found the Antec Tricool that came with my case moves a reasonable amount of air at the 'low' setting but is noisier than the Noctuas which are moving a similar amount of air (free space - no obstructions). At 'medium', I consider the Tricool to be loud while at 'high', it is totally unacceptable except for a server.

Greg

Sunfox · Post by **Sunfox** » Sun Jan 14, 2007 7:52 pm

It's odd about the Noctuas, since I bought all four fans at the same time and both 800rpm units have the exact same click while both 1200rpm models don't have any.

I was surprised at the Antecs, since I'm using 2 of them in my current case on the assumption that they moved more air than the Nexus I have in the front of that case, but in free air at "low" I found them moving next to no air compared to some of the other fans - sort of like the Noctua 800rpm with the ULNA.

Post by **NeilBlanchard** » Sun Jan 14, 2007 7:58 pm

Greetings,

Maybe the [internal] speed control in the 800RPM Noctua's is a PWM, or it is working in a way to cause the motor to click?

[Edit: however they are slowing down the the fan, it seems to be causing the clicking.]

Sunfox · Post by **Sunfox** » Sun Jan 14, 2007 8:58 pm

They were plugged directly into a 5.25" drive plug... no speed control or voltage changes whatsoever.

kaange · Post by **kaange** » Sun Jan 14, 2007 9:55 pm

My guess is that the only difference between the 800rpm and 1200rpm version of the Noctua fans is that the slower speed version have added resistance. The lower rated amperage (0.05A vs 0.12A => 240ohms vs 100 ohm total resistance) seems to support this.

Of course they could fiddle with different windings on the fan motors but I would think for simplicity, they would just use the same motor but throttled back with some resistance (just as with the ULNA adapter).

I wouldn't be surprised if this was also the only difference with the Scythe SFlex D, E and F fans.

Greg

Felger Carbon · Post by **Felger Carbon** » Sun Jan 14, 2007 10:16 pm

kaange wrote:Of course they could fiddle with different windings on the fan motors but I would think for simplicity, they would just use the same motor but throttled back with some resistance (just as with the ULNA adapter). I wouldn't be surprised if this was also the only difference with the Scythe SFlex D, E and F fans.

If this was true, then the minimum RPMs of the D, E, F would all be the same. If different windings are used, then the minimum RPMs would differ. My bet's on different windings. Who wants to measure the minimum RPMs?

EndoSteel · Post by **EndoSteel** » Mon Jan 15, 2007 11:55 am

Felger Carbon
My bet's on different windings. Who wants to measure the minimum RPMs?
They're all the same due to Ð° stupid over current protection circuit. But I too think windings are different.

kaange · Post by **kaange** » Mon Jan 15, 2007 2:48 pm

Does anyone know the starting voltages and full power current of the different Scythe SFlex's? We could then calculate the power being consumed at the starting voltages and if they are different, then most likely resistance is used to alter the speeds - if the starting power is pretty much the same, then the windings are different. This assumes that the starting torque is identical, of course.

I thought over current protection would kick in due to the lack of back emf when the motor was stationary. Why would it be set for a minimum speed?

Greg

Felger Carbon · Post by **Felger Carbon** » Mon Jan 15, 2007 3:06 pm

kaange wrote: - if the starting power is pretty much the same, then the windings are different. This assumes that the starting torque is identical, of course.

Greg, let me say this politely: you are mistaken. You obviously have never designed transformers and seen them built, for instance. For a D vs an F, for instance, there will be many more turns of wire, of a smaller diameter, on the D. Therefore, the starting current will be much lower on the D. This is in the nature of motors and magnetic windings in general. No problem, you aren't an electrical/electronic engineer. I am (retired, though).

kaange · Post by **kaange** » Tue Jan 16, 2007 9:08 pm

Actually, I did train as an EE but work in IT but I was posting in a fairly casual way so let me work through your assumption in this post.

If the D has n x turns of the F (where n is a real number), then the magnetic field produced by the same current flow should induce a torque that is n x that of the F. So if the fans are stalled, then the D should start with 1/n of the current needed to start the F. But at this point, since the motors are not turning, it is purely the coil winding resistance that limits the current flow (as there is no back EMF being generated by the magnets on the windings). Now since there are n times the number of windings of F, there would be n times the coil resistance for the same gauge winding but the thinner gauge wire will make the D have more than n times the resistance of the F. This implies that the starting voltage of the D should be more than that of the F, doesn't it?

But let's assume that they use the same guage wire so that the coil resistance of the D is n times that of the F. Then the starting voltage should be identical.

Then again, they could have parallel windings (ie n is an integer) and then the coil resistance would be much less (1/n for the same gauge wire) so the D should start at a voltage of 1/(n*n) vs the F. I guess by using thinner gauge wire and parallel but fewer windings, you could end up with the starting voltage/speed/power draw variations that are seen across the range.

But in all cases, starting power (V*I) should be pretty similar.

Now if the magnet arrangement for the different fan motors was changed, then this alters everything but that hasn't been mentioned as a reason for speed differences of the Scythe Flex's

Please help me correct my analysis (it has been a long time since Uni)

Greg

Devonavar · Post by **Devonavar** » Tue Jan 16, 2007 9:23 pm

kaange wrote:Does anyone know the starting voltages and full power current of the different Scythe SFlex's?

You should be able to milk this data from SPCR's review.

In any case, I've been told that different fan models generally use different PWM frequencies rather than different windings or different resistors to modulate fan speed â€” so that's another avenue to pursue.

Incidentally, my work with SPCR's fan round-ups has left me with the same impression as arrived at in this thread: The very slowest fans sometimes suffer from clicking that is not present in higher speed models. This was true of our Noctua Samples; to a lesser extent it was true of the S-Flexes, and it was also true of some others I looked at (some NMB & some Coolink IIRC).

Sunfox · Post by **Sunfox** » Thu Jan 18, 2007 11:10 pm

Well, guess I'll give some real world feedback. I ended up using 3 of the Aerocool Turbine 1000 fans in my P180 - one in the lower chamber, one blowing in the front, and one blowing out the top. I also used a Noctura 1200RPM to blow out the back (this moves a fair bit more air than the turbines). So far this seems to be giving me excellent system temperatures, and is pretty quiet (I don't need absolute silence, since I keep my system on the floor, and also since this is a quad core system

.

I had originally used a Noctura 1200RPM in the lower chamber, but with the side panel closed it made a quite annoying "straining" sound as it worked to move air. I tried the 800RPM version, and it was silent, but it moved next to no air. The Aerocool seems to be happy down there and the power supply is running cooler.

I also know it's more traditional to blow out the top hole, however after thinking about the airflow pattern my Zalman CNPS9700 creates I figured I'd try blowing in instead. I also did this since my current P180 system has a CNPS9500 and a fan blowing out the top, and I can never feel ANY airflow in that direction. Blowing inwards I definitely feel air movement, and my system temperatures seem to be doing quite well.

GuyClinch · Post by **GuyClinch** » Fri Jan 19, 2007 1:17 pm

Well I haven't tested THAT many fans but today I got the S-Flex SFF21E (medium) speed fan and it's really quiet compared to my Antec Tri-Cool on medium and technically they are moving the same air. The Tri-Cool I had was LOUD on medium. You could hear it on the other side of the apartment.

So thanks alot SPCR. The only issue I can't seem to read the speed reading from speed fan and it's plugged into the motherboard not the power supply.

Pete

Stereodude · Post by **Stereodude** » Fri Jan 19, 2007 1:22 pm

I replaced 2 120mm Antec Tri-Cools with the 800RPM 120mm Noctua. They Noctua's are definitely quieter without the ULNA than the Antec's on low. They seem to blow a little more air too. I ended up using one with the ULNA adapter because I didn't need much airflow from it. The other I'm running without the UNLA. I didn't notice any ticking, but maybe I haven't listened close enough.

nici · Post by **nici** » Sat Jan 20, 2007 9:43 am

Stereodude wrote:I didn't notice any ticking, but maybe I haven't listened close enough.

Don't do it either. When you start seeking out disturbing noises to make quiet you know you have gone mad.

kaange · Post by **kaange** » Mon Jan 22, 2007 2:57 pm

kaange wrote:
Felger Carbon wrote:
kaange wrote: - if the starting power is pretty much the same, then the windings are different. This assumes that the starting torque is identical, of course.
Greg, let me say this politely: you are mistaken. You obviously have never designed transformers and seen them built, for instance. For a D vs an F, for instance, there will be many more turns of wire, of a smaller diameter, on the D. Therefore, the starting current will be much lower on the D. This is in the nature of motors and magnetic windings in general. No problem, you aren't an electrical/electronic engineer. I am (retired, though).
Actually, I did train as an EE but work in IT but I was posting in a fairly casual way so let me work through your assumption in this post.

If the D has n x turns of the F (where n is a real number), then the magnetic field produced by the same current flow should induce a torque that is n x that of the F. So if the fans are stalled, then the D should start with 1/n of the current needed to start the F. But at this point, since the motors are not turning, it is purely the coil winding resistance that limits the current flow (as there is no back EMF being generated by the magnets on the windings). Now since there are n times the number of windings of F, there would be n times the coil resistance for the same gauge winding but the thinner gauge wire will make the D have more than n times the resistance of the F. This implies that the starting voltage of the D should be more than that of the F, doesn't it?

But let's assume that they use the same guage wire so that the coil resistance of the D is n times that of the F. Then the starting voltage should be identical.

Then again, they could have parallel windings (ie n is an integer) and then the coil resistance would be much less (1/n for the same gauge wire) so the D should start at a voltage of 1/(n*n) vs the F. I guess by using thinner gauge wire and parallel but fewer windings, you could end up with the starting voltage/speed/power draw variations that are seen across the range.

But in all cases, starting power (V*I) should be pretty similar.

Now if the magnet arrangement for the different fan motors was changed, then this alters everything but that hasn't been mentioned as a reason for speed differences of the Scythe Flex's

Please help me correct my analysis (it has been a long time since Uni)

Greg

Felger????

Felger Carbon · Post by **Felger Carbon** » Mon Jan 22, 2007 3:43 pm

The minimum starting power should be constant across the range. All you're doing is overcoming the static bearing friction. With more turns of thinner wire, it takes more voltage to get that minimum power, so the higher RPM fans should start at a lower voltage. But if you're starting the fans at 12V, then you're not using the minimum starting power! BTW, my 800RPM Evercool Ever Green 120mm 800RPM fan will not reliably start below 9V...

s_xero · Post by **s_xero** » Fri Feb 02, 2007 3:59 am

Hmm I've got a new one for the Noctua 1200-version.

At multiple speeds the fan makes a mid-frequency-whining that's audible like hell.
The volume of the whining is lowered acording to fanspeed.

I think it's something like resonation inside the motor-hub.

As far as I can say there's about every mm you move the knob of a Fanmate2, an interfal. It only happens when the fan is used vertically.

Used horizontally, blowing up, it doesn't whine, but a slight shaving noise can be heard; this is due to the bearings.

Can anyone help me with the whines? I guess I should return it..

DonQ · Post by **DonQ** » Fri Feb 02, 2007 11:32 am

NeilBlanchard wrote:Greetings,

Maybe the [internal] speed control in the 800RPM Noctua's is a PWM, or it is working in a way to cause the motor to click?

[Edit: however they are slowing down the the fan, it seems to be causing the clicking.]

That might explain the following situation:

I connected the Noctua 1200 fan to my Gigabyte GA-K8NF-9-RH motherboard's CPU fan header, which has automatic fan speed control based on CPU temperature, through a Fanmate 2.

If I set the fan speed controller to Disabled in the BIOS then I can control the speed of the fan with the Fanmate 2 but if I set the fan speed control to Auto then the fan will spin slower than without the Fanmate, both at slow and fast speeds, but turning the knob on the Fanmate doesn't do anything.

Strange. I read something like this in some other thread but the poster didn't provide specifics.

dfrost · Post by **dfrost** » Fri Feb 02, 2007 4:25 pm

I connected the Noctua 1200 fan to my Gigabyte GA-K8NF-9-RH motherboard's CPU fan header, which has automatic fan speed control based on CPU temperature, through a Fanmate 2.

If I set the fan speed controller to Disabled in the BIOS then I can control the speed of the fan with the Fanmate 2 but if I set the fan speed control to Auto then the fan will spin slower than without the Fanmate, both at slow and fast speeds, but turning the knob on the Fanmate doesn't do anything.

My Papst fan is similarly connected to the Abit AI7 CPU fan header through a Fanmate 2, and the Abit BIOS also controls the header voltage based on user-defined CPU temp when enabled. Pretty sure it is DC variable from 8-12V.

I also have a somewhat strange, but different response to the Fanmate.

When the FM is set to full on, which would be ~11V otherwise, the fan speed varies over a range resulting in fan speeds = ~6 to 11V. When the Fanmate is turned fully down (~5v?), the temperature variation is apparently disabled, and the fan stays at a fixed low speed. I'd guess the point at which temp variation 'turns off' is about half way on the FM knob, but I haven't bothered to test it.

I've chosen to keep the FM at full-on for simplicity (and multiple users in my house), and the Papst fan only ramps up beyond 900 rpm under unrealistic test-only conditions.

nici · Post by **nici** » Fri Feb 02, 2007 5:15 pm

It you feed the fanmate with PWM it wont behave properly, simple as that. If you enable fan control in the BIOS the mobo spits out something that the fanmate thinks is 5-8v DC, and the fanmate can't raise the voltage higher than the input level obviously. If you have the fanmate on full, it reduces the voltage about 1,5V.

Sunfox · Post by **Sunfox** » Sat Feb 03, 2007 6:02 am

One of the annoying things I've noticed about the Noctua is that it makes a strong hum if you make it work to move air. For example in the lower chamber of a P180, or even in the main chamber blowing out when you don't have an equal CFM blowing in. So far the Turbines don't seem to mind working to move air.

I've definitely found the sweet spot for 120mm fans to be about 1000rpm for somewhat silent operation with decent air movement, and up to 1250rpm for a rear exaust fan where a little more noise might be tolerated. Of course 800rpm and what not is practically silent, but it also doesn't do a whole lot to keep high-spec components cool.

What other good 120mm fans are there out there in the 900-1250rpm range at 12V?

Post by **NeilBlanchard** » Sat Feb 03, 2007 9:08 pm

Hello,

If the Noctua's are making noise when they are working against static pressure, then it seems that the blades are "cavitating", maybe?

s_xero · Post by **s_xero** » Sun Feb 04, 2007 1:30 am

I guess. But for you who use it. If you're using an very open heatsink, like the Ninja or HR-01, the fan is actually better than anything else.
I put this fan in place for a Scythe DF(S122512L) which is pretty much on par with the performance of the Nexus and it definetly blows out more air which makes me think that the most of the critisism about the provided airpressure is excaggerated.
So I think that as a case fan, It's good enough. When it gets tighter, then it's performance goes down. I could believe that, but maybe a short review is the way to state at what restriction the performance goes down?

s_xero · Post by **s_xero** » Sun Feb 04, 2007 1:33 am

By the way, I can get a replacement, without to much hazzle. Just fill in your adress etc. state your problem with the Noctua-equipment, add a photograph and send....You'll get a reply that they send a new unit.

Haven't got it though. I'll keep you posted on how the service really is:D
I've just got good hopes.

kaange · Post by **kaange** » Tue Feb 06, 2007 10:49 pm

Felger Carbon wrote:The minimum starting power should be constant across the range. All you're doing is overcoming the static bearing friction. With more turns of thinner wire, it takes more voltage to get that minimum power, so the higher RPM fans should start at a lower voltage. But if you're starting the fans at 12V, then you're not using the minimum starting power! BTW, my 800RPM Evercool Ever Green 120mm 800RPM fan will not reliably start below 9V...

Which is why I asked "Does anyone know the starting voltages and full power current of the different Scythe SFlex's?"

jaganath · Post by **jaganath** » Wed Feb 07, 2007 2:03 am

kaange wrote:
Felger Carbon wrote:The minimum starting power should be constant across the range. All you're doing is overcoming the static bearing friction. With more turns of thinner wire, it takes more voltage to get that minimum power, so the higher RPM fans should start at a lower voltage. But if you're starting the fans at 12V, then you're not using the minimum starting power! BTW, my 800RPM Evercool Ever Green 120mm 800RPM fan will not reliably start below 9V...
Which is why I asked "Does anyone know the starting voltages and full power current of the different Scythe SFlex's?"

Starting voltages I can't help you with, but full power current is stated in the electrical specs:

Model Number:
S-FLEX SFF21D (800rpm)
S-FLEX SFF21E (1200rpm)
S-FLEX SFF21F (1600rpm)

Dimensions:
120 x 120 x 25mm

Specifications:
800rpm Version: 33.5CFM / 8.7dBA / DC12V / 0.10A
1200rpm Version: 49.0CFM / 20.1dBA / DC12V / 0.15A
1600rpm Version: 63.7CFM / 28.0dBA / DC12V / 0.20A

http://www.scythe-eu.com/en/products/fa ... e-fan.html

kaange · Post by **kaange** » Sun Feb 11, 2007 3:15 pm

Ta but without the starting voltages, we can't determine the if the differences with the Scythe Flex fans is due to windings or resistance.

I think Felger has misread my original post on this subject as he agrees that if there is no difference between the Scythe Flex models aside from the windings, then the starting power will be identical but the voltages and hence currents will most probably be different (since the different windings will have different resistance and produce different amounts of magnetic flux).

Now if there was resistance added in series to produce the speed differences, then the starting power would not be the same as some power would be dissipated across the resistor.