Fan Testing Round 4 Complete (Results Summary)

[Martinm210]

I figured I'd clean up my mess of noise exploring and create a new thread with just the results. This is my fourth round of fan testing, and I've tried to explore many different methods. From PQ curves, to Sound Level meters, to Frequency Response curves, only to be left completely unsatisfied with being able to quantify good sound vs bad sound. In the end, I feel that sound level meters are frankly...PISS POOR...at quantifying good vs bad sound. They measure noise levels only and completely miss quality.

I feel sound "Quality" is a combination of noise level, sound pitch, and rythm which you frankly can NOT quantify in any one chart or number. The only way to really tell is to closely listen and compare at like performance levels.

First off, my testing sponsors for this round of testing "High Cost Fans" (no less than our community members right here in these forums):
Gentle Typhoon AP-15 - Hondacity
Noctua P12 - Masika
Silverstone AP-121 - MagisD

Thanks guys



I also sponsored the following fans myself for the "budget series":
The yate I've had for a while, but the rosewills came in a case I just bought. I noticed the egg sells similar models in set of 4 for $10, so I couldn't help but put them up against the tried and true previous budget fan king...the yate loon.



Test Methodology:

-Fan is hard mounted to an XSPC RS120 radiator with anemometer measuring CFM output, fan in push. CFM number is not calibrated and only good for relative comparisons.
-Video - Canon T2i in 1920 x 30fps HD mode
-Audio - Zoom H1 Handy Recorder, mounted in camera hotshoe, aimed at fan. Manual level 90, low cut on, max .wav setting.
-Monitors - In the background is the Anemometer display, two cheap sound level meters, a voltage meter (at plug), and a Skythe Kaze Master to monitor RPM.
-Power - Mastech 0-30V 5Amp lab power supply giving the Kaze master roughly 13V for Vdroop (this allows a full 12.00V to be provided to the fan). 5V input is provided by an old phone charger.
-Warmup - Fans/Anemometer were each allowed a 5 minute warmup at 12V, then testing proceeded.
- Environment - Tested in my walk-in closet at 12-1 AM when ambient noise level is perceived to be the lowest possible.
- Processing - Combined Video/Audio in Sony Vegas, matched up finger snap audio from T2i to Zoom and replaced poor T2i audio with Zoom audio. Audio level held to the original recorded.
-Recorded Ambient section, then through voltage ranges stopping at even 1 Volt increments and even 10CFM increments for comparison.





Products:
Video/Audio as tested
Comparison Video/Audio comparing 10, 20, 30...etc CFM levels


Uses: Suggest careful listening by the end user to make their own determinations. Use CFM anemometer levels to compare like performance levels.



INDIVIDUAL VIDEO RESULTS:
Fan Test 4 Yate Loon D12SM12 Curved Black
http://www.youtube.com/watch?v=vILqRsVDqX8

Fan Test 4 Servo Nidec (Scythe) Gentle Typhoon D1225C12B5AP-15
http://www.youtube.com/watch?v=8AmF1Z0M8JY

Fan Test 4 Rosewill DF1202512SEMN BlueLED Clear
http://www.youtube.com/watch?v=mLXsDcQvRS4

Fan Test 4 Rosewill DF1202512SEMN Black
http://www.youtube.com/watch?v=hE48YxQwnOo

Fan Test 4 Noctua NF-P12
http://www.youtube.com/watch?v=uuTm9XysZVA

Fan Test 4 Silverstone AP121 HA1225L12SF-Z
http://www.youtube.com/watch?v=Cec_HlQK4IQ

COMPARISON VIDEO RESULTS:
http://www.youtube.com/watch?v=rs3hnMGPGi4

Where are the charts you ask?

I'm not doing them, I think sound level meters paint a very inadequate, incomplete, and false picture. They give you some indication of noise level, but they do nothing for quality. You need to listen to the audio to get the real sense of what's good sound vs bad sound...sound level based charts just can't do that. For example the Rosewill fans perform at similar noise levels as the Yate Loon, however the motor ticking that's present in the Rosewill fan is much more harsh/annoying than the Yate. I see similar relations between the GT and Noctua here. At 20CFM levels the GT is CLEARLY lower in noise level, however it also has a higher pitched whine. These noise characteristics vary at different voltage levels. I see the GT being exceptionally strong at 10CFM and 40CFM, but I prefer the Noctua at 20CFM. I'll leave it up to you to decide...

Cheers!
Martin

[MagisD]

Whoot great job martin.

-Fan is hard mounted to an XSPC RS120 radiator with anemometer measuring CFM output, fan in push. CFM number is not calibrated and only good for relative comparisons

I noticed the ap-121 was much higher on rpm to push the same amount of air as the rest. that inbuilt grill straightens but cuts flow I guess.

Strange thing is I found the GT at 40 cfm easier on the ears then 20-30 it had a whine in these tests i don't have that in RL.

[Martinm210]

Whoot great job martin.

Thanks! and thanks for sponsoring the Silverstone..

Glad it helps...

Next on my plan is to use the two Noctua P12's that Masika sponsored and do a similar effort specific to push vs pull vs push/pull. That may not be until next or the following weekend, but I have two P12's slated for that.

I have a feeling that push is better than pull for noise and possibly push/pull is better than one fan...but we'll see. I don't think anyone has actually tried to measure that.

[Martinm210]

I noticed the ap-121 was much higher on rpm to push the same amount of air as the rest. that inbuilt grill straightens but cuts flow I guess.

Strange thing is I found the GT at 40 cfm easier on the ears then 20-30 it had a whine in these tests i don't have that in RL.

Exactly, I personally prefer the Noctua at 20CFM better than the GT. The GT has a strange high pitch noise that's a bit annoying at the 20-30CFM levels. The GT seems to perform best at 10CFM and maxed out at 40CFM, but not everywhere in my opinion.

[Vhaara]

Yeah, i kinda noticed the whine too. I hear people rave about the GT's, and at the lower RPM they where so quiet, but as soon you started upping the RPM i definitely noticed the whine. Any reason why its doing that?

By the way, absolutely fantastic test!

Whoot great job martin.

Strange thing is I found the GT at 40 cfm easier on the ears then 20-30 it had a whine in these tests i don't have that in RL.

[Martinm210]

Yeah, i kinda noticed the whine too. I hear people rave about the GT's, and at the lower RPM they where so quiet, but as soon you started upping the RPM i definitely noticed the whine. Any reason why its doing that?

By the way, absolutely fantastic test!

Not sure, some sort of harmonic thing, maybe specific to the radiator interaction or something...might be different in pull vs push also.

[Monkey Puzzled]

Great testing as always. Have you tried the 1450 rpm vs the 1850rpm gentle typhoons? I seem to remember reading the 1850 models were louder at the same rpm when undervolted.

[jayhall0315]

Hey Martinm, another you might want to include is the Scythe S-Flex series. I own both the GTs and the Scythes, but I tend to run them in the 800 to 1000 rpm range, and in that range, the Scythe D and F models sound better than the GTs. (sound better for the same cfm as well, too)

[Martinm210]

Great testing as always. Have you tried the 1450 rpm vs the 1850rpm gentle typhoons? I seem to remember reading the 1850 models were louder at the same rpm when undervolted.

I haven't. Xbit tested them and seemed to indicate noise levels were higher with the lower models in open air testing. But I'm pretty skeptical of noise levels, the AP-15 tests really low noise levels at 20CFM, but it has a higher pitched whine. Very possible one of the lower models doesn't exhibit that same frequency change.

It seems generally I'm finding most fans are pretty optimized for operation at 12V. The odd ticking and weird harmics usually present themselves when undervolted on some fans.

Hey Martinm, another you might want to include is the Scythe S-Flex series. I own both the GTs and the Scythes, but I tend to run them in the 800 to 1000 rpm range, and in that range, the Scythe D and F models sound better than the GTs. (sound better for the same cfm as well, too)

I tried some in my round 1 and 2 testing. I didn't have nearly as nice a recording instrument back then, but I do have the old videos if you want to check them out here:

Round 2 videos of the S-flex vs GTs:

Sflex-E = Scythe S-Flex SFF21E
http://www.youtube.com/watch?v=RcJ4Yr2krTk

Sflex-F = Scythe S-Flex SFF21F
http://www.youtube.com/watch?v=5ygYCWhwP7c

Sflex-G = Scythe S-Flex SFF21G
http://www.youtube.com/watch?v=rrtRQD2q1vM

Gtyphoon AP15 = Servo Nidec Gentle Typhoon DC1225C12B5AP-15
http://www.youtube.com/watch?v=bLgJomTh74w

They are ok, I just don't think they have dense enough fan blades to be as smooth as the GT or Noctua on a rad.

I've since given them all away...

[MagisD]

Hey Martinm, another you might want to include is the Scythe S-Flex series. I own both the GTs and the Scythes, but I tend to run them in the 800 to 1000 rpm range, and in that range, the Scythe D and F models sound better than the GTs. (sound better for the same cfm as well, too)

See that's why martin post the videos and why i sent him an ap-121 instead of bring it home. The sound is subjective.
Its litterly a matter of personal taste. The settings and setup for comparison matters.

The cfm through rad is measurable along with actually volume.

But what sounds good or easy on the ears is completely subjective in most cases.

Gona wait for my 1850's and "tune" them once there in my case.


Sent from my X10a using Tapatalk

[MagisD]

Another test i just thought of .... in pull how good is the airflow after.the fan. As in rads covering intakes etc how much airflow are you gona get and how far. So you know if its enough to cool passive cooling heatsinks etc.

Sent from my X10a using Tapatalk

[Martinm210]

Another test i just thought of .... in pull how good is the airflow after.the fan. As in rads covering intakes etc how much airflow are you gona get and how far. So you know if its enough to cool passive cooling heatsinks etc.

Sent from my X10a using Tapatalk

Not sure if you mean Rad CFM vs Open air CFM?

I did some of that in round one, here is one example. At the bottom, Orange is open air testing, green is with a radiator. Rad CFM was 15 vs Open which was 45, so you only get around 30% or so of the rated open air CFM.


This is why it's so imperative to have a fan designed specifically for pressure applications.

Unfortunately I think box open air CFM specifications is what most people refer to when picking a fan, so in a way we are encouraging fan makers to design fans for open air applications.

Take a look at the market of fans, they almost all follow the same 7 blade design, which is really not very optimal for radiator setups.

[Masika]

I am finally back home.

Thanks for that Martinm20. Great to see the video to see the evidence for oneself. I look forward to future tests.

[Martinm210]

I am finally back home.

Thanks for that Martinm20. Great to see the video to see the evidence for oneself. I look forward to future tests.

Thanks for sponsoring, I plan to use the two P12s in my next mini test to compare push vs pull vs push/pull regarding noise to CFM ratios..

I'm still trying to think through how I'm going to do that but I've got some ideas. Probably have to test voltage vs CFM in step one to determine the right voltage levels, then remove the radiator and test those voltage levels vs noise. It can be done..

[Masika]

NP at all. I wanted to see the noise and I got it. I have to say that when I mount my noctuas vertically I can not not hear them at all - the push up. Would be an interesting test to change the orientation of the fans from vertical to horizontal.

I am looking forward to next tests. Would be good to see the fans verse cooling on rads. I know there are a few around but noctua P12 are not on many.

[skinnee]

I won't touch the noise scenario as I am a firm believer in what Martin found... we all have an individually subjective ear.

However, with the assortment of fans I have and the thermal test bench for radiators, doing a fan and temp analysis is something I think we would all find interesting. I'll need to get some smaller aquarium heaters though.

[Metroid]

By the tests, the Gentle Typhoon is 1.7x better than the Noctua NF-P12 and doing so at a lower DBA. It is time to get one to check it out. I can always lower the voltage anyway. Once again thanks Martin.

[slaveondope]

Im pretty happy with my cheap Rosewill Blacks

[fgw]

Thanks for sponsoring, I plan to use the two P12s in my next mini test to compare push vs pull vs push/pull regarding noise to CFM ratios.

martinm, may be you also check the effect of using both fans on the same side of the rad. interested to see if this would improve performance in regards of cfm or static pressure? i'm currently building a custom duct do direct a fans airflow to the back of a gtx295. looks like the sflex-g i'm using lacks on static pressure. so i think of using two of them directly attached to each other and see if this improves temps. will take a few days until my test setup is ready, but may be you can make a quick test in your setup too. never tried this before, but i'm curious to see whats the effect of two fans attached together like this.

also,would be interesting to see how the new thermalright ty-140 does here? x-bit stated this fan does very well and is real quiet. i know this is a 140mm fan but at least it has 120mm mounts. unfortunately this fan is not yet available her, so no first hand experience with it.

anyway, great info and even greater testing performed here as all your tests, thanks for all the work!

[Pliskin45]

The gentle typhoons definitely have a sound signature I can't stand, but they move so much air! Probably going to switch to noise blocker np12 or whatever its called.

[MagisD]

martinm, may be you also check the effect of using both fans on the same side of the rad. interested to see if this would improve performance in regards of cfm or static pressure? i'm currently building a custom duct do direct a fans airflow to the back of a gtx295. looks like the sflex-g i'm using lacks on static pressure. so i think of using two of them directly attached to each other and see if this improves temps. will take a few days until my test setup is ready, but may be you can make a quick test in your setup too. never tried this before, but i'm curious to see whats the effect of two fans attached together like this.

also,would be interesting to see how the new thermalright ty-140 does here? x-bit stated this fan does very well and is real quiet. i know this is a 140mm fan but at least it has 120mm mounts. unfortunately this fan is not yet available her, so no first hand experience with it.

anyway, great info and even greater testing performed here as all your tests, thanks for all the work!

It was a couple of years ago but somebody proved that didn't work well unless you spaced them out.
But push/pull worked better if you were useing 2 fans.

Have no idea where i read that on here.

Sent from my X10a using Tapatalk

[Martinm210]

See that's why martin post the videos and why i sent him an ap-121 instead of bring it home. The sound is subjective.
Its litterly a matter of personal taste. The settings and setup for comparison matters.

The cfm through rad is measurable along with actually volume.

But what sounds good or easy on the ears is completely subjective in most cases.

Gona wait for my 1850's and "tune" them once there in my case.


Sent from my X10a using Tapatalk

My thoughts exactly..I know a lot of people want charts to decide from...but this is yet another reason to stick with the videos...sound is simply to complex and subjective to put a number on...the best we can do is record it and let folks listen..

NP at all. I wanted to see the noise and I got it. I have to say that when I mount my noctuas vertically I can not not hear them at all - the push up. Would be an interesting test to change the orientation of the fans from vertical to horizontal.

I am looking forward to next tests. Would be good to see the fans verse cooling on rads. I know there are a few around but noctua P12 are not on many.

Never thought about trying that...Ill see what I can do on the noise test, but I'm not set up for rad thermal testing anymore.

I won't touch the noise scenario as I am a firm believer in what Martin found... we all have an individually subjective ear.

However, with the assortment of fans I have and the thermal test bench for radiators, doing a fan and temp analysis is something I think we would all find interesting. I'll need to get some smaller aquarium heaters though.

That would be great if you have time and interest...I just don't. Want to run thermals..I can't do thermals and noise/cfm at the same time anyhow with the crystalfontz/pc logging requirement on thermals.

You've got the thermal part down pat anyhow...I'd probably just try to catch my house on fire again..lol!

I'll stick with the noise related stuff for now..no thermal testing for me..

martinm, may be you also check the effect of using both fans on the same side of the rad. interested to see if this would improve performance in regards of cfm or static pressure? i'm currently building a custom duct do direct a fans airflow to the back of a gtx295. looks like the sflex-g i'm using lacks on static pressure. so i think of using two of them directly attached to each other and see if this improves temps. will take a few days until my test setup is ready, but may be you can make a quick test in your setup too. never tried this before, but i'm curious to see whats the effect of two fans attached together like this.

also,would be interesting to see how the new thermalright ty-140 does here? x-bit stated this fan does very well and is real quiet. i know this is a 140mm fan but at least it has 120mm mounts. unfortunately this fan is not yet available her, so no first hand experience with it.

anyway, great info and even greater testing performed here as all your tests, thanks for all the work!

I did try that...don't think I documented it well..but it doesn't work. Since air is compressible, fans really disturbe/compress the air so you need some sort of flow spreader like a radiator or a long channel to allow the air to clean up before entering the next fan. At least with yate loons it didn't work...may work a little on something like the silverstone though I suppose. Regardless,, Ill check this again with the noctuas.

Next round will be evaluation of fan/rad noise in differing orientations. Something lie push, pull, push and pull, push and push, horizontal vs vertical.

The only difficulty I have with fan testing is the 1AM requirement. I get up for work at 5, so I can't make runs during the weekdays. Ill see what I can do next wekend..

Sounds like Hondacity has a couple of additional fans he wants to sponsor. PM me if you guys have others. I don't mind doing these if I can keep it to half a dozen fans or so per round.

[OldChap]

The comparison video is a stroke of brilliance .....so obvious It is hard to say why it has not been done before.

Makes me want to replicate your work with my own test rig and a few other, perhaps less obvious, choices of fan

A hearty thank you for your work

[Vapor]

Gotta agree with you here Martin, what the SPL meters are reading seems to have little bearing on how okay the noise is.

Odd that the GT is so good at 25CFM+ and kinda has a weakspot at 18-25CFM. Even at 40CFM though, it's better or as good, to my ear, as any other fan at 25CFM.

In that 18-25CFM range the GT is weak, seems the Yate is the best to my ear. AP121 is pretty bad all over. Rosewills fare well, same with the Noctua, but the Noctua is way expensive and sounds like it's "doing more work" than the Yate to me.

Full-speed GT is really impressive though, the whine is gone and it's moving a ton of air for how little noise there is.

Great stuff, thanks for doing it

Do you have push+pull Yates on the docket?

[Vampiyer]

Yep - if I turn my GTs too low, there is a slight whine, but I don't notice it inside of my case.

I don't ever drop them below about 9v, and above that voltage, my fans do not whine. I wish I had more of them - they're noticeably better than my San Ace fans with regard to noise quality.

I thought the Noctua's sound signature was unoffensive, but noticeably louder across the spectrum. It's a bit like a larger displacement V8 engine just humming along, whereas the Scythe GT is... really in a different league...

I'm hoping that one of these days someone will run a formal test on the Sanyo Silent series (specifically the SA-9S1212L401). Ehume's megahalem's test shows it to be better than the GT-AP15, as does the Sanyo spec sheet (with regard to airflow and static pressure at that noise level).

The new Thermalright TY-140 fan included with the Silver Arrow also looks good. This test: http://www.overclock.net/air-cooling...40-review.html shows it at 1.89mm H20 static pressure (vs GT-AP15 at 2.057mm H20 according to spec sheet) and 70cfm.


Nice work again Martin, as always.