However when you want the closer to the absolute noiseless 140 mm fan -with still some air displacement- Your own measurements price Nanoxia DX-700 as the real winner in lower dBA readings (Lowest dB at Graph 1, pg12).
A feature that is even accomplished at 500 RPM, higher than close alternatives. I do wonder what would be the actual dBA readings at lower RPMs? In fact a detailed low dBA/rmps, airflow measurement comparison is completely missing from your study.

What if some of us use ie. Seasonic Fanless PSU, noiseless SSD and then want the lowest noise CPU/case fan... possibly to work on just exceptional peaks, where obviously it should go unnoticed?

I work with sound and aim for a fanless -and still powerful- set up, which -for safety- will need a fan somewhere. Thus, I'd like to see that indeep review definitely discriminating fan units below 33 dBA, and hopefully below 20 or lower dBA measurements.
Thanks for what it's done already.

Excelent review Sergey! I see you put plenty of midnight effort.

Nice review. It was so interesting to see how tightly they clustered at 1200 - 1300 rpm.

As always, there is a roster of missed fans. But in this case there were only two: the Yate Loon D14SL-12 and the Prolimatech fan that comes with the Aramageddon. Now *that* is impressive.

"Thermalright TY-140 comes with a recommended price tag of $14.95 and is furnished with a one-year warranty."

At $12.99, the best fan in this roundup just got better!
http://www.svc.com/tr-ty-140.html

This makes the SilverArrow cooler a formidable package as it comes with 2 of these fans already.

I agree with ehume that the Prolimatech Vortex should have been included since its rated at 87 CFM @ 18dBA.

Plus, could you confirm if all the tested Fan models are the Brushless type? This is important, because electric motors with brushes produce electromagnetic distortions that are passed onto the powerlines (and into audio and earth planes) where the offenders are plugged into.

The Audio world has long relegated this lesser tech out, hope all fans here pass this test

Can the approximate range of noise readings with the fans off be given to us? Even better, can the actual fan off readings at time of the measurement be given to us?

If I missed it in the article, my appologies.

I suspect that the background level changes a few dBA during the many hours of testing this article took!

If you look at the dBA curves they asymptotically approach a constant value as the RPM (or CFM) is reduced. This would appear to be at least in part due to fan noise approaching background noise and the lower measurement limit of the meter. I believe this is one of the reasons the meter is so close to the fan, if it was further away the noise levels would be too close to the noise floor for more of the RPM range.

You can see what I mean at SPCR for the TY-140 fan where their measurements are with an 11dBA noise floor and this fan shows very little curve over in noise at all RPMs.


http://www.silentpcreview.com/article1039-page8.html


So it is actually critical to know the dBM reading at the time of the test for each fan. The true fan noise can be estimated using this information.

hay xbitlabs, bought the TY-140 year and a half a go on your recommendation, and now it become noisy.

i would have appreciated you making it clear this fan
warranty is only 1 year and that the fan hinge cannot be serviced.. this is crucial data i was not aware of..

i found this after contacting TR
thanks to TR graces & fast customer service im getting a new and TY-140 at 50% off.

i think readers would appreciate getting the full picture about a product and not just performance specs.

Thanks for the review
If I look for fan's of 140mm it's because I wanna combine 2/3 or 4/6 on a double or triple radiator. So the shape of the best fans aren't okkey.

These fans are better just because they have a bigger blade.

Actually I see there is not very big difference with fan's of the same size between the cheapest and the most expensive one.
Very well done for the graphs.
Another site tested and found out that efficiency quite depends also from the direction of the air, so not only the overall cfm.
Maybe it would have been interesting to test the temps if putted on a radiator.

Thanks anyway. Until now I think I pretty much would stay on the Yate loon, a fan that spins until 2400 rpm, but becoming very noisy. Anyway the difference with the same blade size isn't much relevant. The only one that is 15% better than the mass is the thermalright x silent 140. But is limited on the RPM.

I was reading about the corsair fans and realized there are 2 different pourposes:
Airflow
Static pressure
So the one that do more static pressure have bigger blades and have more power to push through (for example) a radiator.

This makes the whole cfm reading, when talking about radiator fans, a wrong way to evaluate a fan.



In my pc all fans are connected through controllers, so I can change the power> and change the speed. So the performance at max speed isn't anymore a useful data. It is more interesting to know that a fan is silent at 30% and makes discrete airflow, then maxing out, it becomes powerful and noisy. I will compare this fan with a quiet one, thinking that if I need to, I have alot of extra room.
And the whole cfm value becomes clearly subjective. If looking for a fan for a radiator I'd like to compare the cooling performance, means the water temperature.


I write here after reading through about 10 reviews, if I'm allowed to, id like to say the most useful one are the one that put on a list with 2 values all the fans, even better if the fan are present 2 times, one with full speed one with half. The data that the user looks for are the noise and the cooling performance. So a list like this ordered by name db and temperature, is the most useful resource when comparing fans.

Thanks for your work in reviewing all sort of hardware.

This is a great review. I would like to note that the "wind tunnel" cfm measurement does allow for some load but not much resistance as I understand would exist for a direct radiator attachment. So I presume this was more for an "average" or case flow cfm. I am looking for top mount exhaust fans and felt these results would be good for my purposes. Thanks!!

Also- for overkill only it would be good some day to test the frequency range as well as the raw dBA's of these fans as we are particularly sensitive to the 400-5000Hz range of normal speech as an evolutionary byproduct. Many manufacturers address this range, but sadly, that is well within the range case fans generate. sidewindercomputers.com has a "Listening Room" that both records and illustrates the point that SPL/dBA is only half the story behind "annoying" sound by using several of their for sale fans. Frequency and the phase of sounds are at least as important as amplitude (the "volume" of sounds).