Viewing page 1 of 2 pages. 1 2 Next
NOCTUA-ROTOSUB ACTIVE NOISE CANCELLATION FAN
Fans and heatsinks are never in short supply at any computer show, but at Computex 2012 last week, there were two
fan developments worthy of serious note for SPCR readers. The first and more
revolutionary of the two is from Noctua, whose Active Noise Cancellation fan
project with RotoSub I reported on a couple of weeks ago. I had the opportunity
to visit the Noctua booth and meet finally with Noctua's Jakob Dellinger, a
man I've been exchanging emails with for years, and also with Lars Stromback,
one of the principals of RotoSub, a Swedish firm. Noctua has a new range of
heatsink/fans poised for launch, but by far the most interesting item on display
was the early working model of the ANC fan.
Active Noise Cancellation is not new, and many consumers have learned about
it through promotional material for headphones using this technology; the big
ad campaign of Bose for their noise-cancellation headphones comes to mind. The
basic technology is to detect the unwanted noise, then create an inverse signal
which cancels the waveform of the noise to create a null at the point where
it is heard. In a headphone, the technical challenge is fairly simple, because
all the action occurs right at the same point: At the ear. The detection of
external noise by a microphone as it comes through the headphone, the processing
to develop an inverse signal, and the generation of that signal by the headphone
Noise cancellation of a fan is considerably more complex, because the noise
must be tackled at the fan rather than at the ear, and the noise itself has
some unique characteristics which makes it difficult to create an inverse of or replicate. A fan creates noise as its blades cut and beat the air, and because
it is rotating, the waveform is a series of spiraling pulses; the number of
spirals depends on the number of blades, the pulse rate depends on the rotational
speed. A microphone placed near the fan can pick up this sound easily enough,
but to cancel the noise with speakers, at least two would be needed, one on
the intake side and another on the exhaust side, because the sound from the
two sides would not be identical but in reverse phase. Naturally, this would
entail high complexity and cost.
RotoSub's solution to the problem is elegant: A single microphone feeds the
sound of the fan into a processor which then uses a complex algorithm to vibrate
the fan blades themselves in such a way as to create an inverse
of its own noise signal. This means there is a secondary "motor" of
sorts, with tiny magnets at the tips of the blades, and a coil that wraps around
the frame of the fan.
The Noctua/RotoSub working ANC fan sample. Note silver bands in the center of the frame facing us this is part of the coil. Also note wind-filtered mic attached to the heatsink, and the electronics for the noise cancellation signal in small case on shelf below. In the final product, the electronics will be integrated into the fan.
New 37mm tall low profile coolers for both Intel and AMD processors are among the new Noctua products coming soon.
Trying to listen to the working ANC fan model at the Computex booth in the Nangang
Exhibition hall was laughable, because even before I checked the rough-and-ready
deciBEL app on my Google phone, I knew it was around 90 dB ambient, just too
noisy. But between sticking my ear right next to the fan, and using headphones
to watch and listen to a video demo recorded a couple days before, I can confirm
that the early working model reduces the tonal noise of the fan, from easily
audible and annoying, to virtually gone. So there's little question that tonality
can be elminated using this technology, and not just at one or two points; Lars
pointed out the many tonal spikes in the spectrum which disappeared when the
ANC was turned on. But does it do anything with the broadband turbulence? The
answer is no.
I commented in the forums that it seemed curious the Noctua NF-F12 FocusedFlow
fan is their starting point. This fairly new fan, which Noctua says took well
over a year to develop for a mere 8% gain in performance, happens to have higher
tonality than previous Noctua fans, at least as we measured and heard it. In
person, Jakob and Lars agreed my assessement was correct, and it is partly why the ANC
technology is being worked into this model. The FocusedFlow aspect of the
NF-F12, some of you will recall, is basically a flow-straightener that helps increase
pressure by directing the airflow into a more focused area. The price paid is
higher tonal noise. So as I conjectured in the forums, ANC will be used here
not to make an already quiet fan silent, but to make a higher airflow, higher
pressure fan sound quiet. In other words, to make something like a 1600 RPM 12cm fan sound more like a quiet 1200 RPM fan.
This will obviously be more useful for a power user who wants to cool hotter
components with less noise, but it will not do much for some of the uber silencers
who live at SPCR. Asked whether the technology could be used to make a 1,200
RPM fan sound like an 800 RPM model, the answer is yes, but the audible gains
are much less dramatic for most users, and the demand for such a product is far lower. Whether
such a model will be developed is too early to say; the main goal is to complete
development on the current model, which will likely take upwards of two years,
and result first in an ANC-fan and microphone with dedicated heatsink. Mic-integrated
ANC fans will come later. Expect the starting retail price of the ANC fan to
be at least double that of Noctua's conventional fans.
Discuss this news in the SPCR forums.
|Help support this site, buy from one of our affiliate retailers!|