Slashdot Mirror
Cancelling Out CPU Fan Noise
Percy_Blakeney writes "After realizing how noisy his computer was, a professor at BYU has created a new CPU fan that uses small microphones and speakers to cancel out its own noise. It isn't perfected yet -- it only nixes the whine, not the whoosh -- but it looks like it could be promising, especially given the professor's background: making jet engines quieter."
I have heard of something like that for cars ages ago, basically replays the engine sound over the car sound to negate it.
There were various addons with such a system so you could add a roar of a 911 or rattle of a clapped out sad wanker boy racer in the car.
Jonty! Neil! Work!!
I think this is how noise cancelling headphone do it - they just feed the external noise back into the earpieces after inverting it.
Its actually used on some heavy earthmovers and tractors simply because its actually cheaper than making a decent muffler.
Don't worry - its just stigmata. Pass me a napkin and don't you dare tell my mother.
What you describe is not that same thing which is mentioned in that article. You describe passive cancellation (i.e. simply reflecting and hoping it will cancel the original noise), whereas the article describes active cancellation (i.e. recording the noise, computing the negating and sending it off) of noise.
Pick up some of the new fluid-drive-bearing units most companies are producing these days. I can hear my 60GB drives when they access, but the 250GB drive is completely without any detectable noise.
Of course, I do need one of the prof's nifty new toys for other parts of my system...
You can never go home again... but I guess you can shop there.
Read my post carefully, I actually talk about both technology.
Although, it would be very very cool to get this technology to work on big loud things, and is very cost effective, for quite pc's, the Voodoo F:50 does a very good job at keeping noise at a minimum, using no fans, only convective heat pipes, and using the entire case as a heatsink. Voodoo claims that their system operates at below 20 dBs, and cannot be measured in a room with regular ambient noise.
Skill is successfully walking a tightrope over Niagara Falls. Intelligence is not trying. -- Anonymous
The headsets also have 2 microphones in series and out of phase - 1 picks up the voice+noise and the other only the noise, so the noise cancels and you get left with the voice. This stuff is used on comms systems for concerts, etc as well.
Rob :)
Not to worry, the cancellation -can be- that complete. In practice, there will be some residual noise, but it will be very quiet.
I think you're referring to a "blast fence." Those have nothing to do with active sound cancellation, they're strictly passive noise control devices that block the path between the noise source and the receiver (just like highway noise barriers). See here or here for examples (the latter is a run-up enclosure, but it's the same principle).
If you can't hear it, it's not hurting you (assuming that it's a frequency that you could normally hear). Typically what will happen with a scheme like this is that the cancellation will only work in one direction. In other directions, it will reinforce rather than cancelling. For instance, you can buy noise-cancelling headphones, but the cancellation only works for your own ears, which gets the sound in just right right phase; to the people around you, there will be a perceptible noise coming from your headphones! Conservation of energy says you can't just destroy the energy of those sound waves. Most likely you're just sending extra-strength sounds waves somewhere else. Theoretically the extra energy could be converted into heat, or electrical energy, but I doubt that's really practical.
Find free books.
how it works it also has more informative links.
"Adaptively Cancelling Server Fan Noise" can be found here. They were able to lower the whine by 30dB and the broadband noise by 20dB.
First, get yourself a copy of Forrest M. Mims III's "Engineer's Mini-Notebook - Op Amp IC Circuits" (RS 276-5011) - probably not very easy to find (I believe it may be out of print - go to ratshack and ask). Alternatively, grab a copy of "The Forrest Mims Engineer's Notebook" (ISBN 1-878707-03-5).
Ok, now - if you have the mini-notebook, look on page 12 - if you have the other book, look on page 93. Basically, what you are looking for is the "Inverting Amplifier" (both circuits are nearly identical - though the original booklet shows a resistor R3 hanging off pin 3 of the 741).
So, anyhow, you run your mic input (exercise left for the reader) into the Vin to pin 2 of the 741, and an inverted waveform will show on pin 6. Run that output through an amplifier, then the output of the amp into a speaker - and there is your basic sound cancellation system.
By taking the input sound waveform, inverting it through the 741, amplifying it, then outputing it through a speaker - the valleys and crests of the two waveforms should nearly match each other (with the exception of a slight delay introduced by the circuit/amps).
Want to take it a step further? Tie an A/D and a D/A on each end - and inbetween set up a really fast DSP or microcontroller, and perform on each sample of the waveform a form of amplitude forecasting - so that you can maybe cancel out the effects of the delay in the circuit by adjusting for them dynamically. Probably would be difficult to do homebrew, but if you grok what I am talking about, you can see how it would help, and why it would be fun to try.
Hope this inspires someone - good luck!!!
Reason is the Path to God - Anon
Bose noise cancelling headphones suck, they aren't even the best active noise cancelling headphones available. Sennheiser has better models for around the same price. Far better than either though is the Etymotic ER-4P, these in ear canal headphone provide over 24dB of isolation, with some nice jazz playing you won't hear anything outside, headphone.com has them for only $219, about the same as the Bose units.
There are 4 boxes to use in the defense of liberty: soap, ballot, jury, ammo. Use in that order. Starting now.
No Slashdot post about computer noise is complete without a link to Silent PC Review.
Losers choose to abuse the use of "loose".
you can rewire your fan for 7v instead of 12v (unless you like to do monster 3D rendering sessions for a few weeks), drops noise drastically
just use the 5v line and the 12v line as + and -
(no earth) the voltage differential is 7v so you wont need to usa a resistor, if you still need large airflow then look at Panaflo fans (made by panasonic) 21db quoted @12v
see
this store for the most common fans
While that would be fine, it actually costs more, and it weighs a TON (or at least a thousandth of one, which is nearly as bad.) If this noise cancellation can quiet a system just as much, for the same price, without the potential of ripping a hole in your motherboard, I'd call it a win.
The noisest part of all my computers i the hard drive, not the CPU fan.
How to reduce drive noise:
1. Take your hard drive off the mounts.
2. Find a 3.5" drive mounting kit and a piece of foam (styrofoam, or packing foam).
3. Mount drive on mounting kit, place on top of foam in the bottom of your case.
4. For completeness, ground the mounting kit to your case.
This will knock out a very large portion of your drive noise that is getting transmitted to the body of your case. It is a little Rube Goldberg, but it is very effective.
Someone send this guy a link to
http://www.silentpcreview.com
Implementing noise cancellation for poor quality whining fans seems ridiculous in comparison to replacing the fans with better quality ones.
Quote from SPCR -
" What is a good inexpensive & quiet general purpose fan?
The 80mm Panaflo FBA08A12L with "HydroWave bearing" is widely used and recommended for its combination of low noise (21 dBA), good airflow (24 cfm), wide availability (but not in Canada where I type this) and low cost. At 7V, it is almost inaudible in most applications. At 5V, it is inaudible but still provides some airflow. We think of it as a workhorse, suitable for use as a case fan, CPU heatsink fan, or PSU fan replacement."
...is to buy Zalman components ( http://www.zalman.co.kr/english/intro.htm )
I built my last PC with their components. When I powered up for the first time I freaked out because I saw the power light go on, but that was it. Then the BIOS came up, thank god. No noise at all...seriously. I mean, I expected quiet, but not noiseless...
I was extremely let down by my hard drive though. Considering Seagate had a great reputation for quiet hard drives, I figured I'd get a Seagate SATA hard drive...well their SATA drives are loud as heck when writing...
01100111 01100101 01110100 00100000 01101111 01110101 01110100 00100000 01101101 01101111 01110010 01100101 00101110
he's a professor at Boston...
Boston? BYU is Brigham Young University, a private school located in Provo, Utah. It is a somewhat prestigious university in its own right, but it certainly isn't any school in Boston.
Moof.
1-2dB at certian frequencies (yes, at the limit of what you can hear different) and with a studio dB metre/spectrum anaylizer (don't remember the brand). It was in a studio, under ground, climate controlled, in the desert in the summer, so fairly consistent conditions.
Also, it eliminates the smell. New B&W speakers produce an odd smell when you shake the air up in them (ie whenever there is any significant amount of bass). Do it enough and for long enough, it goes away.
This solves that and will break in the speakers, assuming they need it.
Look I won't say for certian that the changes were a result of breaking in the speakers. However, I believe it to be the case. Either way, it costs nothing and deals with another problem.
Voodoo shit like buying $1000 speaker cables or spending lots of money on sand pads for my amp that are contrary to what I know about science, I'm NOT going to do without reliable emperical evidence to suggest it does something. However something simple like speaker burn in which is free (other than minimal electric costs), I have some empricial proof (though not conclusive), and DOES appeal to what I know about science, I'll continue to do until I see solid emprical proof that it doesn't do any good.
I have both the MDR-NC20 and the older Bose model. If you get the Bose, make sure you pay the extra money for the newer model, otherwise you're better off with the NC20's. The older model has an annoying dongle that takes two AAA batteries. It also requires the batteries to be operating to play any sound. It also seems to amplify the bass a bit more than I like, but some people might thing that's a good thing.
The newer bose model has a battery up on one of the ear pieces (I believe a single AAA), no dongle, and passes sound without being turned on, iirc. It also has a single cord going to one earpiece, instead of the Y split. Both bose models have quite a bit more padding than the NC20's and are closed earphones.
My biggest problem with the NC20's was that I kept stressing the weird mini-plug thing on the bottom of the cord and it finally broke. Fortunately, I'm handy with a soldering iron. Other than that, I think they are a great pair of headphones.
I believe Bose has a 30-day "test it out" period with their headphones, so you could give them a try.
/ \
\ / ASCII ribbon campaign for peace
x
/ \
Oops, I re-read my post and realized I screwed up on one point. The speaker output isn't zero. The combination of the ambient sound and speaker output at the ambient sound frequency should be zero.
I noticed that my University employs the use of masking. While taking a test in the religion department at my University, I noticed that the air conditioner was a little loud. Looking around, I noticed that there were four speakers placed near the doors of the religion professors. I figured this was to provide privacy for the professors when counseling people, and so passer-bys would not hear any of the comments. Anyhow, since then, I noticed that in the new building that was just completed, there are a lot of these air-condition sounding speakers. It is a lot cheaper to implement a system where you just have a speaker replaying the sound of an air conditioner, than by going out and getting a system that creates white-noise.
The views expressed are mine own and do not express the views of my employer.
I was obviously wrong in my prediction about an audible noise for people not wearing the headphones, but I'll bet both my testicles that it's not because conservation of energy is violated. You'd get the Nobel Prize if you found a violation of conservation of energy.
Your bloodline is safe :-) I'm guessing the energy is dissipated as heat in the speakers and voicecoils themselves. The cone will travel further than normal since it will be 'flowing with' the incoming pressure waves rather than working against the air as normal.
As you say, there's only a small amount of power in the small zone where the sound is deadened, so not all that much extra heat in the speakers.
Try Maxtor's fluid bearing drives. In an old AT box the only way I knew the drive was being acessed was the case fan slowing down. I now have 2 in my current box, along with Panaflo Low case fans (80mm and 92mm) and AMD's own heatsink and fan on a 2000+. Very quiet for the performance. Also check out the Volcano 10 heatsink and fan, thats quieter than AMD.
Professional Politicians are not the solution, they ARE the problem.
- THGs review of the TNN500A
THG sez you can still hear(barely) the HDD and optical, but if you're a noise weenie, do like the govt and replace everything with solid state(HDD and use CF for transportable media)The case is stupid expensive at $1400US and the adventurous could probably build one for less by cannibalizing heat pipes from VGA coolers and stripping heatsinks from dead hifi amps, but there are ways of reducing PC noise without killing yourself or your bank account:
case - antec sonata or slk3700bqe
PSU - antec's yet-to-be-released phantom 350W PSU, or check this list:
using vibration absorbing grommets for everything that vibrates(HDD, Optical, fans, etc.)
quieter fans:
OR, get longer cables and put the machine in an airconditioned closet; with a long USB2 cable and a powered hub, you might never hear your machine again. it'd just be you, your KB, monitor and a 7-in-1 media reader.
"...that's as white as it gets; all the bits are on..."
your only solution is to do what I'm doing,
put all the computers in a rack downstairs, cut a hole in the ceiling and the floor, and run the cables up to the second floor office!
*fire pole sold separately
I live in Orange County. In City of Orange, the 22 at the 55 is *silent*, it's about a half mile of a different type of pavement. When you're driving with the windows down, it goes from a roar to next to complete silence. The parent comment is correct.
The fan controller in the Power Mac G5 is aware of the noise amplifying and canceling effects of running different fans at different speeds in different combinations. It actively uses this information and uses it in decisions on how to cool the G5 in the quietest manner using the 9 strategically placed fans.
Would that be vulcanized asphalt, a concoction of regular asphalt and ground up rubber from used tires? This is something that solves several problems at once. It prolongs road life, because the rubber allows the asphalt to contract and expand easier when temperatures and humidity fluctuates. Even cooler, it repurposes used tires which are piling up around the country.
Naturally, federal and especially state DoTs want nothing to do with this stuff as it has the potential to significantly slash their budgets for road development and maintenence. Yay for politics!
It should be noted that the flow of blood, chewing, other internal sounds, etc. would be a LOT louder if we didn't have the middle ear. One of its (theoretical) functions is to reduce the transmission of bone-conducted sound to the cochlea.
/trying to see if I can actually interject something I learned in my music psychology class into a conversation....
I belong to the ______ generation.
I beleive Alpine and I know MB Quart does this, two of the highest quality car audio speaker manufacturers.
I follow the SDK and GDN principles.. Spelling Dont Kount, Grammer Dont Neither