Slashdot Mirror
BYU Project to Silence Computer Fans
phunster writes "The New York Times has an article about Scott D. Sommerfeldt and his students at BYU who have created a noise suppression system for computer fans (drop of human blood required to read article). The technology is not new, he uses out of phase sound to substantially cancel out the sound of the fan. What is interesting is his implementation of the technique. While other systems place a microphone and speakers in the center of a room, he places four miniature speakers and microphones around the noise source itself. His results are promising."
BYU Project to Silence Computer Fans
We as /. computer fans have been discriminated for so long, that
you'd think that we, as computer enthousiasts, have had quite enough....
An NOW, these people that have been bullying us all along have invented a system that makes us keep our mouths shut... Just great...
Pills... must... take... pills...
Support a Europe-related section on Slashdot!
We've seen this before here.
A love beyond compare...
If the fans inside your CPU are silent you wouldn't notice if one failed, my machine is fairly quiet but i would notice if it booted without the CPU or PSU fan running.
Snowden and Manning are heroes.
This would be well worth it on 90mm tornados. Its pretty shitty having 4 of them right beside your bed trying to sleep while your linux e-penis uptime grows every night ;)
I don't think this industry has come remotely close to making a fan that works. It's sad when people need tubes running to the bathroom just to keep the GPU or CPU cool.
If you create a fan that doesn't need water and guarantees performance of a water cooler, I think it'd be a hit. I have never the gotten blue screen of death from a noisy fan. Look on any forum, people are not complaining about noise. People are whining about overheating...
Blood Donation
I don't know as much about noise cancellation as I would like, but I understand most of the concepts. Although the method described in the article certainly is very cool, I wonder if they couldn't get better results by redesigning the fan. It seems that the fan generates too much random noise. Is it possible to make a fan that has a more predictable noise source? It could even be a fan that is way noisier before noise cancellation...
Another thought on this is that you really shouldn't consider the fan alone. The G5 has a beautiful interior with a ton of fans. Its not terribly loud, however, because the airflow is well designed.
Now if they could just silence the kids that come to my house to tell me about Joseph Smith.
'Same speed C but faster'
Personally its them damn harddisks that piss me off.
If only someone could suppress the disk noise..
http://www.nytimes.com/2004/05/27/technology/circu its/27next.html?ex=1086235200&en=8355c13a3b4a8afa& ei=5062&source=GOOGLE
Yeah, this is much simpler than just making the fans quiet in the first place, right?
Wonderful. "Look, instead of paying an extra 50 cents for a higher quality quiet fan, you can use cheap fans and spend $25.00 in additional parts to make the computer quiet!"
*sigh*
Musicians would care
So would home theater machines
Listening to music on a PC would mean lower required volume
How about being to sleep with a PC on 24/7?
GPL Deconstructed
For a second, I thought that BYU was going to hire big, burly henchmen to "take care" of us Slashdotters...
The perfect sig is a lot like silence, only louder
Will the same technology also work on my girlfriend? She is like one of those movies that is just better on mute.
Evolution or ID?
Maybe we should look at the fan design first. When Brushless-DC fans were first introduced in the early 1980's I evaluated most of the major brands. Blade shape and contour were major contributors to noise, but by far the worst was mounting the fan up against the panel wall with holes or a grill. Running slow also reduces noise. One company's blade design removed the high pitch wind noise and just left a low pitch rumble that sounded quieter than it was. Centrifugal "squirrel-cage" fans were much quieter than axial fans. Sleeve bearings were a little quieter than ball bearings, but had a much shorter life and will "freeze-up" once the oil dissipates. (I actually had this happen to my old computer.)Ball bearings get louder over time, but you'll replace your computer before then.
Power supplies can reduce airflow requirements considerably by better heatsinks and/or using the chassis for moving the heat away from the hot components. Once the real design issues are tackled, the bell-and-whistles approach could then be used to further reduce sound levels as necessary.
Dell mounts one fan deep inside the computer and the PS fan is quiet but near a wall. Six weeks ago I bought a Systemax and a Dell computer. Systemax sounded normally obnoxiously loud. Dell was so quiet, I thought it was not working, so I opened it to see if I could tell what was wrong. I was fooled by the sudden start of noise and then quiet.
I work in a computer music and acoustics research lab and we're always after a quieter PC. We've considered a solutions like this, but we've decided it wouldn't really be necessary for long. Here's why.
Among the many reasons for having a hard drive in every computer, two of the big ones were the Microsoft vision statement, and the fact that the network was much slower than disk. The latter is no longer the case.
The fact that network is now faster than local disk is a MAJOR development.
We've experimented with RedHat 9 with nfs root on older hardware with no disk and no fans, with 100Mb bootable NICs. We found to our surprise that they ran faster than with standard (non UDMA) ide. So, we're trying it now with newer hardware and gigabit, and some BIG heatsinks. So far, so good. We can optimize the central storage for speed, and the systems do, in fact, run noticeably faster in most cases, in addition to being nearly* silent.
We hadn't counted on the added bonuses, but there are many. We can change an entire system disk by moving dirs, reexporting, and booting the machine up. Poof, new system. We can install and uninstall packages on machines while they're off! We no longer have two or three extra gigs on each machine, all our nfsroots are from a single physical filesystem (so far) so they all have the same amount of free space, much more efficient! And if a machine offends you, you can yank the plug out. No local fsck!
*Note that the machine is never truly silent. Without any fans or disks, you can still hear a certain noise that sounds like it's happening when the disk used to seek. It's the toroids in the power supply! The network traffic causes HF noise in the power lines, which is filtered in the power supply and causes the chokes to vibrate slightly. The noise is very low, it would easily be drowned out by the quietest of fans, but in a totally silent room with no other PC sound, it's quite audible. There is also some low and infrequent clicking while the machine is warming up and cooling down, due to the thermal expansion of the heat sinks. This doesn't happen during use, when the temperature is more or less constant.
I'm supposed to document all this and I've been lazy, so if you want the rundown on booting redhat 9 without a hard drive, write to me and I'll finish the page and send you the link.
Obviously you never studied acoustics.
Noise-cancellation as implemented by these systems relies on matching an equal in amplitude, but opposite in polarity, waveform, to the incoming acoustic wave. By combining a compression wherever there's a rarefaction and a rarefaction wherever there's a compression, you wind up with blissful silence. However, the nature of these systems dictates that the interference only happens at specific places; where the waveforms match exactly. If you could place your cancelling radiator at the same location in space as the unwanted radiator, with the exact same radiation characteristics, it would be great, but you can't. Instead, you get cancellations at certain locations and intensification of the noise at other locations.
Basically, due to acoustics, getting closer to the noise won't do you as much good as getting closer to the receiver of the noise. This is why NR headphones work great, and aren't hard to do, but NR for open environments is hard to do and doesn't work very well. No matter what you do, unless you can colocate the cancelling radiator, you will always make some parts of the freespace environment worse.
---
Mod me down, you fucking twits. Go ahead. I dare you.
(I read with sigs off.)
Why can't they replace the bearings with magnets? And shape the fan blades to avoid the wasteful noise entirely, at least in the audible band?
--
make install -not war