DIY Active Noise Cancellation

[Al Skierkiewicz]

phrontist,
Yes it is possible for you to build noise cancelling headphones. What makes them expensive is all of the auto adaptive circuitry reguired for varying acoustic conditions. Effectively what you need to do is sample the noise, phase shift it so it arrives at your ear 180 degrees out of phase with the noise and at the same level and there you are. The auto circuitry samples the environment, determines what is noise and only amplifies/shifts that part of the spectrum with a little AGC thrown in for level adjusting. With DSP on the rise, it is really easy to get a good sample, analyze what is audio and correct for the rest. If you have a good idea of what the spectrum of noise is, then you can build a simple filter to just pass the noise to the phase shifter and amp. Throw in a tuning control and level control (knobs on the box) and you don't need to add the auto circuitry. Just tune out the noise as best you can.
So here is what you need...Mic preamp, tuneable filter, phase shifter, power amp. You should be able to put it together with parts from Digikey into a case that fits into your pocket. LM324 opamps should do nicely for the everything but the power amp. LM386 for power amps (you need two for stereo phones.) Have fun! There is a simple book "Op Amp Cookbook" that ought to be a big help.

[Andrew]

There used to be a big "hobbyist" electronics movement in this country. I have watched this steadily decline over the past several years. You may be able to find a hobbyist area on the internet and get plugged in.

There were two magazines (Popular Electronics & Electronics Something) which merged into Electronics Now about four years ago. These magazines published circuits especially for audio applications. If you can dig back into that archive, you'll find some audio circuits to do the various tasks that Al mentioned.

You need to start with a condensor microphone element (available from Radio Hut or from Digikey). A condensor mic takes sound waves and converts them into a change in capacitance. You need a preamplifier circuit to take this change in capacitance and convert it into a voltage. Good luck on finding such a circuit. There were many in the Popular Electronics. Since Pop Elec is referenced in Applied Science and Technology database, you can probably do a real lit search and find one.

You might be able to find a data sheet for the condensor microphone which gives you a circuit to preamplify it.

The output side is the driver is best done with an LM386. I'm pretty sure this chip is not obsolete. If you want to find the circuit to do this, look up the Data Sheet and Application Note for the LM386 on National Semiconductor's web site (if they haven't spun off this part of there business to another chip maker).

You can find the "in between" stuff by looking up the patent for the noise cancelling ear phones. Somewhere out there is a patent search database (on the .gov net). You can acquire patents from any public library. If they don't have them, they can get them through interlibrary loan.

In terms of getting everything working, if you don't have an oscilloscope, I would abandon ship right now.

The LM386 part of the puzzle is the easiest to get working. You can get the parts and wire it up in a couple of weeks. To test, you can take the output of any CD device (CDROM drive for instance) and connect that as your circuit input. Connect the output of your audio driver to a speaker. If you get sound (and the LM386 doesn't heat up and start a fire), then you're good to go. Otherwise, check your solder joints and connections!

[Adam Y.]

There used to be a big "hobbyist" electronics movement in this country. I have watched this steadily decline over the past several years. You may be able to find a hobbyist area on the internet and get plugged in.

Heheh. I caught the tail end of the electronics movement. I managed to get two of Forrest M. Simm's books from Radio Shack before the company discontinued them. I didn't understand them until I bought more electronics books. Im not too worried though. Servo and Nuts and Volts are still running strong and there are now new hobbies that Im reading about. In fact you might want to try out their message board at www.nutsvolts.com They range from Space Exploration to Robotics. I have to dig through my magazines but I remeber there was a headphones set that could cancel out noise for only twenty dollars.

[Al Skierkiewicz]

Originally Posted by Andrew

You need to start with a condensor microphone element (available from Radio Hut or from Digikey). A condensor mic takes sound waves and converts them into a change in capacitance.

You might be able to find a data sheet for the condensor microphone which gives you a circuit to preamplify it.

Andrew,
Not exactly...True condensor mics have built in circuitry that make the conversion from capacitance to voltage internally and they are very expensive ($2000-5000). Most mic elements that you find (cheap at Digikey $2-$4) will be electret condensor mics. These elements have a permanent electrostatic field and as the plates of the condensor (capacitor) vibrate with sound the element outputs a real tiny voltage. This voltage is at a very high impedance so there is an amplifier to lower the impedance and add some gain built right into the mic. For this reason it is necessary to add a little DC power to the mic. Electret spec sheets will have typical power supply wiring, but it is usually very simple, a resistor to power supply is usually all it takes. Follow up with a coupling capacitor to a preamp and you are all set. LM324 are quad op amps optimized for single ended power. You still have to add bias wiring but they are cheap and available and have reasonable gain and distortion.

[phrontist]

Excellent advice Mr.Skierkiewicz! I've been reading a bit about op-amp theory and this looks like an excellent project. Don't have an o-scope, but my physics theacher has a really nice one. Surely this will fill those otherwise wasted summer hours

[DanArgent]

Phrontist is just jealous of my new senn's. Although, the grados that he has are decent. I think this is a great project phrontist and I would love to hel pyou out in the shop. I could see D "the Team" geting into it too once we get the gear box out of the way.

On a side note, are you sure, that Condensers are $2000+ I work with a few sure condensers that I think were more in the range of $1000.

[Al Skierkiewicz]

Originally Posted by DanArgent

On a side note, are you sure, that Condensers are $2000+ I work with a few sure condensers that I think were more in the range of $1000.

Dan,
The Shure's are electret, are you using the SM81? That is an electret. True, non electret condensors will be from manufacturers like AKG, Sennheiser, and of course the Neuman U67 tube mic. Many of them will be in large cases to hold the electronics needed to convert the condensor output to a useable signal. Many mics are coming out that look like classic mics with big cases and heads that are electrets made to look like a classic. Remember that each mic has it's own distinctive sound and good audio people can tell a mic (model) by listening to it.

[Pat McCarthy]

Does anybody know if there is an active noise cancellation program you can download that uses your computer's mic and speakers/headphones? If so, could you post a link to the site were it can be found?

-Pat

[Bharat Nain]

Originally Posted by PCPat

Does anybody know if there is an active noise cancellation program you can download that uses your computer's mic and speakers/headphones? If so, could you post a link to the site were it can be found?

-Pat

I don't know how possible it is for a program to cancel noise, just because it has no input for what noise is going into our ear. It doesnt know what to cancel. I'd really be amazed to find a program that actually cancels noise.
-Bharat

[Al Skierkiewicz]

Techno et al,
There are all sorts of ways to detect noise and cancel it out or pull data from it. Modern DSP (digital signal processing) can do marvelous things. For just about one hundred dollars there are DSP add ons for ham radios that reduce noise 20-30 dB. In these applications, the DSP is not noise canceling as much as signal enhancing. Noise cancelling for me means multiple loud sounds only one of which you want to hear. You filter out the noise, invert it and add it back in out of phase to "cancel" the unwanted sounds.
Now there are other methods for pulling signals out of a mixed environment. NASA uses these techniques in order to communicate with distant objects like the Mars explorers. It is easy when you have an unlimited power supply or huge antenna array, to crank out enough power to overcome the losses of distance in radio communications. But when you are limited by power and size, you must come up with other ways to insure "no data radio" will hamper you project. By using these techniques, NASA is able to demodulate signals that are below the noise floor. For standard voice communications, you need about 35-40 dB signal to noise for the parties to be understood. If both of you know that a limited vocabulary will be used or are willing to repeat messages until understood, then less than 30 dB S/N is tolerable. For data and morse code (with no DSP) a few dB S/N is OK. When you get to the point where the signal is 10-20 dB below noise, you better have a powerful DSP set and high gain antennas to get the data through.

[TBNAudioEnginer]

Hello All,

Well, this is interesting, Robot Comps? I think thats what I've stumbled upon...(scratches head) Anyway, My name is Jonathan, I'm the Audio Engineer for a TV Station in Canton, Ohio. I've been having a phase interference problem with trying to use 4 omni-directional microphones (on an interview set). Thus, I've been doing constant research about phasing concepts. Along the way, I found this forum. You might find a google search of "noisebuster" to be useful.

[Al Skierkiewicz]

Jonathon,
There is no real way to counteract the phasing you are experiencing other than riding the mics. The phasing is occurring simply by the length of time sounds takes to travel. So as one person speaks, his voice is picked up by the closest mic first and then the next closest, etc. Since all the mic signals are mixed together you are hearing four different signals slightly out of time alignment with each other. At audio frequencies, these slight time differences convert the mixed signal into a comb filter(s). As with any notch filter, phase changes rapidly going into and out of the notch frequency and that is what you hear. The depth of the notch(es) is determined by the level of each of the contributing sources and the frequency is determined by the acoustic distance between the mics. Two mics at equal level can add in phase to only +6dB but out of phase can add to an infinite depth notch. (i.e. >85dB) If you need any further info, please PM me. I am a broadcast maintenance engineer for WTTW/Chicago with specialty in audio.