Generate more than 120dB of ultrasonic sound?

[bunger]

The end result would be to create an amplified "dog repellent" using the Arduino. To do this, I would need to generate sound in the ultrasonic range ( greater than 20kHz ) and louder than 120dB. Unfortunately, I have no idea where to start.

Once I get past that, I would like to investigate sending those signals down a wire ( think perimeter wire ).

Any thoughts or suggestions would be greatly appreciated!!

[adafruit_support_mike]

Here's a reference circuit based on a 555 timer:

http://electroschematics.com/1026/dog-r ... t-circuit/

In general, you have two issues: generating a signal higher than 20kHz, and projecting that sound at the volume you want. With an Arduino, the first is just a matter of timing. An Arduino can't put out enough power to do the second at all, so you'll need an audio amplifier. Any of the (many) portable amps based on the LM386 should do the job, because the '386's gain/frequency curve is flat up to about 100kHz.

As for setting up a perimeter wire, it's either 'no problem' or 'impossible' depending on what you mean. If you just want to put some distance between the Arduino that generates the signal and the speakers that pump out the sound, you just need some wire. If you want a 'wall of sound' all the way around a patch of ground, you'll either need a whole lot of speakers or another universe where forcefields are possible.

The 'electric fence' products out there work by putting a speaker in the dog's collar along with a short-range radio receiver. The buried wire is basically a long antenna carrying a weak radio signal, and the collar emits sound when it gets close enough to the antenna to pick up the signal.

[john444]

M Stone,

Any of the (many) portable amps based on the LM386 should do the job

The LM386 may be a little weak in the volume-capability department. 120-dB loudness is on the order of a jet engine backwash.

Bunger,

I would need to generate sound in the ultrasonic range ( greater than 20kHz ) and louder than 120dB.

Ultrasonics of that volume also causes significant distress even to those unable to 'hear' it. In addition, it is not a trivial nor inexpensive task to generate audio of that magnitude.
But, good luck anyway.

[adafruit_support_mike]

The LM386 may be a little weak in the volume-capability department. 120-dB loudness is on the order of a jet engine backwash.

It's an edge case.. the LM386 can supply a Watt of power through a 32 ohm speaker, and 120db of sound corresponds to about a Watt of power. Push comes to shove, you could double them up.

I also think 120db is more power than necessary, given that the dog whistles I looked at spec 110db. Decibels are a logarithmic scale, so I can't see any reason to increase the noise by a factor of 10. 110db of sound (lawn mower loudness) is about a tenth of a Watt of power, which the LM386 can do.

[john444]

M Stone,

My mistake. When you indicated 120-dB, I thought you were referring to audio volume levels instead of electrical watts.
See: http://www.industrialnoisecontrol.com/c ... amples.htm

In the case you site, the audio volume produced by the speaker is very dependent on the speaker efficiency which is low. In your example, 1-watt electrical power to the speaker will be just like what you can get from a set of desk-top computer speakers. Computer speakers sometimes use the LM386. Those speakers might be about 60-dB of audio volume. However, 60-dB is a long way from the 110-dB dog-whistle.

I am reminded of this analogy: Incandescent light bulbs are rated in watts. We have come to think of light intensity (lumins) in terms of watts. But incandescent bulbs are very inefficient light producers.
It is kind of the same thing where the LM386 is rated in watts but a common speaker cannot come anywhere close to 1-W of acoustic power.

Sorry, John

[adafruit_support_mike]

My mistake. When you indicated 120-dB, I thought you were referring to audio volume levels instead of electrical watts.

It's not really your fault.

Decibels are, by definition, a ratio scale. "120 decibels" isn't an absolute quantity, it's a statement that X is a trillion times larger than Y. For sound, the reference volume (0db) is a sound pressure of 20 micropascals, but nobody actually knows that, so confusion reigns supreme.

Also, there's no difference between electrical Watts and and the work-energy in a sound wave. It's all Joules-per-second in the long run. On that score, you're right that I assumed a speaker would convert eletrical power to sound pressure with 100% efficiency, which is utterly wrong and highly dependent on the speaker. My bad there.

To make things even more complicated, we have to take the inverse square law into account. Sound pressure drops with the square of the distance from the source, so a truly complete description would be something like: "120db above 20uPa at a distance of X meters from the source." An iPod's earbuds have a rated maximum output of 103db (some people claim they can reach 130db), but they're like, 1/2" away from your eardrum.

Touching back to the original specs, I still think 120db is massive overkill. The audio pain threshold is around 100db, which represents a hundred times less energy. 120db is into "momentary exposure can cause permanent hearing loss" territory.

[john444]

Bunger,
IMHO actually generating 120dB of audio or ultrasonic audio is inadvisable. It is not only fairly difficult to accomplish but would also be somewhat expensive. Not to mention, hazardous. This is a 'major-league' project and I would not encourage you to pursue it.

M Stone,

you're right that I assumed a speaker would convert eletrical power to sound pressure with 100% efficiency

I think we are on the same page here.

If Bunger wished to try the LM386 or even 100-W amplifiers, then we are playing in a different ball-park.
Or, to continue the analogy, this would be more like we are playing a game of catch in the front yard.
It might be an interesting to see if he could get the dog's attention. However, the sound level would be no where near 120dB.

John