Freesound Forums

I'm looking for a speaker or module, that I can control with the Pi, in order to generate sound frequencies between 10kHz and 60kHz.

I've been reading a lot about piezo buzzers that you can control connected to a GPIO lead (3.3v) and a ground lead. However these only seem to go up to around 20-25 kHz from what I can find. They also seem to be relatively quiet and I'm looking for something with more volume.

When I start searching for "ultrasonic transmitter" instead, I only get results that are modules for ultrasonic distance detectors. Or other "continuous sounding" buzzers that only operate at nominal frequencies of 40KHz.

Basically, what speaker and power supply does the ultrasonic transmitter of this product use? I would like only the ultrasonic transmitter piece of this product and want to control the frequency using the Pi.
https://smile.amazon.com/Ultrasonic-Repellent-Repeller-Squirrel-Chipmunk/dp/B01M6EBICB

Thank you!

Most ultrasonic transducers that have high acoustic output are limited to one frequency only, the designed resonant frequency. The cheap "buzzers" are simply a little slab of piezo crystal glued to a plastic or paper cone. If you try to drive them at a different frequency the output and power handling drastically falls.

How much acoustic power you need? Short distance or far throw? The AMT tweeter used in ADAM studio monitors is quoted as operating up to 50kHz. The Pi connected to an audio "chip" amplifier connected to a wide bandwidth tweeter would be a reasonable approach for a variable frequency ultrasonic setup. There are caveats of course, but it all depends on what you're trying to do.

I think dheming has come up with the only practical (Read affordable - off the shelf components) solution.

I agree there are caveats, which unfortunately make task more a lot more complicated.

Firstly it requires a bigger amplifier than ceramic transducers and will not work driven directly from a GPIO output.

Tweeters dome or ribbon AMT, are only maybe 10-20% efficient which means that the bulk of electrical energy put into them is turned into heat. The usual cause of failure of a tweeter is thermal. The coil or ribbon element heats up like an electric fire and eventually burns out like a fuse(See edit note*). (To visualise the problem, imagine holding a lit 15 Watt Pygmy light bulb in your hand. Ouch - HOT! That's only 15W of heat, but it's enough to make a tungsten element glow white hot! Think what 15w would do to an aluminium element of similar mass.)

An AMT tweeter (or any other) might be specified for a 20W or 100W amplifier, but it's not designed to handle continuous high level signals. In practice with a musical signal in a properly designed system, a 100W amplifier might typically only deliver 2 to 10 watts to the tweeter averaged over a second or two, with very short peaks up to 50 or 60 watts.

To reduce the chances of burning out the tweeter there are three things to watch.

1. Do not drive with ultrasonic continuously but in short bursts. For example, a 200ms burst every 2 seconds reduces the heating by 90% but should still squeak the sh** out of squirrels.

2. Guard against DC and all low frequency’s below, say, 10kHz with a high-pass filter. All this energy would be turned directly into heat.

3. Never drive the Tweeter directly with square waves, either from the output of the pi or due to clipping in the amp. This could be achieved with a low-pass filter set @ ~50-60kHz. A square wave will produce about 1.4 times more heat than a sine wave of equal peak magnitude.

I've not been able to find any Power Handling V frequency specs. for AMT tweeters, but I suspect they need to be strongly de-rated at the higher frequencies. The typical figures I've used are pure guess work, even though I'm confident the principles are sound, so no guarantees. You might have to blow up a tweeter or two to find the actual limits...

(About me - Before I retired and sold my shares I used to design the electronics & amps for Quested Studio Monitors but my experience is mainly with conventional dome tweeters more than AMTs)

Anyway, good luck with the project,

Wibby

EDIT note*.
A conventional tweeter with a voice coil will actually fail way before the voice coil blows like a fuse because the "glue" fixing the coil to it's former melts and burns. The voice coil distorts and rubs itself to death against the magnet. A ribbon or AMT tweeter wont fail in this way, but is still able to die because of thermal stress.

Okay, I was fooled. Buchanon is another spammer with cut and paste from other blogs with spam links added. I wasted quite a lot of time trying to be helpful with this one, as did dheming, bless him/her/them!

- https://freesound.org/forum/off-topic/40039/?page=1#post88572 Different spam link but same problem!

Wibby.

I thought it was weird that they never responded. Lame..

a (bad) haiku

Spam comes and goes
I have learned of AMT
Springy Diaphragm.

Thanks Wibby et al, I learned new stuff!

strangely_gnarled wrote:
Okay, I was fooled. Buchanon is another spammer

I am not a spammer, sorry for not responding sooner! I'm just trying to digest this stuff as I'm a complete audio noob.

strangely_gnarled wrote:
Firstly it requires a bigger amplifier than ceramic transducers and will not work driven directly from a GPIO output.

I'm still trying to understand the relationship between power/voltage, sound volume, and frequency. Frequency controls pitch, where as putting voltage into the system controls volume?

strangely_gnarled wrote:
To reduce the chances of burning out the tweeter there are three things to watch.

1. Do not drive with ultrasonic continuously but in short bursts. For example, a 200ms burst every 2 seconds reduces the heating by 90% but should still squeak the sh** out of squirrels.

2. Guard against DC and all low frequency’s below, say, 10kHz with a high-pass filter. All this energy would be turned directly into heat.

3. Never drive the Tweeter directly with square waves, either from the output of the pi or due to clipping in the amp. This could be achieved with a low-pass filter set @ ~50-60kHz. A square wave will produce about 1.4 times more heat than a sine wave of equal peak magnitude.

#1 Makes a lot of sense.
#2 I don't plan on using this at low frequencies like that, but I will read up on high-pass filters which I'm unfamiliar with.
#3 Just read up on this as well, and it also makes sense about square waves driving more heat.

Thank you all for the responses, I'm just trying to figure this all out myself as well. It sounds like I'll need a power source of some time and many more hours of reading.

dheming wrote:
How much acoustic power you need? Short distance or far throw? The AMT tweeter used in ADAM studio monitors is quoted as operating up to 50kHz. The Pi connected to an audio "chip" amplifier connected to a wide bandwidth tweeter would be a reasonable approach for a variable frequency ultrasonic setup. There are caveats of course, but it all depends on what you're trying to do.

I'm trying to scare off animals with a 25khz-50khz SHORT distance throw. Maybe 10 foot max.

My full apologies buchannon, and thanks for responding. it's nice to get feedback when one tries to help. I hope my posts in other threads explain why I jumped to conclusions, even if in this case I was clearly wrong.

To add a bit more info that might be helpful -

Speakers are designed for typical audio, which has a high dynamic range. That is, in order not to compress and distort the peaks a tweeter might need to handle, say, 100W peaks when the mean(average) power level is only 2 or 3 Watts. Your application doesn't need to drive those peaks, but it still requires the tweeter to handle the heat generated by the average signal level, so you are probably looking at an AMT that is rated between 20-100W, but an amplifier that only needs to deliver 1-2W (providing it is not driven too far into clipping). With a typical 8 Ohm speaker that would suggest a battery or power supply of 6v minimum. Up to 12v (or +/- 6v) would make the thing go four times louder. With a continuous sine wave a 12v peak signal would give about 9watts into 8ohms which is close to or over the limit for a 100watt tweeter, but with a pulsed signal say 200millisec on, 2seconds off this would come down to under 1watt which should be safe for the speaker but still very loud. You would have to look at the sensitivity tech spec of the tweeter to work out just how loud.

Again, my figures are "back of the envelope" but I hope they are a good pointer.

Good luck with your project, and I hope you keep us informed on progress.

Regards Wibby.

strangely_gnarled wrote:
it's nice to get feedback when one tries to help.

Yep, for sure. My fault for not replying right away. I went down a google rabbit hole with those terms and then just plain forgot to get back.

strangely_gnarled wrote:
an amplifier that only needs to deliver 1-2W (providing it is not driven too far into clipping). With a typical 8 Ohm speaker that would suggest a battery or power supply of 6v minimum.

6v minimum. Hmm.

I know the pi has a 5v output integrated into the board, I wonder if I could get away with that. I may order a cheap speaker/amp and try it out.

Thanks again for the suggestions!

An interesting variation on the standard, electromagnetic AMT is one built using piezopolymer plastic film. Taket of Japan utilizes this technology in all of their products. http://www.taket.jp/

The efficiency is relatively low, but the frequency response is purported to be greater than 150kHz. Their BATPRO 2 would likely be the most appropriate starting point.

dheming wrote:
The efficiency is relatively low, but the frequency response is purported to be greater than 150kHz. Their BATPRO 2 would likely be the most appropriate starting point.

buchannon wrote:

dheming wrote:
The efficiency is relatively low, but the frequency response is purported to be greater than 150kHz. Their BATPRO 2 would likely be the most appropriate starting point.

Wow, that looks perfect but it's hard to find one for sale online. Apparently they cost $499 USD as well. Wowza.

Taket is a very small company and I think they make everything in house. So you won't find their products being resold in many places. I've been wanting to try out their headphones for many years now.

As an aside, Audax once made a piezopolymer dome tweeter called the HD-3P. The gold-sputtered film was held in an elliptical dome shape by a pressurized chamber filled with nitrogen. The gas slowly leaked out over time on many units and treble performance would then suffer. Not really relevant to your needs here, but that unique tweeter is what started my fascination with piezopolymer audio transducers.

buchannon wrote:

strangely_gnarled wrote:
an amplifier that only needs to deliver 1-2W (providing it is not driven too far into clipping). With a typical 8 Ohm speaker that would suggest a battery or power supply of 6v minimum.

6v minimum. Hmm.

I know the pi has a 5v output integrated into the board, I wonder if I could get away with that. I may order a cheap speaker/amp and try it out.

My knowledge of the pi is limited to a quick skim on google search and Wikipedia, but it seems to me the cheapest way getting the signal you want is from the PMW ports - if they can be configured to work up to 60kHz. All the signal processing can be programed digitally saving external circuitry which could otherwise get complicated. Whichever port you use the pi will never give out more power than what you might expect from an ipod phone jack so the amplifier will be needed.

Wibby.

strangely_gnarled wrote:
the cheapest way getting the signal you want is from the PMW ports - if they can be configured to work up to 60kHz. All the signal processing can be programed digitally saving external circuitry which could otherwise get complicated. Whichever port you use the pi will never give out more power than what you might expect from an ipod phone jack so the amplifier will be needed.

Thanks.

Yes. The amplifier isn't just about volts, it's about amps. My semipro mixing desk will give out 28 volts but if I connected my speakers to it directly I wouldn't hear a squeak. At 28 volts my speakers would need nearly 4 amps to drive them properly and the mixer would balk at around 2 tenths of an amp. The speakers would simply short the mixers output and the 28v would drop down to nothing. It's like trying to turn the starter motor on a car with AA cell batteries.

Anyhoo, good luck buchannon,

Wibby.