Hello !
I am interested in a small diy solution in order to transmit a high frequency sound to a long distance (approx. 50-70m).
My first thought was, well that is not going to be that hard, you have a good amplifier and speakers, find an app that creates a 20-25khz signal and done (yes, as others in here, i try to solve a dog barking issue unfortunately), but that seems not to be the case.
Reading several threads here, i found out that my solution are not normal coil speakers but piezos in combination with a ultrasonic generator (like Kemo M048N Ultrasonic Generator) . As my knowledge on electronics is really limited, except of the basics about speakers/amps wattage/impedance etc, i got lost in the information found here. Like the need of a resistor or capacitor that i read in some topics. On this topic i would like your help. Allow me to ask a few things and excuse my ignorance on this topic.
1) I got an old 2x100 Watt 4ohm amplifier. If i would connect the signal generator and the correct piezo speaker on this amplifier, would i end up with the results i want or i will fry my amp?
2) If the first option with my amplifier is not possible, what is alternative simple solution for this purpose ? How else can i drive the piezo speakers to output in such high volumes ?
3) What specs to look on the piezo speakers for this scenario? Any recommendations ?
4) Is the Kemo M048N Ultrasonic Generator fine to create those frequencies ?
If possible, give me some details of what i will need after all to buy. Is there a simple explanation in this forum of how to connect all those components together? This would be really helpful.
Thank you!
I am interested in a small diy solution in order to transmit a high frequency sound to a long distance (approx. 50-70m).
My first thought was, well that is not going to be that hard, you have a good amplifier and speakers, find an app that creates a 20-25khz signal and done (yes, as others in here, i try to solve a dog barking issue unfortunately), but that seems not to be the case.
Reading several threads here, i found out that my solution are not normal coil speakers but piezos in combination with a ultrasonic generator (like Kemo M048N Ultrasonic Generator) . As my knowledge on electronics is really limited, except of the basics about speakers/amps wattage/impedance etc, i got lost in the information found here. Like the need of a resistor or capacitor that i read in some topics. On this topic i would like your help. Allow me to ask a few things and excuse my ignorance on this topic.
1) I got an old 2x100 Watt 4ohm amplifier. If i would connect the signal generator and the correct piezo speaker on this amplifier, would i end up with the results i want or i will fry my amp?
2) If the first option with my amplifier is not possible, what is alternative simple solution for this purpose ? How else can i drive the piezo speakers to output in such high volumes ?
3) What specs to look on the piezo speakers for this scenario? Any recommendations ?
4) Is the Kemo M048N Ultrasonic Generator fine to create those frequencies ?
If possible, give me some details of what i will need after all to buy. Is there a simple explanation in this forum of how to connect all those components together? This would be really helpful.
Thank you!
The primary question you haven't addressed is the SPL required at 50-70 meters. How large an area you have to cover is another relevant data point. Whether the signal you want to use is even effective is also important - that would be the first one I'd try to answer before building something extravagant.
The Kemo M048 is made to drive up to 5 piezo speakers directly. Their datasheet doesn't say what the output voltage is. You might need to decrease its level if it overdrives the input of the amplifier. A cheap oscilloscope should answer that question.
Datasheet for M048
https://static.rapidonline.com/pdf/130243_an_ml_02.pdf
Resistor additions to piezo circuits are often to make them sound better for audio purposes or to improve their behavior with crossovers and response shaping circuits. Capacitors on their own with piezos don't do what they do with a normal speaker, so they wouldn't normally be used without additional resistors (unless you are just trying to decrease level - which you aren't interested in for this application). Typical piezo speakers can be connected directly to most transistor amplifiers without trouble as long as the capacitance is reasonable (that is, you aren't trying to drive something weird like an industrial transducer or a bunch of piezo tweeters). If you want to power a large array of piezos or push things hard, adding some resistance may make the load more friendly for the amplifier and protect the piezos. Some transistor amps are less stable into capacitively reactive loads, so there's always a bit of variability there.
The Kemo M048 is made to drive up to 5 piezo speakers directly. Their datasheet doesn't say what the output voltage is. You might need to decrease its level if it overdrives the input of the amplifier. A cheap oscilloscope should answer that question.
Datasheet for M048
https://static.rapidonline.com/pdf/130243_an_ml_02.pdf
Resistor additions to piezo circuits are often to make them sound better for audio purposes or to improve their behavior with crossovers and response shaping circuits. Capacitors on their own with piezos don't do what they do with a normal speaker, so they wouldn't normally be used without additional resistors (unless you are just trying to decrease level - which you aren't interested in for this application). Typical piezo speakers can be connected directly to most transistor amplifiers without trouble as long as the capacitance is reasonable (that is, you aren't trying to drive something weird like an industrial transducer or a bunch of piezo tweeters). If you want to power a large array of piezos or push things hard, adding some resistance may make the load more friendly for the amplifier and protect the piezos. Some transistor amps are less stable into capacitively reactive loads, so there's always a bit of variability there.
Last edited:
mattstat, thank you so much for your input.
You have a good point here ! I would assume that 95-100db would be sufficient for an area of about 10 meters or wider.
that is some great information. now the youtube video i saw makes also sense ! I guess then i should not worry about those components as i am not interested in audio quality.
No, surely i don't plan for any big array. 1 or 2 speakers max.
You have a good point here ! I would assume that 95-100db would be sufficient for an area of about 10 meters or wider.
that is some great information. now the youtube video i saw makes also sense ! I guess then i should not worry about those components as i am not interested in audio quality.
No, surely i don't plan for any big array. 1 or 2 speakers max.
Achieving your wanted SPL at that distance with just 1 or 2 typical piezo speakers doesn't seem feasible.
Normal speaker voltage sensitivity is rated at 1 meter and 2.83 volts (often just listed as SPL on spec sheets). If we extrapolate distance to 70 meters and use 3 dB spreading loss per distance doubling, that's 10*log(1/70), which is about -18.5 dB for the distance difference. More on this assumption later.
Unfortunately, you're also likely to lose as much or more to atmospheric absorption at 20 kHz.
https://www.frontierlabs.com.au/post/sound-attenuation-through-the-air
Google suggests an atmospheric absorption loss of 0.4-1.3 dB per meter, but we'll assume an 18.5 dB absorption loss because it makes the math easier and the speaker counts something achievable. At the upper end of Google's suggested absorption range, your scenario seems impossible if my math is correct.
Your 100 watt amp should be about 20 dB above 1 watt (same 10*log equation from earlier is used for two different powers).
So in round numbers, your tweeters will need to give you 117 dB at 2.83 volts and 1 meter and be able to take the full amp power to give the output you want.
117 dB base sensitivity + 20 dB for amp at full power = 137 dB at 1 meter. -18.5 db for distance and another -18.5 dB for absorption = 100 dB at 70 meters best case.
These kinds of numbers aren't too crazy. If you have 25 tweeters rated at 90 dB you can get there. In 95 dB tweeters, you'd need 13 of them. But many piezo tweeters are rated for 50 watts, which is another issue.
So far I've been assuming piezos don't fall off at 6 dB per distance doubling. This could be true due to the high frequency and the horn loading typical of piezos, but I wouldn't count on it without measuring. If you wind up with 6 dB loss per distance doubling, the spreading equation changes to 20*log(1/70), which is -36.9 dB, and combined with absorption loss would make the required voltage sensitivity at 1 meter about 135 dB to achieve 100 dB at 70 meters at full amplifier power. That would take about a hundred 95 dB sensitive tweeters.
Calculator for coherent sources:
https://sengpielaudio.com/calculator-coherentsources.htm
Normal speaker voltage sensitivity is rated at 1 meter and 2.83 volts (often just listed as SPL on spec sheets). If we extrapolate distance to 70 meters and use 3 dB spreading loss per distance doubling, that's 10*log(1/70), which is about -18.5 dB for the distance difference. More on this assumption later.
Unfortunately, you're also likely to lose as much or more to atmospheric absorption at 20 kHz.
https://www.frontierlabs.com.au/post/sound-attenuation-through-the-air
Google suggests an atmospheric absorption loss of 0.4-1.3 dB per meter, but we'll assume an 18.5 dB absorption loss because it makes the math easier and the speaker counts something achievable. At the upper end of Google's suggested absorption range, your scenario seems impossible if my math is correct.
Your 100 watt amp should be about 20 dB above 1 watt (same 10*log equation from earlier is used for two different powers).
So in round numbers, your tweeters will need to give you 117 dB at 2.83 volts and 1 meter and be able to take the full amp power to give the output you want.
117 dB base sensitivity + 20 dB for amp at full power = 137 dB at 1 meter. -18.5 db for distance and another -18.5 dB for absorption = 100 dB at 70 meters best case.
These kinds of numbers aren't too crazy. If you have 25 tweeters rated at 90 dB you can get there. In 95 dB tweeters, you'd need 13 of them. But many piezo tweeters are rated for 50 watts, which is another issue.
So far I've been assuming piezos don't fall off at 6 dB per distance doubling. This could be true due to the high frequency and the horn loading typical of piezos, but I wouldn't count on it without measuring. If you wind up with 6 dB loss per distance doubling, the spreading equation changes to 20*log(1/70), which is -36.9 dB, and combined with absorption loss would make the required voltage sensitivity at 1 meter about 135 dB to achieve 100 dB at 70 meters at full amplifier power. That would take about a hundred 95 dB sensitive tweeters.
Calculator for coherent sources:
https://sengpielaudio.com/calculator-coherentsources.htm
Last edited:
ChatGPT is getting better at these kinds of things, so I also asked it: given typical speaker spreading loss and atmospheric absorption at 20 kHz, is it possible to achieve 100 dB of ultrasonic output at 70 meters?
Its math and logic were similar to my last post, but it used the more pessimistic estimate for spreading loss and an even more pessimistic 1.5 dB per meter for absorption. This led it to the conclusion that: "Achieving 100 dB of ultrasonic output at 70 meters at 20 kHz is not feasible with typical speaker technology, as the required source level would exceed the capabilities of any current sound generation equipment."
I then asked it if a horn would help, and it said: "Without any horn, the spreading loss at 70 meters would be 39 dB (from the inverse square law). A well-designed horn can improve the directivity, possibly reducing the loss by 10-20 dB, depending on the size and design, meaning the spreading loss could be reduced to something closer to 20-30 dB instead of 39 dB."
. . .
"The absorption is still going to be significant due to the high frequency, which is a fundamental physical limitation."
. . .
"Conclusion:
A horn-loaded piezoelectric transducer can definitely reduce the spreading loss, and you might see a notable improvement in the sound pressure level at a given distance. However, even with a horn, achieving 100 dB at 70 meters at 20 kHz would still be very difficult due to the combination of spreading loss and atmospheric absorption. The horn helps, but the required source level would still be impractically high to overcome the losses."
Its math and logic were similar to my last post, but it used the more pessimistic estimate for spreading loss and an even more pessimistic 1.5 dB per meter for absorption. This led it to the conclusion that: "Achieving 100 dB of ultrasonic output at 70 meters at 20 kHz is not feasible with typical speaker technology, as the required source level would exceed the capabilities of any current sound generation equipment."
I then asked it if a horn would help, and it said: "Without any horn, the spreading loss at 70 meters would be 39 dB (from the inverse square law). A well-designed horn can improve the directivity, possibly reducing the loss by 10-20 dB, depending on the size and design, meaning the spreading loss could be reduced to something closer to 20-30 dB instead of 39 dB."
. . .
"The absorption is still going to be significant due to the high frequency, which is a fundamental physical limitation."
. . .
"Conclusion:
A horn-loaded piezoelectric transducer can definitely reduce the spreading loss, and you might see a notable improvement in the sound pressure level at a given distance. However, even with a horn, achieving 100 dB at 70 meters at 20 kHz would still be very difficult due to the combination of spreading loss and atmospheric absorption. The horn helps, but the required source level would still be impractically high to overcome the losses."
mattstat thank you so much for your detailed answered. No matter if this diy project is more complicated than expected and its goals are not so easy to achieve, i surely learned many things today from your comments. Really appreciate your effort.
i'd still like to ask a more generic question. We previously discussed about 100db at 70m. What could be a common SPL lets say at 1m for simple portable ultrasonic anti-barking devices found on the market ? If for example such a device delivers 50db at 5 meters (lets assume those 5 meters as max range for this device to still be effective), then maybe i don't need 100db at 70m, but only 50db to match the specs a smaller device on the market
Many devices sold for this purpose don't specify the SPL. ChatGPT says: "certain high-end devices might list the SPL in the range of 100–130 dB at the device's effective range, but these details are often left somewhat vague or generalized." I tried to get specific values from it, but it made up answers based on its effectiveness assumptions instead of using actual sources. So the AI Overlord still has some issues on specifics at this time.
A few data points from patents:
https://patents.google.com/patent/WO2014011195A1/en
"There are several commercially available dog deterrent devices such as the DAZER (TM) and DAZER Π (TM), manufactured by Dazer International, Peterborough, England and distributed by K-II Enterprises of Syracuse, NY, emit high frequency "ultrasonic" sound bursts of 25 kHz (25,000 Hz) at a decibel level of 115 dB (SPL) at 0.5 meter reference. The DAZER devices are advertised as having an effective range of about 15 feet and increasing effectiveness as the dog becomes closer to the sound source. Most "ultrasonic" dog deterrent devices provide high frequency "ultrasonic" sound burst to produce a startle effect to interrupt a dog's barking or the dog's aggressive behaviour, and their deterrent effectiveness depends upon the dog's reaction to the strange sound. However, such devices are not effective on all dogs. For example, the dog's reaction depends upon its age, breed, mental state, temperament, health and intelligence. For instance, smaller dogs react better to "ultrasonic" noise, whereas medium to large sized dogs are more responsive to "audible" blasts of sound."
https://patents.google.com/patent/US5061918A/en
"Another anti-bark device available in the prior art is one invented by the inventor of the present invention and marketed by Humane Technology, Inc., College Station, Texas, under the trademark "PeaceMaker". The PeaceMaker utilizes CMOS integrated circuits and piezoelectric crystal elements to produce a single high-intensity burst of sound with a decrescendo pattern shortly after detection of a dog's bark. Vocalization discrimination circuitry distinguishes between a dog's barking and extraneous sounds of brief duration. The PeaceMaker is small enough to be fitted onto a dog's collar, and most of the high-frequency sound produced is inaudible to humans."
. . .
"A device for discouraging the barking of dogs designed in accordance with the present invention would produce approximately 116-dB of sound at the dog's ear"
------------------
Comment from @cowanaudio in a related thread
www.diyaudio.com/community/threads/building-a-powerful-ultrasonic-dog-deterrent-which-works-behind-wooden-fence.389425/page-2
"I've actually had a bit of experience using ultrasound to train a neighbour's dog, and learnt a few things along the way.
The signal doesn't have to be all that loud, I achieved around 90dB at the dogs ears. A sine wave only works for a brief time. I guess either the dog starts to ignore the signal or goes deaf over the small frequency range. In the end I clipped a mains freq signal, fed it through a differentiator then high passed above 20KHz. The engage switch was in the master bedroom of the In-Laws house. The neglected german shepherd was retrained within a couple of days. It might be more difficult with a dumb dog. Just be consistent with hitting the noise button when the dog barks, at least for a few days."
------------------
I was looking at the above thread because I had a vague recollection of talking about delayed playback of the dog's barks as a deterrent. That's another approach that might work if you can't get enough ultrasonic output or if the dog just doesn't respond to it.
ChatGPT's suggestion on delay: "The time delay for a dog bark deterrent that plays back the dog's own barks typically ranges from 1 to 3 seconds. This delay allows the dog to associate the sound of its own barking with a mild form of correction or a deterrent. The idea is that the dog hears its own bark right after it occurs, which may confuse or interrupt the barking behavior.
Too short of a delay might not allow the dog to connect the sound with its action, while too long of a delay could reduce the effectiveness of the deterrent."
A few data points from patents:
https://patents.google.com/patent/WO2014011195A1/en
"There are several commercially available dog deterrent devices such as the DAZER (TM) and DAZER Π (TM), manufactured by Dazer International, Peterborough, England and distributed by K-II Enterprises of Syracuse, NY, emit high frequency "ultrasonic" sound bursts of 25 kHz (25,000 Hz) at a decibel level of 115 dB (SPL) at 0.5 meter reference. The DAZER devices are advertised as having an effective range of about 15 feet and increasing effectiveness as the dog becomes closer to the sound source. Most "ultrasonic" dog deterrent devices provide high frequency "ultrasonic" sound burst to produce a startle effect to interrupt a dog's barking or the dog's aggressive behaviour, and their deterrent effectiveness depends upon the dog's reaction to the strange sound. However, such devices are not effective on all dogs. For example, the dog's reaction depends upon its age, breed, mental state, temperament, health and intelligence. For instance, smaller dogs react better to "ultrasonic" noise, whereas medium to large sized dogs are more responsive to "audible" blasts of sound."
https://patents.google.com/patent/US5061918A/en
"Another anti-bark device available in the prior art is one invented by the inventor of the present invention and marketed by Humane Technology, Inc., College Station, Texas, under the trademark "PeaceMaker". The PeaceMaker utilizes CMOS integrated circuits and piezoelectric crystal elements to produce a single high-intensity burst of sound with a decrescendo pattern shortly after detection of a dog's bark. Vocalization discrimination circuitry distinguishes between a dog's barking and extraneous sounds of brief duration. The PeaceMaker is small enough to be fitted onto a dog's collar, and most of the high-frequency sound produced is inaudible to humans."
. . .
"A device for discouraging the barking of dogs designed in accordance with the present invention would produce approximately 116-dB of sound at the dog's ear"
------------------
Comment from @cowanaudio in a related thread
www.diyaudio.com/community/threads/building-a-powerful-ultrasonic-dog-deterrent-which-works-behind-wooden-fence.389425/page-2
"I've actually had a bit of experience using ultrasound to train a neighbour's dog, and learnt a few things along the way.
The signal doesn't have to be all that loud, I achieved around 90dB at the dogs ears. A sine wave only works for a brief time. I guess either the dog starts to ignore the signal or goes deaf over the small frequency range. In the end I clipped a mains freq signal, fed it through a differentiator then high passed above 20KHz. The engage switch was in the master bedroom of the In-Laws house. The neglected german shepherd was retrained within a couple of days. It might be more difficult with a dumb dog. Just be consistent with hitting the noise button when the dog barks, at least for a few days."
------------------
I was looking at the above thread because I had a vague recollection of talking about delayed playback of the dog's barks as a deterrent. That's another approach that might work if you can't get enough ultrasonic output or if the dog just doesn't respond to it.
ChatGPT's suggestion on delay: "The time delay for a dog bark deterrent that plays back the dog's own barks typically ranges from 1 to 3 seconds. This delay allows the dog to associate the sound of its own barking with a mild form of correction or a deterrent. The idea is that the dog hears its own bark right after it occurs, which may confuse or interrupt the barking behavior.
Too short of a delay might not allow the dog to connect the sound with its action, while too long of a delay could reduce the effectiveness of the deterrent."
Last edited: