I'd like to build an ultrasonic rat deterrent with some power. To that end, I drove a 20 watt tweeter with a 9 volt square wave. At 4 ohms I expected a little over 2 amps and 18 watts. The tweeter died. I expect I was too close to the rating. But is this a reasonable thing to try? If I get a 100 watt tweeter should I expect to be able to push an ultrasonic square wave through it directly with, say, 12 volts = 3 amps = 36 watts?
Question is, do ultrasonic rat deterrents actually work? Do Ultrasonic Pest Repellers Work? | Terminix
Yeah, that's true. I bought one and the cat came in and laid down next to it. But ... commercial ones are low power. Would I live with a 60 cycle hum of 20 dB? Sure. Might even be soothing. But how about 120 dB? That would repel me. So, how about 20 or 30 watts of ultrasonic? Who knows?I have cameras to find out. And I have a rat that's climbing in my truck and is not going in the trap. This is a worthwhile experiment if I can drive a tweeter as I describe.
I've used a little Kemo oscillator to drive two KSN1005 knockoffs in hopes to repel dogs and protect cats. Dogs and cats paid little attention, there was a young UPS delivery guy who could hear it 50 feet away 😀
M161 Ultrasonic Power Cannon
L010 Piezo Loudspeaker
Kemo also has a relatively low priced piezo speaker which appears to have built-in impedance matching coil and rated 105dB 1M
L003 Piezo-tweeter approx. 8 Ohm 50 mm
If driving a piezo from an audio amp, then a stepup/matching transformer would be of merit as a real KSN1016 would be around 103dB sensitivity with 2.83v input.
With a big bunch of them in parallel and current limiting resistor so the load doesn't look like a dead short at UHF, a simple highpass network or none would suffice.
Good luck
Here's a couple of plots of the GRS KSN1005 knockoff
Looking for advice on adding a super tweeter to high efficiency mono speaker
Kemo sells pest control piezo speakers - dunno how good
M161 Ultrasonic Power Cannon
L010 Piezo Loudspeaker
Kemo also has a relatively low priced piezo speaker which appears to have built-in impedance matching coil and rated 105dB 1M
L003 Piezo-tweeter approx. 8 Ohm 50 mm
If driving a piezo from an audio amp, then a stepup/matching transformer would be of merit as a real KSN1016 would be around 103dB sensitivity with 2.83v input.
With a big bunch of them in parallel and current limiting resistor so the load doesn't look like a dead short at UHF, a simple highpass network or none would suffice.
Good luck
Here's a couple of plots of the GRS KSN1005 knockoff
Looking for advice on adding a super tweeter to high efficiency mono speaker
Kemo sells pest control piezo speakers - dunno how good
Last edited:
I'm not using an audio amp, and I hope to push 12 volts into it from a power transistor. The specs on the KSN1005 suggest 1 watt at 3 volts. That ain't gonna do it.
there are ultrasonic transducers; tweeters usually can't handle continuous high power; they aren't designed for that
- can't find a spec sheet for Motorola's KSN1005 piezo (~0.13uF), but looking at KSN1041 (~0.3uF), long term continuous power was rated around 15 volts.
The late DJK (Dennis Kleitch) posted some stuff on piezo use. I'm almost sure he commented on the original piezo's top voltage limits. The Chinese knockoffs may not be that good. Maybe a bank of them in series parallel would run off of your oscillator and play loud enough to discourage the boogers (?)
What does your circuit look like ?
Many pro tweeters don't have a lot of power handling plus won't reach 20K
Here's some piezo stuff - I know its not what you're looking for but
might be handy for some hi fi applications. Sometime, I'll try to gather
a lot of DJK's posts into one thread.
https://www.grc.com/acoustics/CTS_Piezo_Tweeter_AppNote.pdf
Re: piezo impedance and other.....
• Posted by djk (M) on May 28, 2002 at 02:58:28 In Reply to: Re: piezo impedance and other..... posted by mr.x on May 27, 2002 at 17:19:12:
What you need is a book that doesn't exist. Jon Risch has been threatening to write one, but doesn't have the time. Any book on speakers will have some filter theory in it. I hear the one by Alden at Radio Shack is OK. Weems is another popular author.
While anyone can follow a cookbook formula and build a crossover that works, ie makes noise, even experienced designers can come up short if you are looking for good sound and adequate power handling. It seems to be as much of an art as it is a science. You can have state of the art test equipment and a PHD EE and still make a bad sounding speaker that blows up at the blink of an eye.
But I would still encourage you to try your hand at it. Sometimes an amateur will come up with something that an experienced hand will 'know' is impossible.
As regards piezo tweeters. Used properly I think they're great. Print out the contents of the piezo link. Either return for credit (if you can) or throw away any KNS1176A tweeters and KSN1151A horns. The best for hi-fi are the 4Khz group, 0.13µF per element, KSN1005, KSN1016, KSN1056, and the dual element KSN1177, and the 2Khz group, 0.30µF per element, KSN1025, including the lightbulb protected KSN1141, KSN1165.
I use a 62R/2W in series with each 0.13µF element. The 0.30µF KSN1025 comes with an internal 30R/2W and the protected ones have a lightbulb, I frequently add a 15R~22R/2W in series with these. The added series resistor does a couple of things. The capacitive bender element 'rings' a bit at its self resonant frequency. The series resistor helps to reduce the Q (or sharpness) of this resonance. It also isolates the amplifier from this capacitance, some amps are unstable without it. It also rolls off the input power to the piezo above 20Khz. The piezo still looks like a 1000R load to the amplifier at 1Khz, that's where the 8R resistor comes in. Its main function is to present a load for the crossover network. For a 2K crossover a 25W part is fine, for a 4Khz crossover a 10W part is OK. The minimum crossover slope for the 2Khz piezo is 12dB/oct. The 4Khz piezo likes 18dB at 4Khz, 12dB at 7Khz, and 6dB at 10Khz. Select the piezo and slope that best matches the roll-off of your midrange driver.
You can stack the KSN1016 and KSN1025 vertically to increase sensitivity, or use a matching transformer. Use a concave array to reduce comb-filtering. A pair of stacked KSN1016 sound better than a single KSN1177 if you tilt their axis across each other even just a little bit. A matching transformer requires a Zobel, in the case of the piezo this is an inductor in series with a resistor. For home use the transformer matched KSN1016 can keep up with pair of horn loaded 15" in an Altec A4 or JBL 4550 type system.
The late DJK (Dennis Kleitch) posted some stuff on piezo use. I'm almost sure he commented on the original piezo's top voltage limits. The Chinese knockoffs may not be that good. Maybe a bank of them in series parallel would run off of your oscillator and play loud enough to discourage the boogers (?)
What does your circuit look like ?
Many pro tweeters don't have a lot of power handling plus won't reach 20K
Here's some piezo stuff - I know its not what you're looking for but
might be handy for some hi fi applications. Sometime, I'll try to gather
a lot of DJK's posts into one thread.
https://www.grc.com/acoustics/CTS_Piezo_Tweeter_AppNote.pdf
Re: piezo impedance and other.....
• Posted by djk (M) on May 28, 2002 at 02:58:28 In Reply to: Re: piezo impedance and other..... posted by mr.x on May 27, 2002 at 17:19:12:
What you need is a book that doesn't exist. Jon Risch has been threatening to write one, but doesn't have the time. Any book on speakers will have some filter theory in it. I hear the one by Alden at Radio Shack is OK. Weems is another popular author.
While anyone can follow a cookbook formula and build a crossover that works, ie makes noise, even experienced designers can come up short if you are looking for good sound and adequate power handling. It seems to be as much of an art as it is a science. You can have state of the art test equipment and a PHD EE and still make a bad sounding speaker that blows up at the blink of an eye.
But I would still encourage you to try your hand at it. Sometimes an amateur will come up with something that an experienced hand will 'know' is impossible.
As regards piezo tweeters. Used properly I think they're great. Print out the contents of the piezo link. Either return for credit (if you can) or throw away any KNS1176A tweeters and KSN1151A horns. The best for hi-fi are the 4Khz group, 0.13µF per element, KSN1005, KSN1016, KSN1056, and the dual element KSN1177, and the 2Khz group, 0.30µF per element, KSN1025, including the lightbulb protected KSN1141, KSN1165.
I use a 62R/2W in series with each 0.13µF element. The 0.30µF KSN1025 comes with an internal 30R/2W and the protected ones have a lightbulb, I frequently add a 15R~22R/2W in series with these. The added series resistor does a couple of things. The capacitive bender element 'rings' a bit at its self resonant frequency. The series resistor helps to reduce the Q (or sharpness) of this resonance. It also isolates the amplifier from this capacitance, some amps are unstable without it. It also rolls off the input power to the piezo above 20Khz. The piezo still looks like a 1000R load to the amplifier at 1Khz, that's where the 8R resistor comes in. Its main function is to present a load for the crossover network. For a 2K crossover a 25W part is fine, for a 4Khz crossover a 10W part is OK. The minimum crossover slope for the 2Khz piezo is 12dB/oct. The 4Khz piezo likes 18dB at 4Khz, 12dB at 7Khz, and 6dB at 10Khz. Select the piezo and slope that best matches the roll-off of your midrange driver.
You can stack the KSN1016 and KSN1025 vertically to increase sensitivity, or use a matching transformer. Use a concave array to reduce comb-filtering. A pair of stacked KSN1016 sound better than a single KSN1177 if you tilt their axis across each other even just a little bit. A matching transformer requires a Zobel, in the case of the piezo this is an inductor in series with a resistor. For home use the transformer matched KSN1016 can keep up with pair of horn loaded 15" in an Altec A4 or JBL 4550 type system.
9v square wave at what frequency? and was this a 0 .. 9 v square? 50% time at 9v DC ? Yeah, sounds deadly to a tweeter. Can you adjust your duty cycle on your square wave? Does it have to be square, can you use a triangle or sin to be much more gentle on the tweeter? Can you get an AC wave, +/- vs pulsing DC ?
The best way to get rid of rats is to kill them, and to not give other rats a reason to come around. This is easily done with a rat "trap" which you can buy at Home Depot for about three dollars. You probably have more than one rat, so get three or four traps. Continue killing rats until they're gone. They'll eat almost anything. So, put a little cheese and a couple of birdseeds in the hole. You'll have dead rats in the morning!
To deter additional rats from coming around, don't leave out anything which may be food for them.
Better yet, get the traps that 'overbalance' when the rat walks inside to eat the food (not poison), trapping them inside. They can't resist peanut butter.
Then you can release them into a local park or bush land where they will be a tasty treat to a bird, snake, fox or similar.
Then you can release them into a local park or bush land where they will be a tasty treat to a bird, snake, fox or similar.
Freddi-
Thanks for the long technical reply. I think most of it doesn't apply to my situation, since I am not reproducing music nor am I using an amplifier. My circuit is a single chip microprocessor programmed to toggle an output bit at whatever rate I choose. I can choose low frequency to test operation, or any operating frequency I choose up to and beyond 60,000 Hz.
The output bit feeds a power transistor (not unlike the transistor output of power amplifier perhaps) that connects the tweeter to a nine volt power supply. So I am switching zero to 9 volts directly into the 4 ohm tweeter. Any harmonic distortion in the piezo element is immaterial to me since it's only likely to make it less soothing to "those dirty rats". It's not clear to me why this would be substantially different from the "high" voltage output of a power amplifier, except that would operate in the linear range of the transistor and I'm just pushing it to rails. Maybe slamming it from rail to rail is destructive to the element, in which case .01 ufd easily turns it into a sawtooth. But I'm no audio engineer and that's why I'm posting here.
Thanks for the long technical reply. I think most of it doesn't apply to my situation, since I am not reproducing music nor am I using an amplifier. My circuit is a single chip microprocessor programmed to toggle an output bit at whatever rate I choose. I can choose low frequency to test operation, or any operating frequency I choose up to and beyond 60,000 Hz.
The output bit feeds a power transistor (not unlike the transistor output of power amplifier perhaps) that connects the tweeter to a nine volt power supply. So I am switching zero to 9 volts directly into the 4 ohm tweeter. Any harmonic distortion in the piezo element is immaterial to me since it's only likely to make it less soothing to "those dirty rats". It's not clear to me why this would be substantially different from the "high" voltage output of a power amplifier, except that would operate in the linear range of the transistor and I'm just pushing it to rails. Maybe slamming it from rail to rail is destructive to the element, in which case .01 ufd easily turns it into a sawtooth. But I'm no audio engineer and that's why I'm posting here.
As for trapping rats, in this field I am expert. I have trapped at least 100 rats in the past year using live and electric traps. I know exactly what bait works best on our dusky-footed wood rats, and I know how they respond to the traps because I have night vision cameras on the traps (and on the truck). That's how I know that there is at least one rat active now, that it's climbing into the truck, and that it won't go into the traps. A perfect solution would be to trap him. But I haven't been able to. It is for this reason that I have undertaken the challenge of discouraging him. And anyway, even if I am able to trap this one, there is an endless supply of them. They chewed wires in my last truck and left acorns and rat **** in my car. This one built a nest in the air intake, and every time I start the truck the ventilation fills the cab with rat musk. A "no rat zone" would be valuable to me.
Deterrent? Coyote urine. Works well in areas that have Coyotes. Rats have a habit of looking for accommodation elsewhere. No need for ultrasonic transducers that don't work.
As with ultrasonic deterrence, feedback on coyote urine is a very mixed bag. It seems to be mostly used for deer, and lots of people say it doesn't work at all. It's more expensive than a piezo loudspeaker, but I may try it if the current path fails. We certainly have coyotes. My rat camera has also revealed a red fox that walks under the truck.
If you switch between 0V and 9V with 50% duty cycle, you actually create a 4.5V amplitude square, which is 5W in 4 ohm, but you also get a 4.5V DC on the tweeter, which is another 5W at DC. To avoid that DC component, you may need a slightly more complex circuit to drive the tweeter. First option is to use a complementary driving stage (with N and P power transistors), and a capacitor in series with the tweeter. Another option could be to drive an H bridge configuration from the microcontroller, so you'll get rid of the DC without needing a capacitor, and you can reach 4x as much AC power from the same 9V supply.