Piezo Tweeter for Dogs

[vacuphile]

Cats like fish, that is where it comes from.

 

[restoremysanity]

Hello there, sorry to bump a new thread

What is being discussed in this thread is becoming a very common theme across the internet for more and more desperate people looking for peace.
I hope we can have a discussion without angering people or getting anyone banned. If I breach any rules, can someone please warn me.... so apologies in advance, on with the show....

I'm looking to build a hand based 'dog silencer' which will not permanently hurt the dog, but I definitely want it to damn well know about it. The dinky $30 hand ones on Amazon don't cut it.


I need to fire the signal 22 metres (65' maybe) I'll have 'kind of' line of sight (some trees but no direct obstacles)
I'd like to keep it under $200 but I want the damn thing to work

Would using 2 of these.
Piezo Siren Bargain - 120dB | Pro Sensor, Sirens & Strobes | Home & Commercial Alarms | Alarm & Asset Protection | PRODUCTS | LA5258 | Jaycar Electronics
plus 2 of these
Kemo L010 Ultraschall Piezo Lautsprecher Loudspeaker 1X | eBay
work?

I was intending to put them inside tubing to give me some directional ability, perhaps a pringles can and some PVC piping, all ducted together into a 4 barrel configuration. Power supply needs to be 'semi' portable, I was looking at a 12V DC motorbike battery 3Mah?

So where I need help is understanding if this is even capable of what I intend? Am I buying too big or too small? I don't want to do a half job and have it not work.
I've been told by another expert, it would be best if "tube size resonates at the frequency of the piezo transducer." How can I ensure this?
Do I need to install some kind of modulation dial to adjust the frequency? Do I need very particular length pipes? Etc



If you fast forward to 57:50 of this internet radio show, you'll see a gadget guy who knows his stuff who built a similar device (shorter range) many years ago.
https://twit.tv/shows/security-now/episodes/248

I would *really* appreciate some help on this, seriously - there's a market for something like this.
Thanks again, seriously thank you all for ANY help
♥

 

[phase_accurate]

In order to gain narrow directivity I would build an array rather than using a tube or the like. This way you will also gain SPL.
I don't know about the quality of this Kemo thingie. I guess you will have to try it by yourself.

Regards

Charles

 

[restoremysanity]

When you say an array, you mean multiple units of the same type side by side? Like some of these?
https://www.google.com.au/search?q=tweeter+array&hl=en&tbm=isch&gws_rd=cr&ei=j2U8V7SyM6OMmgX_xpvwDw

I thought the tube would help with direction but you guys know better than I, would 4 not constitute a reasonable array?

Thanks again

 

[nofar679]

Sorry to bump this old thread, but has anyone made any progress with this? I am desperate for this device.

 

[freddi]

what do you want to do? - I have a little kemo module directly driving a couple of ksn1005 knockoffs - its pretty loud where I can hear - up above that, dogs and cats seem to ignore the noise. A young UPS worker heard if from pretty far (!)

 

[Caramello]

This old thread interests me, as I have been living next to a neighbour with several yappy dogs, which have access to the garden through a large flap in the door. I get on with the neighbour, but they have no control over the yapping, and even though I get on with them and mentioned the yapping many times, their attempts to control the yapping are inconsistent. Inconsistency towards stopping anything to do with dogs has no effect. However, without going into details, when the neighbours are not in, if the dogs come out, I have conditioned them to quickly go back in when I bang the fence (if you want detail, send me a PM).

But, for the last year, I have been thinking of trying ultra-sound with a remote, to save me having to bang the fence, and I'm hopeful that it will also work when the neighbours are there.

The starting point for calculations is the data sheet for a suitable driver. There are several mentioned in this thread, either piezo or conventional coil.

I am thinking of going for the Eminence APT-80 , which has the best combination of SPL at 1W and RMS power input of 32W . Whilst the sound emissions will be brief, it is best to stick to the RMS power to avoid distortion.

The datasheet goes to 20 kHz, 103 dB @1W @ 1metre. Each doubling of power adds a further 3 dB, so 32W adds 15 dB, taking it to 118 dB @ 32W @ 1 metre. But sound falls away at 6dB for each doubling of distance, so 112 dB @ 2 metres, 106 dB at 4 metres, 100 dB at 8 metres. If I use a conventional stereo amplifier and buy 2 of the APT-80 drivers (one per channel), that doubles up, which adds another 3 dB, making it 103 dB at 20 kHz at 8 metres.

I have an old Quad 405 power amp, and an old Gould signal generator.

Before starting, I would like to buy a 'calibrated' microphone for measuring SPL up to 25 kHz. Any ideas anyone ?

I'd also want to make a crossover/filter to put between the Quad 405 and the APT-80, any ideas how to start on that ?

 

[Pano]

A piezo is likely to play higher/louder than the APT-80. Most affordable calibrated mics are calibrated only up to ~20K. You will get signal from them above that, but you won't be sure of the level.
Crossover? If it's just tones above 10K stick a 2.2uF in front of the driver and be done with it.

 

[Caramello]

There are good datasheets on 3 KSN piezos

KSN 1005A http://146.88.69.205/piezo/files/KSN-1001A-1005A-Datasheet.pdf
KSN 1038A http://146.88.69.205/piezo/files/KSN-1038A-Datasheet.pdf
KSN 1165A http://146.88.69.205/piezo/files/KSN-1165A-Datasheet.pdf

The 1165A is probably the one that is best and also probably easier to get.

From the datasheet
Sensitivity: 93 dB at 1m/1W [10 dB lower than the APT-80]
Maximum Power Handling Capacity: 400 W (EIA RS426) 8 Ohm system reference
[ EIA RS426 is a continuous test (not peak) ]
Typical Impedance: Appears as a 0.3 µF capacitor in series with a 30 Ohm resistor

400 W is about 28dB gain on 1W, so total output at 400 W = 121 dB, only 3 dB more than the APT-80 at only 32W

Unless I'm mistaken, the problem is that getting that much power into the piezo will require either very large amplitude voltages from the amplifier output, or a complex crossover ?

I'm also puzzled by "8 Ohm system reference" ? Does that mean when it's in a system with a crossover that presents 8 Ohms to the amplifier ?

 

[Caramello]

^ slight error in the above, should read

400 W is about 26dB gain on 1W, so total output at 400 W = 93 + 26 = 119 dB, only 1 dB more than the APT-80 at only 32W

Considering the marginal difference, I think that driving the coil APT-80 should be easier from a conventional power amplifier.

Out of interest, how would one get a conventional power amplifier to drive a KSN 1165 piezo ?

Also, I'll settle for a calibrated microphone up to 20kHz, and ideas on where to get one ?

 

[Aridace]

Long ago I was faced with a similar problem. The piezo horn was fried in tests but a tweeter with voice coil fared better and even didn't have a problem with 22 KHz square waves (easier to generate). More important was the delay between "bark, bark!" and human response: dogs "didn't get the message". So I used a mike with an amp fitted with BPF to detect "bark, bark!" and that was fast enough to make dogs immediately think of a cause-effect relationship.

 

[Caramello]

That makes sense, I might have to get to that stage.

I'll be taking it one step at a time, first step (after getting the tweeters and capacitor) is to check the SPL output at low powers indoors, by using the signal generator at various frequencies. Hence the need for a calibrated microphone.

I'll also be checking the power input as the impedance may vary. To do that, I intend to get a high precision 1 Ohm resistor and connect in series, and put an oscilloscope across the resistor to see the current output at various frequencies.

 

[Pano]

Calibrated microphone. In the USA from Parts Express. They have several. I even have the one that plugs into a smart phone.

 

[Caramello]

In the USA from Parts Express. They have several. I even have the one that plugs into a smart phone.

Have you got a link or a part number , then I can either find out delivery to UK, or search for an equivalent in the UK

 

[Pano]

From the datasheet
Sensitivity: 93 dB at 1m/1W [10 dB lower than the APT-80]
Maximum Power Handling Capacity: 400 W (EIA RS426) 8 Ohm system reference

But look at the frequency response plots. The APT is down lower than the piezo near 20K. From the looks of it you'd need more voltage for the APT. I do agree that it will be easier to drive than the piezo. The 405 amp does OK into capacitive loads, it was designed to, so that should be OK of you did go that way.

Interesting about the delay from bark to ultrasound. That's useful feedback.

 

[Aridace]

SPL is a big issue indeed, as high frequencies don't propagate well in air. Tweeters often are directional too. Power loss due to propagation can be overcome by concentrating power in a certain direction but then you would need more mikes, to get the direction for the tweeter and a control system aiming it. That would make it a project for graduation ;-)

 

[Caramello]

It's a tough choice, and this sort of dialogue is very helpful.

To discuss the issues some more, I've taken screen shots
- for the KSN 1165 from here http://146.88.69.205/piezo/files/KSN-1165A-Datasheet.pdf
- for the APT 80 from here https://www.eminence.com/pdf/APT80.pdf and here https://www.eminence.com/pdf/Eminence_HF_Components.pdf

Frequency response of KSN 1165, showing SPL 93/94 dB @ 20kHz

Off-axis response of KSN1165, shows SPL 93 dB @ 20 kHz @ 0 degrees, falling away rapidly past 15 degrees @ 20 kHz

Frequency response of APT 80, showing SPL 103 dB @ 20 kHz and impedance about 9 Ohms @ 20 kHz

Same as above

Interesting description of how the measurements of the APT-80 were made

Without an off-axis respnse for the APT-80, it is feasible that it achieves the 10 dB higher 1W SPL by having a sharper focus (in RF antennas, this is called the forward gain, which is usually achieved by making the vertical response much sharper). However, it could be that a coil is more efficient at moving air than a piezo ?

Either way, they both achieve the same SPL at their full power, but since I will be using a Quad 405 amplifier, I can only achieve 100 W RMS. Pertinent spec of the Quad 405 is 100W RMS into 8 Ohms, implying working voltage of 80 volts peak-to-peak, which I've verified in the past by looking at the voltage across Rogers Studio speakers with signals from the Gould sig gen. Power rails are +/-50 volts, and output does start to clip as voltage across load reaches 80 pk-to-pk.

Unless and until I find that the APT-80 has a much sharper off-axis roll-off, I think that it should be easier to set up.

I've googled "calibrated microphone" and found the Dayton Audio iMM-6 Calibrated Measurement Microphone for iPhone iPad Tablet and Android Dayton Audio iMM-6 Calibrated Measurement Microphone for iPhone iPad Tablet and Android: Amazon.co.uk: Musical Instruments

So before committing to buying any high output tweeters, I need to set up a microphone and make measurements of the Quad and Rogers, using the Gould sig gen. When I'm confident that this setup works, I can then make the choice between the two tweeters, but presently, the APT-80 looks better with the available data.

 

[Pano]

Good comparison. The graphs I saw for the APT-80 showed lower output than yours, or so I thought. I'll go with you graphs.

That Dayton mic is the one I have. Works well but be careful not to lose it! I've lost 2.

 

[Caramello]

Yep that first APT:80 graph (curve in black) is easy to mis-read, the 105 on the left axis looks like a 100, until you zoom in really close. The second graph (curve in red) is much clearer, and the link where that is from has all the specs. In the link there are also some line array waveguides that bolt onto the driver, which should give a wider horiontal dispersion with narrower vertical dispersion, maybe even some "gain" in SPL.

I'll only be using the microphone in and outside the house, so it should be ok. This will be the first time that I've ever started to measure audio

 

[weltersys]

I
1)Frequency response of KSN 1165, showing SPL 93/94 dB @ 20kHz

2)Off-axis response of KSN1165, shows SPL 93 dB @ 20 kHz @ 0 degrees, falling away rapidly past 15 degrees @ 20 kHz

3)Frequency response of APT 80, showing SPL 103 dB @ 20 kHz and impedance about 9 Ohms @ 20 kHz

4)Interesting description of how the measurements of the APT-80 were made

5)Without an off-axis respnse for the APT-80, it is feasible that it achieves the 10 dB higher 1W SPL by having a sharper focus (in RF antennas, this is called the forward gain, which is usually achieved by making the vertical response much sharper). However, it could be that a coil is more efficient at moving air than a piezo ?

6)Either way, they both achieve the same SPL at their full power, but since I will be using a Quad 405 amplifier, I can only achieve 100 W RMS. Pertinent spec of the Quad 405 is 100W RMS into 8 Ohms, implying working voltage of 80 volts peak-to-peak, which I've verified in the past by looking at the voltage across Rogers Studio speakers with signals from the Gould sig gen. Power rails are +/-50 volts, and output does start to clip as voltage across load reaches 80 pk-to-pk.

7)Unless and until I find that the APT-80 has a much sharper off-axis roll-off, I think that it should be easier to set up.

Caramello,

I have used the APT-80 in a number of loudspeakers. Decent driver for the price, but I don't think the UHF output would be near that of an array of KSN 1165.
I would not use KSN 1165 for any "Hi-Fi" use, but for annoyance factor, they rate very "high" ;^).

1) Since the KSN 1165 is around 30-50 ohms at 20kHz, around 8 volts would be needed for "1 watt", the "93dB sensitivity" is more equivalent to 100 dB. Using 8 drivers would increase on axis sensitivity by around 9dB, near 109 dB.
Some good information here about similar piezo drivers:
Pulsar Developments Ltd - Piezo Tweeter Application Note

2) The relatively high VHF sensitivity is due to narrowing directivity.

3)All of the dozen or so APT-80 drivers I owned had rapidly falling response past 17kHz.
The look more similar to the APT50 measurement on the H290H horn, which has around 95dB sensitivity at 20kHz after the 17kHz peak.
The measurements below were of a APT-80 with no crossover (raw) and of two different crossover designs used in stage monitors.

4) Fairly typical, though other manufacturers "sensitivity" may use 2.83 volts for lower impedance drivers. As mentioned in #1, Eminence would use a much higher voltage to measure a driver with equivalent impedance to the KSN 1165.

5) The APT-80 VHF dispersion is similar to the KSN 1165, the throat angle of the APT:50driver itself being roughly the dispersion at 20kHz. In fact, the horn used on the driver will have no effect on the VHF response, for use at 20kHz the phase plug assembly virtually is the horn.

6) The APT-80 coil will burn if used much more than it's rated 35 watts, about 17.75 volts RMS at 20kHz. The voice coil heating will raise it's impedance considerably at that drive level, reducing sensitivity, "power compression".

7) For wide VHF coverage, an array of either driver would be needed.

Cheers,
Art