Need ultrasonic sound

[Ken]

I have a pest repeller that can drive an external speaker. It has
three separate settings: rodents, birds and dogs. I want to repel
rodents (squirrels in my attic, to be precise).

Where can I find a speaker that will not cost an arm and a leg that
will put out a strong, modulated ultrasonic signal? My guess is we
are talking 5 watts at and around 35 kHz.

In the alternative, where can I find a piezo buzzer that will put out
a big 35 kHz sound when DC is applied?




Ken
(to reply via email
remove "zz" from address)

 

[Bob Masta]

On Tue, 27 Apr 2004 16:12:22 GMT, Ken <cprstn54zz@att.net> wrote:

I have a pest repeller that can drive an external speaker. It has
three separate settings: rodents, birds and dogs. I want to repel
rodents (squirrels in my attic, to be precise).

Where can I find a speaker that will not cost an arm and a leg that
will put out a strong, modulated ultrasonic signal? My guess is we
are talking 5 watts at and around 35 kHz.

In the alternative, where can I find a piezo buzzer that will put out
a big 35 kHz sound when DC is applied?

Ultrasonic signals are commonly used as stimuli
for auditory experiments with rodents. We used to
use piezo horn tweeters, those plastic jobs about
3-4 inches across. They were made by Motorola,
and we bought them at Radio Shack.

Note that you will not get anything like flat
frequency response. We measured the response
(a trick in itself at these frequencies) and tried to
work in the "hottest" regions. But they were
cheap and fairly powerful. (Flat response
pretty much requires electrostatic drivers. These
are not only really expensive, but because they
have a big, flat radiating surface they tend to
beam the sound in a tight pattern.)

Note, however, that all these "pest repeller"
gadgets are long on extravagant claims and
really short on hard evidence. I've never heard
of an independent study that shows they work.

Hope this helps!





Bob Masta
dqatechATdaqartaDOTcom

D A Q A R T A
Data AcQuisition And Real-Time Analysis
www.daqarta.com

 

[Terry Pinnell]

NoSpam@daqarta.com (Bob Masta) wrote:

Ultrasonic signals are commonly used as stimuli
for auditory experiments with rodents. We used to
use piezo horn tweeters, those plastic jobs about
3-4 inches across. They were made by Motorola,
and we bought them at Radio Shack.

Note that you will not get anything like flat
frequency response. We measured the response
(a trick in itself at these frequencies) and tried to
work in the "hottest" regions. But they were
cheap and fairly powerful. (Flat response
pretty much requires electrostatic drivers. These
are not only really expensive, but because they
have a big, flat radiating surface they tend to
beam the sound in a tight pattern.)

Note, however, that all these "pest repeller"
gadgets are long on extravagant claims and
really short on hard evidence. I've never heard
of an independent study that shows they work.

Any observations on their effectiveness as dog detterents/calmers, as
in the Dazer?
http://www.dazer.com/dog-deterrent.jsp

I'm still thinking of getting one, following my failure to build an
effective equivalent. My design is shown at
http://www.terrypin.dial.pipex.com/Images/DogDeterrent.gif
and I was quite pleased with it initially because its audible setting
(one way of the 2-way switch) gave an ear-splitting sound with a fresh
pair of PP3 NiMH batteries. I therefore (rashly) assumed its
ultrasonic setting would have a suitable effect on any savage dogs I
encountered while walking in remote areas.

Happily, unlike the previous year in Tuscany, I had no cause to use it
in anger when I took it on a walking holiday in Poland in 2002. But
curiosity had to be satisfied, so I discretely tried it as I strolled
through the crowded streets of Zakopane. No visible effect on the
unwitting victims whatsoever. Even a poodle at 2 metres apparently
remained unaware of it. In fact, I convinced myself after a
few such abortive experiments that it was broken or that the batteries
had failed. But switching fleetingly to the 'human' setting made that
look very unlikely.

I reckon it's just too low-powered. Which is why I'm keen to hear from
any Dazer users. On that walking holiday in Tuscany (when I was
attacked by two dogs, prompting the project originally), another
member of the group demonstrated the Dazer to me (on a barking
Alsatian behind gates). It definitely calmed it down and made it back
off, so its design is clearly superior to mine. Almost certainly down
to power output, I assume.

Terry Pinnell
Hobbyist, West Sussex, UK

 

[Soeren]

Hi Terry,

curiosity had to be satisfied, so I discretely tried it as I strolled
through the crowded streets of Zakopane. No visible effect on the
unwitting victims whatsoever. Even a poodle at 2 metres apparently
remained unaware of it. In fact, I convinced myself after a
few such abortive experiments that it was broken or that the batteries
had failed. But switching fleetingly to the 'human' setting made that
look very unlikely.

Which frequency was it ?

Years ago I made a small ultrasounder from a piezo-disk (about 2kHz
natural resonance) which was very ineffective in terms of power output,
but my parents dog (in lieu of a real ginnea-pig) nevertheless reacted
with a surprised listening attitude (if you know what I mean).

I reckon it's just too low-powered.

About what level ?

Perhaps you tested on a deaf dog

I guess anything around a couple of Watts woud be plenty, if used
efficiently (the right piezo-tweeter and the right frequency being the
main concerns).

The non-electronic solution would be a small can of mace to ensure your
hiking enjoyment


--
Regards,
Soeren

* If it puzzles you dear... Reverse engineer *

 

[Terry Pinnell]

Soeren <Look@iNO-SPAMt.dk.invalid> wrote:

Hi Terry,


curiosity had to be satisfied, so I discretely tried it as I strolled
through the crowded streets of Zakopane. No visible effect on the
unwitting victims whatsoever. Even a poodle at 2 metres apparently
remained unaware of it. In fact, I convinced myself after a
few such abortive experiments that it was broken or that the batteries
had failed. But switching fleetingly to the 'human' setting made that
look very unlikely.

Which frequency was it ?

I guess you didn't look at the circuit? As stated in the notes, it's a
swept range 24 - 30 kHz.

Years ago I made a small ultrasounder from a piezo-disk (about 2kHz
natural resonance) which was very ineffective in terms of power output,
but my parents dog (in lieu of a real ginnea-pig) nevertheless reacted
with a surprised listening attitude (if you know what I mean).


I reckon it's just too low-powered.

About what level ?

As per notes, 15V pk-pk into the piezo tweeter. Power? Wish I knew.
The tweeter impedance is specified as 1k at 1 kHz and 20 ohms at 40
kHz. So suppose it's say an average of 40 ohms at my output
frequencies. Then, unless I've slipped somewhere, approximating with
((Vrms)^2)/R implies a very low power of about 700mW. Which would
explain the pathetic results. But using the same calculation, the
power output of the human-audible signal would be at a greatly *lower*
level, of the order of 60 mW! That just doesn't square with the
ear-splitting shriek I actually hear. Resonance effects presumably
account for some of this? But how would you determine the power
output?

Perhaps you tested on a deaf dog

I guess anything around a couple of Watts woud be plenty, if used
efficiently (the right piezo-tweeter and the right frequency being the
main concerns).

The non-electronic solution would be a small can of mace to ensure your
hiking enjoyment

Regrettably illegal in UK, and probably in rest of EU by now <g>.

 

[Soeren]

Hi Terry,

I guess you didn't look at the circuit? As stated in the notes, it's a
swept range 24 - 30 kHz.

Yes. Sorry for my laziness, but I am looking at it now

As per notes, 15V pk-pk into the piezo tweeter. Power? Wish I knew.
The tweeter impedance is specified as 1k at 1 kHz and 20 ohms at 40
kHz. So suppose it's say an average of 40 ohms at my output

A linear regression on these numbers says 270 to 422 Ohms (at your
frequency range). In reality, the curve will not be all that linear, but
at least it gives a clue.

frequencies. Then, unless I've slipped somewhere, approximating with
((Vrms)^2)/R implies a very low power of about 700mW. Which would
explain the pathetic results.

I guess it is just a _drawing_ error that pin 2 is connected to pin 7 ?

The RMS of a 15V_pp 50% squarewave is 7V5, 7V5^2/40 is 1W4 (at least in
my country

With 270 to 422 Ohms however, yeilds only 133mW to 208mW, so I am not
all that surprised of your results.

But using the same calculation, the
power output of the human-audible signal would be at a greatly *lower*
level, of the order of 60 mW! That just doesn't square with the
ear-splitting shriek I actually hear. Resonance effects presumably
account for some of this? But how would you determine the power
output?

I think resonance is accountable for the major part of the difference.
Most piezo disks of around 1˝" diameter I have played around with has
had their resonance around 2kHz, what is the diameter of your tweeter ?
There are tweeters with good as well as bad hi freq. responce (and
luckily the cheaper ones might be better for your purpose although they
might be a bit shreaky when used for hi-fi.

To get a real blast from a piezo, you need to increase the V_pp fed to
it. A Hi-Q LC-circuit in resonance at the frequency you want, should
enable you to get some real power, 5W to 6W should be reasonably easy to
get.
It will rule out the sweep, of course, but perhaps an intermittent tone
of a single frequency is just as effective (test with low power a couple
of dogs for the best frequency).

Regrettably illegal in UK, and probably in rest of EU by now <g>.

Sure, but so is letting your mongrels attack people

Anyway, nobody will find out until you are attacked, so which duo you
prefer... Hospital or a fine ?

Besides, it might give you peace of mind, as being attacked by dogs can
give you a real fobia towards those filthy bastards (been there
myself as a kid)... But... unprovoked dog attacks are _extremely_ rare,
so perhaps you could get some advice from an animal psykologist about
behaviour around dogs. In most cases of threatening dogs, a bit of
understanding of their mind is a very good help...

By nature, dogs want to dominate, not kill, so if you are not able to
dominate them, submissive behaviour will get you minimal damage (so lomg
as we are not talking of a trained dog gone whacko), at least a couple
of scientists studying animal behavour have verified it the hard way, by
lying on their back and letting the dog _pretend_ to bite their throat
(they must have huge balls to do that, personally I would rather kick
the dogs than test mine that way

Hope you get a solution you are comfortable with, so you can hike
without worrying


--
Regards,
Soeren

* If it puzzles you dear... Reverse engineer *

 

[Terry Pinnell]

Soeren <Look@iNO-SPAMt.dk.invalid> wrote:

Hi Terry,


I guess you didn't look at the circuit? As stated in the notes, it's a
swept range 24 - 30 kHz.

Yes. Sorry for my laziness, but I am looking at it now


As per notes, 15V pk-pk into the piezo tweeter. Power? Wish I knew.
The tweeter impedance is specified as 1k at 1 kHz and 20 ohms at 40
kHz. So suppose it's say an average of 40 ohms at my output

A linear regression on these numbers says 270 to 422 Ohms (at your
frequency range). In reality, the curve will not be all that linear, but
at least it gives a clue.

Can't put my hands on the specs now, but I recall it was heavily
skewed, so my figure would be much closer than yours.

frequencies. Then, unless I've slipped somewhere, approximating with
((Vrms)^2)/R implies a very low power of about 700mW. Which would
explain the pathetic results.

I guess it is just a _drawing_ error that pin 2 is connected to pin 7 ?

No, it's your optician's error - pin 2 is correctly connected to pin
SIX!

The RMS of a 15V_pp 50% squarewave is 7V5, 7V5^2/40 is 1W4 (at least in
my country

I think it was closer to a sine than a square wave. But let's
compromise on say 1W for 40 ohms, and correspondingly less for higher
impedances. Anyway, bottom line is that it was plainly inadequate in
practice!

With 270 to 422 Ohms however, yeilds only 133mW to 208mW, so I am not
all that surprised of your results.


But using the same calculation, the
power output of the human-audible signal would be at a greatly *lower*
level, of the order of 60 mW! That just doesn't square with the
ear-splitting shriek I actually hear. Resonance effects presumably
account for some of this? But how would you determine the power
output?

I think resonance is accountable for the major part of the difference.
Most piezo disks of around 1˝" diameter I have played around with has
had their resonance around 2kHz, what is the diameter of your tweeter ?

As you can estimate from the 3.5mm jack socket in the picture I
provided, that's pretty well full size (on a 17" display), so I
suppose tweeter was about 6 cm diameter.

There are tweeters with good as well as bad hi freq. responce (and
luckily the cheaper ones might be better for your purpose although they
might be a bit shreaky when used for hi-fi.

FWIW, here's the audible output.
http://www.terrypin.dial.pipex.com/Images/DoqAudible.wav

To get a real blast from a piezo, you need to increase the V_pp fed to
it. A Hi-Q LC-circuit in resonance at the frequency you want, should
enable you to get some real power, 5W to 6W should be reasonably easy to
get.

Assuming my 40 ohm guess is about right, and that wave is somewhere
between square and sine, that would probably mean a voltage supply of
around 45-50V, compared with my present 18V (2 X PP3). Seems
impractical. A DC-DC step-up converter would add complexity and reduce
efficiency.

It will rule out the sweep, of course, but perhaps an intermittent tone
of a single frequency is just as effective (test with low power a couple
of dogs for the best frequency).


Regrettably illegal in UK, and probably in rest of EU by now <g>.

Sure, but so is letting your mongrels attack people

Anyway, nobody will find out until you are attacked, so which duo you
prefer... Hospital or a fine ?

Besides, it might give you peace of mind, as being attacked by dogs can
give you a real fobia towards those filthy bastards (been there
myself as a kid)... But... unprovoked dog attacks are _extremely_ rare,
so perhaps you could get some advice from an animal psykologist about
behaviour around dogs. In most cases of threatening dogs, a bit of
understanding of their mind is a very good help...

By nature, dogs want to dominate, not kill, so if you are not able to
dominate them, submissive behaviour will get you minimal damage (so lomg
as we are not talking of a trained dog gone whacko), at least a couple
of scientists studying animal behavour have verified it the hard way, by
lying on their back and letting the dog _pretend_ to bite their throat
(they must have huge balls to do that, personally I would rather kick
the dogs than test mine that way

Hope you get a solution you are comfortable with, so you can hike
without worrying

Thanks. Meanwhile, I've taken the line of least resistance, and
ordered a Dazer from the site I mentioned <g>.

--
Terry Pinnell
Hobbyist, West Sussex, UK

 

[Terry Pinnell]

Soeren <Look@iNO-SPAMt.dk.invalid> wrote:

Hi Terry,


Can't put my hands on the specs now, but I recall it was heavily
skewed, so my figure would be much closer than yours.

OK.


I guess it is just a _drawing_ error that pin 2 is connected to pin 7
?

No, it's your optician's error - pin 2 is correctly connected to pin
SIX!

I just complained to my optician, but he also claims that pin 2 is
connected to pin 7 - on the right 555 in the drawing
(Or to pin 6 through an 82k if you like).

Apologies to you and your optician - I checked only the left 555!

I think it was closer to a sine than a square wave. But let's
compromise on say 1W for 40 ohms, and correspondingly less for higher
impedances. Anyway, bottom line is that it was plainly inadequate in
practice!

OK.


As you can estimate from the 3.5mm jack socket in the picture I
provided, that's pretty well full size (on a 17" display), so I
suppose tweeter was about 6 cm diameter.

OK.


FWIW, here's the audible output.
http://www.terrypin.dial.pipex.com/Images/DoqAudible.wav

URL:http://www.terrypin.dial.pipex.com/Images/DogAudible.wav> works
better.
Hard to judge the actual sound pressure from a recording, but I guess it
would annoy a dog as weel as a person ?


To get a real blast from a piezo, you need to increase the V_pp fed
to it. A Hi-Q LC-circuit in resonance at the frequency you want,
should enable you to get some real power, 5W to 6W should be
reasonably easy to get.

Assuming my 40 ohm guess is about right, and that wave is somewhere
between square and sine, that would probably mean a voltage supply of
around 45-50V, compared with my present 18V (2 X PP3). Seems
impractical. A DC-DC step-up converter would add complexity and
reduce efficiency.

No step-up is needed.
The piezo is a capacitor, so all you would have to add, is a coil with a
tap.
+---+ <--- *NOT* connected to Vcc !
| |
) |
L1a ) |
Out ) O
from O--* Piezo Tweeter.
555 ) O
L1b ) |
) |
| |
Gnd O---+---+

The value of the coil will have to be calculated to get resonance at the
frequency you want and from the piezos capacity at that frequency.
The Q of the coil and the impedance of the piezo will determine the
voltage you get.

Thanks. I'll try that soon.

Thanks. Meanwhile, I've taken the line of least resistance, and
ordered a Dazer from the site I mentioned <g>.

Yeah, outsourcing is the key

Now I need to scour the neighbourhood for unwitting strays, or lie in
wait for the regular dog-walkers...<g>

--
Terry Pinnell
Hobbyist, West Sussex, UK

 

[Soeren]

Hi Terry,

Can't put my hands on the specs now, but I recall it was heavily
skewed, so my figure would be much closer than yours.

OK.

I guess it is just a _drawing_ error that pin 2 is connected to pin 7
?

No, it's your optician's error - pin 2 is correctly connected to pin
SIX!

I just complained to my optician, but he also claims that pin 2 is
connected to pin 7 - on the right 555 in the drawing
(Or to pin 6 through an 82k if you like).

I think it was closer to a sine than a square wave. But let's
compromise on say 1W for 40 ohms, and correspondingly less for higher
impedances. Anyway, bottom line is that it was plainly inadequate in
practice!

OK.

As you can estimate from the 3.5mm jack socket in the picture I
provided, that's pretty well full size (on a 17" display), so I
suppose tweeter was about 6 cm diameter.

OK.

FWIW, here's the audible output.
http://www.terrypin.dial.pipex.com/Images/DoqAudible.wav

<URL:http://www.terrypin.dial.pipex.com/Images/DogAudible.wav> works
better.
Hard to judge the actual sound pressure from a recording, but I guess it
would annoy a dog as weel as a person ?

To get a real blast from a piezo, you need to increase the V_pp fed
to it. A Hi-Q LC-circuit in resonance at the frequency you want,
should enable you to get some real power, 5W to 6W should be
reasonably easy to get.

Assuming my 40 ohm guess is about right, and that wave is somewhere
between square and sine, that would probably mean a voltage supply of
around 45-50V, compared with my present 18V (2 X PP3). Seems
impractical. A DC-DC step-up converter would add complexity and
reduce efficiency.

No step-up is needed.
The piezo is a capacitor, so all you would have to add, is a coil with a
tap.
+---+ <--- *NOT* connected to Vcc !
| |
) |
L1a ) |
Out ) O
from O--* Piezo Tweeter.
555 ) O
L1b ) |
) |
| |
Gnd O---+---+

The value of the coil will have to be calculated to get resonance at the
frequency you want and from the piezos capacity at that frequency.
The Q of the coil and the impedance of the piezo will determine the
voltage you get.

Thanks. Meanwhile, I've taken the line of least resistance, and
ordered a Dazer from the site I mentioned <g>.

Yeah, outsourcing is the key


--
Regards,
Soeren

* If it puzzles you dear... Reverse engineer *

 

[Soeren]

Hi Terry,

Thanks. I'll try that soon.

You are welcome

Now I need to scour the neighbourhood for unwitting strays, or lie in
wait for the regular dog-walkers...<g

Place it behind a hedge and use a PIR-sensor to trigger it for a few
seconds when someone walks by. If you see a lot of people hanging after
stampeeding dogs, you know you have a winner


--
Regards,
Soeren

* If it puzzles you dear... Reverse engineer *

 

[Terry Pinnell]

Soeren <Look@iNO-SPAMt.dk.invalid> wrote:

I think resonance is accountable for the major part of the difference.
Most piezo disks of around 1˝" diameter I have played around with has
had their resonance around 2kHz, what is the diameter of your tweeter ?
There are tweeters with good as well as bad hi freq. responce (and
luckily the cheaper ones might be better for your purpose although they
might be a bit shreaky when used for hi-fi.

To get a real blast from a piezo, you need to increase the V_pp fed to
it. A Hi-Q LC-circuit in resonance at the frequency you want, should
enable you to get some real power, 5W to 6W should be reasonably easy to
get.
It will rule out the sweep, of course, but perhaps an intermittent tone
of a single frequency is just as effective (test with low power a couple
of dogs for the best frequency).

I opened up my Dazer. It includes a miniature transfomer; presumably
an audio-type, but the only identification I can see is the number
'2145'. The very simple PCB has a couple of 'KSP2222A' transistors,
which I'm guessing are equivalent to 2N2222A, and a TIP110 medium
power transistor. There's also a small 94V zener. The piezo tweeter
itself is only 1" diameter, and the markings I can see are
'MASSA-HINGH' and 'TR89...'

The contrast in power with my own DIY circuit is apparent from this
shot of the output waveform, which shows a frequency of about 25 kHz
and a pk-pk amplitude of 85V.
http://www.terrypin.dial.pipex.com/Images/Dazer1.gif

Anyone care to suggest the likely circuit approach from all that? I'm
guessing a two-transistor multivibrator drives the transformer primary
via the power transistor, with output from the secondary to the
tweeter. Would the zener be directly across that?

And, given the pk-pk of 85V, what other factors would be needed to
estimate the actual output power please?

--
Terry Pinnell
Hobbyist, West Sussex, UK

 

[Robert C Monsen]

"Terry Pinnell" <terrypin@dial.pipex.com> wrote in message
news:ka2n90513n1ednnceq8eimoen01ot4dniu@4ax.com...

Soeren <Look@iNO-SPAMt.dk.invalid> wrote:

I think resonance is accountable for the major part of the
difference.
Most piezo disks of around 1˝" diameter I have played around with
has
had their resonance around 2kHz, what is the diameter of your
tweeter ?
There are tweeters with good as well as bad hi freq. responce (and
luckily the cheaper ones might be better for your purpose although
they
might be a bit shreaky when used for hi-fi.

To get a real blast from a piezo, you need to increase the V_pp fed
to
it. A Hi-Q LC-circuit in resonance at the frequency you want,
should
enable you to get some real power, 5W to 6W should be reasonably
easy to
get.
It will rule out the sweep, of course, but perhaps an intermittent
tone
of a single frequency is just as effective (test with low power a
couple
of dogs for the best frequency).

I opened up my Dazer. It includes a miniature transfomer; presumably
an audio-type, but the only identification I can see is the number
'2145'. The very simple PCB has a couple of 'KSP2222A' transistors,
which I'm guessing are equivalent to 2N2222A, and a TIP110 medium
power transistor. There's also a small 94V zener. The piezo tweeter
itself is only 1" diameter, and the markings I can see are
'MASSA-HINGH' and 'TR89...'

The contrast in power with my own DIY circuit is apparent from this
shot of the output waveform, which shows a frequency of about 25 kHz
and a pk-pk amplitude of 85V.
http://www.terrypin.dial.pipex.com/Images/Dazer1.gif

Anyone care to suggest the likely circuit approach from all that?
I'm
guessing a two-transistor multivibrator drives the transformer
primary
via the power transistor, with output from the secondary to the
tweeter. Would the zener be directly across that?

And, given the pk-pk of 85V, what other factors would be needed to
estimate the actual output power please?

--
Terry Pinnell
Hobbyist, West Sussex, UK

Use an inductor and a power mosfet to kick the voltage up. put your
transducer in parallel with the inductor.

Regards