Directing Ultra-Sonic Transmission to Limit Side-Spread

Factual Questions
1

OK, Let’s see where this one goes.

I have several ultrasonic generators (neighbor has chihuahua which likes to bark while I putter in my yard and especially when I try to sleep).

These things have curious quirk - if placed near each other, one will trigger the other - at a slight delay. The result is continuous 7/24/365 operation.
Even I won’t do that to a dog, even a nasty chihuahua.

Each has 2 transducers with very shallow cones.

Is there some readily available material that will reduce a 160 degree cone of sound to a cone of 90 degrees?

For those who remember, what I need is the (ultra) sonic equivalent of a horse’s blinders.
Hit the dog, not the device 10 feet away. Right now I can’t place them within 8 feet - and even getting that range requires moving one behind the other.

I can deal with wood, plexi, sheet metal, fiberglass with or without foam core (styrofoam or urethane).

Yes, I most likely WILL ask the manufacturer.
Google “acoustic horn” leaves much to be desired, even if you add “-speaker”

Thanks!

2

You are on the right track with a horn. Most horn design issues revolve around a balance of directivity (which you want), efficiency, bandwidth, and distortion. For the most part you are not concerned with any of the latter, so things can be pretty simple. However you will want to ensure that the coupling of the horn to the transducer is good, and this can require some exacting effort.

But before embarking there are a few questions and issues. Why does one trigger the other? I assume one transducer is listening for the dog, and the other blasts it. It sounds as if the real design flaw is in the listening stage not having enough selectivity to avoid being triggered by the other. The devices are probably just never intended to be used together.

Anyway, a simple conical horn will do the job. Make it as you would expect - angle of horn cone determines the spread. Mate the horn to the transducer carefully - you want the new horn to exactly surround the transducer in much the same manner as the existing horn does. The longer the horn, generally the better it will perform. You could go fancy any worry about exponential or constant rate of curvature profiles, but you really don’t need to for this application. A slight flare/rounding at the end of the horn will slightly ameliorate some of the inevitable lobing effects, but you won’t really care. You are really looking to build a tiny megaphone.

3

Thanks!

This sounds like I would need to remove the existing horns (think bass speaker) and attach the new horn directly to whatever the tiny thing at what would be a coil in a speaker.

This is a bit outside my comfort level of expertise.
Could the existing cone and housing be left as-is and a strictly exterior devise be used?
I’m thinking of adding “stuff” ™ to the existing cones to re-shape them, as opposed to totally replacing them (the transducers appear permanently mounted in solid cylinders.

4

Or slitting the current cones and reshape them to a tighter cone - reshape the 140 degree cone to a 90 degree cone?

This would leave the existing attachment of the material, just changing the shape.

I am concerned that the cones are carefully glued to the perimeter of the housing. I fear the performance depends on the seal.

And, since I can’t hear the output (or lack thereof), any mods made could kill the performance.

5

That could work. The coupling (the “throat”) of a horn is the most critical, after that is is much more tolerant. The actual transducer is almost certainly some sort of piezoelectric device. It should be a standalone capsule that can be separated from the existing horn. If your new horn is a cone with a truncated apex that neatly fits around the transducer capsule you should be good.

The longer your cone the better the directivity. At the mouth your horn is acting a great deal like a single slit for diffraction effects. The more wavelengths wide the mouth is the less beam widening there will be. Assuming the thing runs at about 40kHz you won’t have much difficulty.

6

I’d investigate shielding the microphones to reduce the likelihood of triggering by the peer device. It might be somewhat easier and not as permanent.

I’d start with foam or polystyrene baffles.

7

Good point. Note that a horn works both ways, and the same directivity for output works for input. Given the high frequencies involved (and thus the really rather small wavelengths) simple baffles may not work as well as one might hope. Energy can diffract all over the place in unexpected ways. Foam is the good answer, and the best foam is the foam with the highest areal density. Foam adsorbs sound by the dissipation of energy as the air follows over the surface. The more surface, the better the adsorption. A really nice, easy to find foam is reticulated melamine. BASF market it as Basotec, but you can buy what is essentially the same thing in small lumps as magic eraser. A brick sold as a kitchen cleaning aid is probably a great start.

8

All good stuff.

You might calculate the wavelength and consider your designs on the basis of wavelength. You want a cone shape, neatly coupled to the transducer, smooth, and of narrow angle. But shape features small compared to the wavelength don’t matter.

You could instead use a parabolic dish. One intended for a solar collector would also work fine for sound.

You could look at images of directional microphones. Similar designs would work for what you want to do. One difference, though, would be that directional microphones that collect sound from a large area can harvest a bigger signal, while a directional speaker doesn’t get that particular advantage from a large area.