You have to be a little careful. Continuous ultrasound can lead to headaches or trigger migraines in humans. A reason for commercial ultrasound insect repellers being low power and intermittent could be due to regulations in a particular country prohibiting higher or continuous power on medical grounds.

In addition, soundwaves in air become more directional as you go up in frequency. This makes it unlikely that a single transducer could effectively cover a whole house, so they are normally limited to one room.

Many years ago I designed a small ultrasonic transmitter that you could carry with you and place on your table at outdoor eating places. It issued bursts of ultrasonic tones at various frequencies, carefully designed to repel different types of insect. I built a few prototypes, and they appeared to work, but I couldn't get a manufacturer interested in taking it on. You can nowadays buy packs of 4 plug-in repellers for £10 from Amazon (not my design!).

No pets in the house? Or the neighbor's?

Thankyou for the advice!

I don't think ultrasonic pest controls could even reach the neighbours even if I tried to be honest.

I've recorded bats, mice and several insect species that produce sounds over 100 kHz. There's no way to get a feel for the level of ultrasonic sounds. The transducers can be extremely efficient. Not to mention animal ears can be very efficient too. Mice, when fighting go up to 220 kHz, fi.

I've also tried the little ultrasonic mosquito and mice repellants. They work, but seldom for all species of a genus. There's always a bug that seems to like the noise. Mice also get accustomed to the noise. I wish I'd measured the spectrum of these devices.

Making a very loud "all species" repellant seems very hard. And I'm pretty certain our neighbor's dog can hear the noise. I can switch on his barking with a piezo tweeter easily, for instance.

Producing loud ultrasonic noise is easy enough. That makes for an extended range, but doesn't yield better -long-term- repelling efficiency.

In my day job we tried ultrasonic systems for bird control (mostly geese and pigeons). The sounds were pulsed based on movement sensors. When the a movement senor in the rafter was activated the 30Khz square wave signal would pulse start. You could hear it, it was annoying.... so there was some harmonics in the human range. It also only worked initially, eventually the birds learned to ignore it.

When I researched the subject of insect hearing for the repellent device I mentioned earlier, there were two salient points that came up:

(1) No single tone (either pure or with its associated harmonics) would work for all species, so it was necessary to cycle round a series of tones that had their fundamentals at different frequencies.

(2) For it to act as a repellent, the tones had to be pulsed and not continuous. The pattern of pulsing is as important as the tone frequencies. In addition, the intervals between groups of pulsed tones should have a random element.

I'm sorry that I'm not going to give you the frequencies and pulse patterns, as that part of it was my intellectual property. However, I can say that it turned out to be surprisingly complex, but I did get it to work reasonably well. The one characteristic that ran in my favour was that I specified the device was for taking with you to outdoor eating places, so it was likely that the local wildlife would not have the chance to learn to ignore it. This was the aspect that a couple of the companies that turned it down picked on - that it might not work so well in your own garden after a day or so of use.

I've never found a reasonably certain answer to the stupid question: "How high do you need to go?"

My gear is limited to 192 kHz SR, so practially speaking, the highest frequency I can record, is

I can tell Boswell spent more time on the subject than I did. He understands the pitfalls.

Same story with colored light. Blue attracts, yellow repels, IIRC, for mosquitos. Heterocera otoh, like blue light. A never ending story...