0

I am trying to design an electronic circuit which does the job of a dog whistle. So basically when a human whistles using the mic of the circuit, the circuit boosts the frequency to at least 50kHz or so. I am thinking to use op-amps in the circuit to boost the frequency but I am not sure if it will boost the frequency. Can anyone help me with this please?

Hamza Hussain
  • 9
  • 4
  • 4
    Search for ultrasonic and "Steve Gibson" or the "Quiet Canine" and modify for your purpose. or make/buy an ultrasonic wind whistle – Tony Stewart EE75 Nov 27 '16 at 14:16
  • 8
    Let me explain something about circuit design: before you can be a designer you have to look at and understand the work of others. The fact that you think you can use op-amps in the circuit to boost the frequency shows that you haven't. Have you seen a circuit which does that ? So why do you think you can do that ? What you need can be done with a very basic oscillator. Others have done what you're trying to make. Look at what others did. – Bimpelrekkie Nov 27 '16 at 14:37
  • @FakeMoustache But one can make an oscillator out of op-amp by connecting feedback resistor to the +ve input of the op-amp. Will that not work for the purpose of boosting frequency? – Hamza Hussain Nov 27 '16 at 17:07
  • 1
    Folks, perhaps Hamza's intention is a keyed oscillator, triggered by (turned on by) an audio tone into a microphone. A logic-gate type RC-type keyed ultrasonic oscillator might be turned on by an op-amp circuit that detects the audio tone from the microphone. – glen_geek Nov 27 '16 at 17:33
  • But one can make an oscillator out of op-amp by connecting feedback resistor to the +ve input of the op-amp You should study some opamp oscillator circuits and learn that it is not as simple as that. Will that not work for the purpose of boosting frequency? If you think it works, show us a circuit doing just that. Your "frequency boosting" is a vague term, what do you want, multiply the frequency, mix it to a different frequency ? There are many ways, if you want to design something you need to understand this also on a mathematical level. – Bimpelrekkie Nov 27 '16 at 21:24
  • @FakeMoustache All I want is to multiply the input frequency. Ideally human voice frequency being multiplied 1000s of times to make it ultrasonic. I am new to electronics so I don't know a lot of stuff but I heard my school teacher saying that if you connect the feedback resistor to the non inverting terminal of the op amp then it will do the job of an oscillator. – Hamza Hussain Nov 28 '16 at 00:59
  • 1
    @FakeMoustache Tolerance, Mr Moustache. Tolerance. Slapping new users doesn't create a good impression of SE. – Paul Uszak Nov 28 '16 at 13:08
  • @PaulUszak I was not "slapping", I'm sorry if the bare truth offended you or anyone else. I'm Dutch and many Dutch are direct, we say it like it is. If you like the "soft" approach, feel free to comment in that way or add an answer. – Bimpelrekkie Nov 28 '16 at 14:07

4 Answers4

3

An op-amp will not of itself boost the frequency - it amplifies voltage, not frequency. Two possible approaches are either to use a gated oscillator or to use a phase-locked loop to change the frequency. As dog whistles are effectively a tuned circuit running at an ultrasonic frequency, the gated oscillator is probably a good bet. Try a phase-shift CMOS circuit such as a CD4011 with delays between the gates and feedback from the third to the first gate (you need an odd number of gates to make it oscillate); the fourth gate can be a buffer to drive the output transducer. You do not need a mic for this; a pushbutton will do.

Ross Henderson
  • 31
  • 1
1

No linear element can add frequencies to a signal that aren't in the original. A opamp when used properly is usually a linear element. In any case, the purpose of a opamp is to provide voltage gain, not to introduce frequencies.

Then there is the question of what exactly the very vague "boost frequency" means. Does this mean, for example, to shift all input frequencies up by maybe 20 kHz or so, multiply them by some fixed value, or something else? Should the entire input waveform be "boosted" in frequency, or the primary frequency identified, then a higher single frequency produced that is a function of the primary input frequency?

There are too many unanswered question to be able to give a reasonable answer about how to proceed. Check out "product modulation" or "heterodyning". These can be used to shift whole ranges of frequencies. A microcontroller looking at zero crossings, after some filtering, can identify the primary input frequency and produce some other frequency as a result.

Wouter van Ooijen
  • 48,572
  • 1
  • 63
  • 136
Olin Lathrop
  • 313,258
  • 36
  • 434
  • 925
1
  1. A dog whistle doesn't "boost" frequencies. It produces a sound whose frequency depends on the whistle's physical dimensions, like a wind musical instrument. You're not whistling into one end of it and having your whistle frequency boosted. Your role in generating the sound is analogous to that of the power supply for an electronic oscillator.
  2. You could make a 23 to 54 kHz (Wikipedia on dog whistles) oscillator embodying a 741, and use it to drive a speaker. Not the best choice: not very loud, among other things. Your reference to using the + input suggests that you know of one of the popular configurations for relaxation oscillators. That configuration wouldn't be a bad starting point for your design.
  3. The earlier two answers are not just correct. They have worthwhile points for you to consider and research.
0

There have been interesting frequency doublers that use fullwave rectification of a clipped signal to get the loudest signal component doubled in frequency.

With some work you may be able to do this. Use an ALC, suitable band pass filter (say 8 to 14 kHz) an amp and clipper, fullwave, level shift, amp and HiFi tweeter.

The other alternative in the modern world is to use a DSP or fast micro-controller and do continuous Fast Fourier Transform on an input signal and adjust gain to keep the calculations sane.

EDIT:
I forgot to add in the DSP option you would have to shift the spectrum in the frequency domain and then do an inverse FFT. The signal quality does suffer but that is normal for digital time<>frequency convolutions.

Also another option was to see if any of the old pitch shifting ICs or gadgets were still to be found. They were popular as voice changes and I seem to remember some could do simple pitch changes. I donot recall what method they used, perhaps a simple DSB or AM modulation and then back again at an ofset frequency. The signals (voice) would be garbled pretty badly as the upper and lower sidebands would overlap badly depending on there the frequency fold occured.

A possible analogue domain solution is also to modulate the signal into a SSB at some handy intermediate frequency with high quality sideband reduction filter and then modulate it back down again to an offset frequency and then high pass filtering to clean it up, the quality would be better than the DSB/AM option but still hard on the ears, perhaps the dog will not mind. This would be more involved but is sure to have been done before. This (and the previous) technique woud allow for frequency shifting as opposed to doubling.

I just remembered that the fix for the very poor performance of the fullwave rectified solution was to use an analogue squaring/self-multiply circuit as this would preserve the amplitude somewhat. This is sensitive to amplitude and signal noise but a bandpass filter and AGC to preceed might get something out. This would be a frequency doubler like the fullwave rectified system.

Community
  • 1
KalleMP
  • 4,079
  • 1
  • 14
  • 32