Group Actions
Engagement
  • Author Author: kmikemoo
  • Date Created: Date Created
  • Views 6739 views
  • Likes 14 likes
  • Comments 11 comments
Related
Recommended

Disappointing Ultrasonic Dog Deterrent (DUDD)

kmikemoo

I was tempted not to post this project because it didn't turn out how I wanted.  So what changed my mind?  I was reading some older posts and one of the community veterans had commented that projects don't always turn out right. Yup.  So I'll share what I learned along the way and maybe someone else will build the better mouse trap - or dog yap trap and I can sleep past 6am on a Saturday.

 

Like many a tinkerer, the 555 has a special place in my heart.  It's one of those MacGyver chips that seems to have endless uses.  Okay, not endless - but A LOT.  When I saw a 555 H-Bridge Push-Pull dog deterrent circuit, I had to try it.  The original circuit used BD679 and BD680 Darlington transistors.  I ordered some from China but... I may see them in April.  Maybe.  I substituted TIP41's and TIP42's because I could get them.

image

Schematic corrected March 12, 2020.

For a speaker, I chose some cheap piezo horns and modified one.  I removed the horn and adapted a 1" PVC coupler in its place.  The plan is to connect a 24" barrel and make the sound as directional as possible.

imageimageimage

Since 30kHz is outside of the human hearing band, one of the piezo speakers was to be used as a receiver to test the output.  A multimeter connected to the receiver piezo does respond to varying sound levels.  Since I was going for an elevated output... I was okay with the test system not being all that sensitive.

 

So... I breadboarded the circuit and this is what it sounded like...

Needless to say, this was not what I expected.  I played with the potentiometer initially - until I noticed the smell of wires getting really, really warm.  I'm not audio expert, but I'm going to say that buzz is that the speaker can't respond to whats going into it.  Playing with the potentiometer didn't seem to affect the output sound.

 

I let the circuit cool and played with it a bit more.  The output frequency is not stable at all and pretty much runs away (escalates) as the transistors get hot.  I could heat sink them, but I don't have faith that it will change much.

 

So I tried to go a different direction.  I connected a standard 555 circuit up and connected it to the piezo.  My Fluke says that it starts at 29.3kHz and climbs slowly.  The major issue?  I can hear it.

imageimage

I shouldn't be able to hear it.

Thanks to the timely arrival of my new tool, I was able to grab this screen shot.  Thank you again to element14 and the judges.

This also allowed me to check my original circuit.  I substituted the trigger circuit from the simple 555 circuit in for the original 1k / 100k circuit.

image

While I was expecting the amplitude, I wasn't expecting the frequency.  Time to heat sink the TIPs and slow the initial 555 by a factor of 10.

 

Yup.  Heat sinks didn't really change anything.  How can the transistors get that hot that quickly with only a 2 amp power supply?

I have rebuilt the circuit for 20kHz, 15kHz, 10kHz and 3kHz a couple of times each with inconsistent output results.  Why?  Well...

1.  One of the things I discovered was that my wall wart makes noise - enough on the last build that with one power lead (either one) disconnected I was still getting audible output on the piezo.  I measured about 5mV peak to peak at an unstable 210kHz-ish.  At the piezo, it was a clear 60.0Hz.  A filter cap on the power input didn't clean it up.

2.  On the original circuit, Pin 5 on the 555's floated.  I could get variations in output frequency just by getting a finger or wire near the pin.  A 100pF capacitor acted more like an antenna than anything else.  A 1nF capacitor tying the pin to ground made the 555 output consistent.

3.  Maybe it was the breadboard, but the wave forms became more distorted the more times I rebuild the circuit.  The hiss from the video was always present.  I did go back to the single 555 output just to make sure that I hadn't damaged the piezo.  I didn't beak it - but I could also still hear it at 33kHz.  That ain't right.

4.  A few times, I did get input frequency as output frequency so the concept is functional.  I never saw another smooth waveform like in the picture above.  More often, I saw rail to rail with a leading spike - and lots of harmonics.

 

I'm going to call this one BUSTED.  I do believe that the switching noise from the power supply created some of the hiss.  If ultrasonic was happening, it didn't matter because the audible noise was also present.  I guess I'll keep looking for a better dog yap trap.

  • The 15kHz makes sense.  I hooked up my audio generator last night to half of my H-Bridge - which I think is an A-B Amplifier.  I could hear the pitch rise up to around 15kHz to 17kHz according to the scope.  Oddly enough, I believe the tone then started to decrease - slightly - as I continued up to around 30kHz.  I feel hit the nail on the head with intermodulation distortion, electrical and mechanical.

    It took me a bit to catch on, but I like your microphone suggestion.  Just because I think I'm putting 30kHz into the piezo doesn't mean that's what I'm getting out of it.  I abandoned my detector because the transmitter still had audible noise.  I should pursue it still.

  • in reply to kmikemoo

    "...The sound level falls off as I approach 30kHz - which I assume is my hearing..."

     

    You will be lucky to hear anything above 15kHz with your adult ears.

    Young children can hear up to about 22kHz but few adults can.

     

    Hence the suggestion of using a microphone to see what frequencies are actually coming out of the tweeter unit for a given input signal.

  • If you want to see how you're torturing the power supply, you could place a small value resistor in series with one of the supply lines and look at the voltage across it with your 'scope. That would then show you the current it's supplying and how it varies. Pick a low value, so that the supply voltage doesn't vary too much, but high enough that you get reasonable resolution on the 'scope. For instance, something like 100mOhm would give you 100mV per amp and not impact the supply voltage too much on a 12V circuit. Try to go for a non-wirewound one if you can. Don't run it for too long if you see a high current and the resistor wattage isn't quite up to the job.

     

    Be careful if you want to measure voltages on a second channel at the same time. One way to arrange that would be to put the resistor in the ground return to the supply like this so that the 'scope grounds were common (that will lead to the current trace being upside-down but, if you're not comfortable with that, 'scopes usually have the means to invert a channel):

     

    image

  • in reply to kmikemoo

    A Darlington Pair is a classic BJT arrangement that results in a gain multiplication across the stages. Because of this, a Darlington would require less base current to achieve the same collector current. Using ordinary transistors instead of a Darlington might not work well if the circuit is expecting the high gain, as the single BJT would not be biased as "hard on" by that current. But I don't know enough to know whether this applies to your particular circuit.

     

    Indeed, mains noise leaking through could be one thing, but there could also be a subharmonic resonance mechanically in the unit, or some other modulating effect at play (e.g. microphonics, magnetostriction). It's always fascinating to hear something when you didn't expect to ... it even happens in optics (e.g. UV-C LEDs outside the visible range do appear a very dull pale blue due to optical noise emission and sometimes other colours when reflected off objects due to fluorescence - don't try this at home!).

     

    - Gough

  • in reply to Gough Lui

    The original circuit actually had Darlingtons.  Would that make a difference?  I also suspected what I would call harmonics.  My electrical power background is stronger than my radio background. image  The intermodulation of the noise from the power supply - then amplified - certainly could yield something close to what I saw.

    I'll try the cap.  It's a better direction than the "Uh... I dunno." I had.