Sonic lost key finder -- how does it work?

Discussion in 'General Electronics Chat' started by spulka, Aug 2, 2010.

  1. spulka

    Thread Starter New Member

    Aug 2, 2010
    8
    0
    Hello,

    I've seen lost key finders that respond to sound by beeping. Does anyone know how they work? Are they responsive to a specific frequency? Is it hard to build your own (I don't have much experience with analog circuits)? Could it be made to respond only to non-audible frequencies?

    Thanks
     
  2. kingdano

    Member

    Apr 14, 2010
    377
    19
    it wouldnt be a terribly complicated circuit to design - having no experience may not help though.

    what you need to do is pick a non-audible frequency (dog whistle might work, if you dont have any dogs nearby) and design a receiver circuit with a microphone.

    the hard part would be implementing a narrow band-pass filter to trigger the beeping cycle when the signal is detected.

    the beeping could be controlled with a 555-timer and a piezo-speaker i would assume, but i dont know for sure.
     
  3. tom66

    Senior Member

    May 9, 2009
    2,613
    214
    You'd also want a low sleep current (for when no whistling is detected), because otherwise the battery will go dead in the keychain quickly. I would suggest a microcontroller with a sleep mode.
     
  4. kingdano

    Member

    Apr 14, 2010
    377
    19
    good idea!


    also, if you have access to MATLAB and their FDA (filter design) tool you can implement a filter relatively quickly inside a uC - i am fairly sure they generate C code for you.

    be forewarned though, as these types of applications (filtering the whistle noise) are best suited to a DSP (digital signal processing) device, not a regular uC.
     
  5. spulka

    Thread Starter New Member

    Aug 2, 2010
    8
    0
    Thanks for your replies. I am fairly familiar with AVR ATmega8, that could be used for the ringtone part I think. What I am not sure about is waking up the μC at the right time. Any suggestions?
     
  6. tom66

    Senior Member

    May 9, 2009
    2,613
    214
    The ATmega8 consumes 500nA in sleep mode, which should be low enough. You'd probably use interrupts to do what you want. Unfortunately, I'm not familiar with AVRs, so I can't help you much.
     
  7. spulka

    Thread Starter New Member

    Aug 2, 2010
    8
    0
    But how much could the rest of the circuit consume?
    Also, I found only an electret microphone for 40Hz–20kHz. Do you think that is enough?
     
  8. retched

    AAC Fanatic!

    Dec 5, 2009
    5,201
    312
    40Hz to 20kHz is the human audible spectrum. You will want a ultrasonic transducer or the like to listen for non-audible tones.
     
  9. tom66

    Senior Member

    May 9, 2009
    2,613
    214
    If designed properly, the rest of the circuit should consume virtually nothing. I'm thinking of some kind of piezo transducer, with a band pass filter and a few protection components to prevent high voltage spikes from damaging the micro.
     
  10. spulka

    Thread Starter New Member

    Aug 2, 2010
    8
    0
    Thanks. As for the sound frequency, I think it doesn't have to be necessarily inaudible (it's going to be followed by the sound of the receiver). But I want to tune the receiver to a frequency that is very uncommon so that your keys don't start ringing every time you push that creaky door handle.
    I wonder whether a cell phone speaker can emit such a high-frequency sound that would trigger the ringtone (as one is less likely to lose his cell phone than a dog whistle or another sound-emitting gadget :)).
     
  11. BMorse

    Senior Member

    Sep 26, 2009
    2,675
    234
    Here is a similar circuit, this circuit was used to start robots in a sumo competition by a whistle blow, it is set to detect a certain pitch sound usually from a piezo buzzer...

    [​IMG]

    B. Morse
     
  12. spulka

    Thread Starter New Member

    Aug 2, 2010
    8
    0
    Is it better to use a piezo transducer or a electret microphone for this purpose? Could you please recommend me any reading to help me better understand the matter?
     
  13. tom66

    Senior Member

    May 9, 2009
    2,613
    214
    Any such circuit designed for a lost keychain will need a long battery life, so that excludes any kind of active amplifier. BMorse, your circuit will draw a few mA from the batteries, so it will stop working in days to weeks if run off a coin cell battery. You'd need a device which when moved produces a voltage. An electret microphone does not do this, it requires a supply voltage, while a piezo transducer does.
     
    kingdano likes this.
  14. spulka

    Thread Starter New Member

    Aug 2, 2010
    8
    0
    That's what I thought too. However, I don't know what voltage level does a piezo transducer produce but I am sure that it needs to be amplified before it's used to cause an interrupt. What do you recommend then?
     
  15. tom66

    Senior Member

    May 9, 2009
    2,613
    214
    A piezo will usually produce a VERY high voltage, when pressed I've seen upwards of 300V. This is of course at very low current, so there is little danger of getting electrocuted!

    You will need a protection device for your micro, or you will blow it to smithereens. I suggest a zener. The output from a piezo is AC at the sound frequency usually. A band pass filter could be used to then produce an interrupt signal.
     
  16. BMorse

    Senior Member

    Sep 26, 2009
    2,675
    234

    So who is going to be "pressing" the piezo while he is looking for his keys?? :D

    A piezo sensor does not create usable voltage if used for detecting "sound", you will have to amplify the output to get any usable voltage out of it.... They are great for vibration or movement sensors, but when it comes to trying to pick up a certain sound frequency, the output needs amplifying.....

    B. Morse
     
  17. tom66

    Senior Member

    May 9, 2009
    2,613
    214
    Good point, I haven't tested it. But, I reckon even audio levels ought to produce a reasonably high voltage. It depends on your distance from the device and sound levels of course. One other way would be to figure out a way to power the circuit from a piezo if that is even possible so when you whistle it activates the micro, or a way of using a NPN transistor which only needs 0.7 volts to switch on.
     
  18. BMorse

    Senior Member

    Sep 26, 2009
    2,675
    234
    this is what wiki says:
    So I guess if the OP can blast the piezo with a powerful sound wave, he might get enough juice out of it to turn on a transistor, (Although it would be kind of funny seeing him walk around with a 12" woofer and an amplifier to blast the surroundings with some Bass to find his keys... :) )

    B. Morse
     
  19. tom66

    Senior Member

    May 9, 2009
    2,613
    214
    I discovered that Maxim may have the solution:

    http://www.maxim-ic.com/datasheet/index.mvp/id/2559

    Yes, that's right, 750nA! That's active current draw, not in standby, per amp. They make a 4 channel version as well, so about 3µA draw, with 4.8µA specified as a maximum.

    The question is however, can the input side with the 10k-47k divider be optimised? Because, at the moment, that will draw a massive ~52 µA from a 3V coin cell, with a typical capacity of ~100mA·h means it will be dead in under 80 days, which is not that long...
     
  20. BMorse

    Senior Member

    Sep 26, 2009
    2,675
    234
    With most "sophisticated" key finders that I have seen (and taken apart) are usually based on some low power low pin count micro's..... So maybe the op will have to delve into the world of microcontrollers if he wants to build something similar....


    B. Morse
     
Loading...