1

I've recently purchased 40 kHz transceivers and I tested its frequency response in anechoic chamber using a B&K microphone located 90cm away from the emitter. The response provided in the datasheet is

enter image description here

whereas my finding is

enter image description here

enter image description here

What can be causing this discrepancy? The center frequency is clearly not 40 kHz as it was claimed to be.

Is a standing wave causing the notch that is at ~38 kHz?

I am pretty sure the experiment setup is functioning well.

csg
  • 161
  • 10
  • The scale on the first graph is suspect. Why is the 2 kHz from 38 to 40 kHz narrower than 2 kHz from 43 to 45 kHz? – The Photon Oct 31 '18 at 05:43
  • 1
    It would also be easier to compare your result if you plotted it on the same scales: 35 to 48 kHz on the x-axis, and dB scale on the y-axis. – The Photon Oct 31 '18 at 05:47
  • 1
    It appears that the first graph is in dB. The second graph is labelled "amplitude." If you convert the second graph to dB, maybe it will match the first one? – user57037 Oct 31 '18 at 05:47
  • 1
    what impedance were you driving the transducer with. What impedance does tyeh data sheet say to drive it with? – Neil_UK Oct 31 '18 at 05:51
  • Normally, in order to have a standing wave, you would have to be in a room that can produce echoes. Also, the wavelength at 38 kHz is less than 1cm. Are you calibrating out the frequency response of the microphone? Could it have a null at 38 kHz? – user57037 Oct 31 '18 at 05:53
  • Eyeballing it from the graph, it seems like the 3dB bandwidth may be about the same in both graphs. But that doesn't explain the notch. – user57037 Oct 31 '18 at 06:01

1 Answers1

3

The linear Vs Db scale thing accounts for much of this, and these are not exactly precision devices so the centre frequency being slightly off is not a huge surprise.

The notch is only about 3dB, and may well be two transducer modes being excited, you may also be seeing a null due to a reflection from the housing or something, try rotating the transducer very slightly off axis and see if the notch moves.

If reflections causing standing waves are a concern, you can try setting up a time gate on your measurement (Emit a pulse, gate the measurement for when you expect the signal to arrive).

Dan Mills
  • 17,266
  • 1
  • 20
  • 38