Measurement of the resonance curve (Acoustics)

[solid]

Hi,
here is somewhat modified version with indication of the resonance frequency...

   

[fbouquet]

Nice curve!

I see in the code that you played a bit with the analysis sleep time, presumably to increase the speed of the experiment. There is a snag if you want to pursue this: the code tries to detect the first instance of the analysis in order to know that data should not be saved (since the measurement is made for the frequency that was calculated in the previous analysis run). Unfortunately detecting the first run is not that easy because Android and iOs handle it differently (there is a thread on that topic somewhere in the forum). In the end, I had to use a combination of factors, one of which being the elapsed time. Changing the sleep time will certainly have side effect on the detection of the first run...

One way would be to have two different programs, one for iOs and one for Android, and there you could play with the sleep time more easily.
the other way would be to find a better first run determination, one that does not use time, but still works on iOs and Android. I would be happy to learn about it!

Cheers,

Fred

[Sebastian Staacks]

One way would be to have two different programs, one for iOs and one for Android, and there you could play with the sleep time more easily.
the other way would be to find a better first run determination, one that does not use time, but still works on iOs and Android. I would be happy to learn about it!

Yes, I intent to change the iOS implementation to follow the Android version more closely and to allow setting attributes to determine when analysis modules should be executed. The idea is to allow you to specifically define that an analysis module is only executed before start is being pressed, in the first run after start has been pressed or in any subsequent run. I hope to get it into the next update along with the new tone generator and a rather long list of minor optimizations.

[solid]

Also, you can now try a different technique that might be interesting to those of you who are working with resonances. The beta version allows you to generate white noise in the audio spectrum experiment (Settings, below the number of samples), so you can feed the white noise into your resonator and can see the amplified resonant frequencies in the audio spectrum.

So I have a version for such a technique. A single spectrum of the response to the white noise collected during one second (in order to have one Hz resolution) is too noisy, but an averaging over 100 runs gives the result very similar to the result obtained by the frequency sweep. I have limited the spectrum by 2500 Hz. Some problem appeared for the so large frequency sweep, the tone generator is switching off after some hundreds points... (To be discussed in the feedback forum).

[solid]

In the phyphox version 1.1.8-beta2 the tone generator is repared !
In the upper part of the figure the acoustic spectrum of a PVC tube is obtained by FFT transformation of the response to the white noise (time of measurements 100 s). In the lower panel of the figure the same spectrum is obtained by the frequency sweep (time of measurements 2500 s).

[paul.davies@york.ac.uk]

In the phyphox version 1.1.8-beta2 the tone generator is repared !
In the upper part of the figure the acoustic spectrum of a PVC tube is obtained by FFT transformation of the response to the white noise (time of measurements 100 s). In the lower panel of the figure the same spectrum is obtained by the frequency sweep (time of measurements 2500 s).

This is superb, I have had some trouble with the white noise generator. It seems to have some artifacts in the spectrum and responds differently with different phones. But a little bit of playing and I found that 'crinkling' a sheet of paper in place of the phone generated noise produces an excellent spectrum. I've attached a picture below. The tube was the cardboard inner from some Christmas wrapping paper (about 0.7m long and .03m diameter). As you can see below the spectrum is lovely and after only a few runs. 

I plan to have my students try this at home and we will measure the speed of sound. 

Thanks again, Paul

[Jens Noriʇzsɔɥ]

I have had some trouble with the white noise generator. It seems to have some artifacts in the spectrum and responds differently with different phones.

Could you provide us with examples so that we could have a look into it. We would suspect audio (speaker) characteristics of smartphones rather than a software bug…

[solid]

I have had some trouble with the white noise generator. It seems to have some artifacts in the spectrum and responds differently with different phones.

Could you provide us with examples so that we could have a look into it. We would suspect audio (speaker) characteristics of smartphones rather than a software bug…

The Fourier transformation technique with the white noise excitation gives the same spectrum as the frequency sweep technique, so I think everything is OK with the white noise generator. The spectrum depends on the positions of the source and the detector. The best results are obtained when the source (earphone) and the microphone are inside the tube at a distance of approximately 3 cm from its end..

---

We have started to prepare the acoustic resonance experiments for large first year student labs and found some faults in the programs posted here at #36: the frequency scan experiment gave interrupted signal for iOs and the white noise Fourier experiment gave too high peaks at 0, 50 and 100 Hz for some phones. In order to correct the first fault the tone generator was configured with loop="true". For the second the Fourier experiment was limited by 200 Hz at low frequencies.

All is OK for Android, iOs is still to check...

[Jens Noriʇzsɔɥ]

Results on iOS for a stack of toilet rolls, height roughly 40cm: Resonance via FFT…

[Jens Noriʇzsɔɥ]

… and Acoustic Resonance.