Bounce Characterization (Count, Time)

[Charles Haase]

I'd like to create an experiment which will look for bouncing phenomenon in the Z direction between 4 and 6 Hertz.  When such a bounce occurs, I'd like to log the intensity (peak acceleration) and the time at which it occurred.  Optionally, I'd like to also graph these over time.

The Editor is a bit of a mystery to me.  Any help would be much appreciated!

-CH

[Sebastian Staacks]

This might be possible, but this highly depends on what the recorded signal looks like and if we can come up with a way to detect these "bounces" in a reliable way with the means provided by phyphox. The magnetic ruler expriment does something similar by detecting peaks in the magnetometer data and plotting their times. Do you have example data? Is it just a short acceleration peak that occurs at a rate of about 4-6 Hz or did I misunderstand you there? If so, does the "acceleration stopwatch" work for you (check the "many" tab)? If so, is it only the count you are missing?

[Charles Haase]

This might be possible, but this highly depends on what the recorded signal looks like and if we can come up with a way to detect these "bounces" in a reliable way with the means provided by phyphox. The magnetic ruler expriment does something similar by detecting peaks in the magnetometer data and plotting their times. Do you have example data? Is it just a short acceleration peak that occurs at a rate of about 4-6 Hz or did I misunderstand you there? If so, does the "acceleration stopwatch" work for you (check the "many" tab)? If so, is it only the count you are missing?

So the purpose of this is to identify what is causing my desk to bounce up and down at work.  I am on the second floor of a building, and there are various pieces of equipment on the floor below me that run at different times.  I measured the bounce using phyphox Acceleration Spectrum, and the peak frequency during the bounce is either 4.72Hz or 5.12Hz.  What I'm looking for is to be able to record data for some time period (ideally a full day) and characterize during which times there is a significant spike at that frequency.  Then I can work with my facilities group, for example, to find out which machines were running during that time.

Acceleration stopwatch just tells me when the acceleration is above a certain threshold, but not the frequency.  So as far as I can tell, it would be difficult to isolate the regular bouncing from (for example) typing on my keyboard, or setting my phone on my desk.  Either of those events generates acceleration spikes larger than 0.17m/s2, which is the threshold necessary for the bouncing to trigger a result in Motion Stopwatch.

[Sebastian Staacks]

Ah, I see. So, indeed I misunderstood you at first.

Unfortunately, I do not see a simple way to implement this in phyphox. I mean, in theory, one could use the acceleration spectrum (maybe with some more data points than the default to increase the spectral resolution) and save the times at which the amplitude in this frequency range is increased, but since you can expect other acceleration "noise" as well, it will be difficult to detect this reliably. It might be easier to just record the data and analyze it later with a script.

Also a little warning: Phyphox is not yet well designed to deal with long measurements. Depending on your phone, you might run out of memory if you try to record the entire day. If you run into this problem, you can set up a version which only records the relevant axis (I would assume it's z) and maybe at a lower rate.

[Charles Haase]

Ah, I see. So, indeed I misunderstood you at first.

Unfortunately, I do not see a simple way to implement this in phyphox. I mean, in theory, one could use the acceleration spectrum (maybe with some more data points than the default to increase the spectral resolution) and save the times at which the amplitude in this frequency range is increased, but since you can expect other acceleration "noise" as well, it will be difficult to detect this reliably. It might be easier to just record the data and analyze it later with a script.

Also a little warning: Phyphox is not yet well designed to deal with long measurements. Depending on your phone, you might run out of memory if you try to record the entire day. If you run into this problem, you can set up a version which only records the relevant axis (I would assume it's z) and maybe at a lower rate.

Thanks for clarifying. I understand. Thanks for a great app, and the detailed support. Keep up the good work!