Using Phyphox as a "crash test dummy"

[Don H]

Hi. I am a volunteer for a program that has scientists/engineers visit schools to inspire them into STEM, largely by doing experiments in the classroom with the students. My post here is a contribution to an experiment idea in this context, but I also have some questions.

The experiment, or design challenge, targeted to 10-12 year old students, is to have the students design effective crash barrier structures out of newsprint, which will assessed by running a (weighted) toy truck down a ramp that crashes into the barrier that they made which is placed it in front of a hard stop.  A smartphone fixed to the toy and running the Phyphox acceleration experiment will be used to record the maximum acceleration of the impact, and thus rate the performance of the barrier.  I think the timed run with the beeping countdown will add a nice touch to the experiment. This would be a fancy version of the classic egg drop experiment, but rather than a potentially messy pass/fail measurement, would actually provide a numeric ranking suitable for this purpose.  

I am still in the process of getting familiar with how the phone (a Samsung Galaxy A5) reacts under different impact events.  So far I notice that the maximum recordable acceleration seems to be about 90 m/s2, and it doesn't take much of an "impact" of the toy on the hard stop to reach that.  Is this range a physical limitation of the sensor?   Also, there seems some non-repeatability in the measurements at the higher acceleration values (for two similar hard impacts, one may record 90 and one 40 m/s2).  Perhaps an effect of the sampling rate or missed sampling points? For more moderate impacts, the measurements seem repeatable, but would this type of experiment benefit from more, or perhaps less, data smoothing?  Hopefully most students will make barriers soft enough to be in the working range of the phone.   In any case, I think it will be fun for the students.

[solid]

Hi,
It is always a fun to crash something. Unfortunately accelerometer in your phone is too slow (its rate of measurements is about 100 Hz). Consult https://phyphox.org/sensordb/ to find a faster phone. Anyway there is a maximum recordable acceleration which is not shown in the table but you can verify in the information about your phone given by phyphox app. So you are limited by soft barriers, not so fun..

[Jens Noriʇzsɔɥ]

Thank you for answering the post, solid.

AFAICS, one variant of the Samsung Galaxy A5 is SM-A520F with an accelerometer rate of 200 Hz, however, there are quite some Android devices with about 500 HZ. If you select an entry in the table at https://phyphox.org/sensordb/, you could get more detailed data. For the SM-A520F you find, for instance,

Accelerometer: Available
Name K6DS3TR Accelerometer
Vendor STM
Range 78 m/s²
Resolution 0.0024 m/s²
Rate 195.4 Hz

Technically, there are different modes for the accelerometers with varying ranges, resolutions, and rates. There is no way to change these by the smartphone OS API though. Typical ranges are 8 or 16g…

[Don H]

Thanks for the replies, this is helpful.
I think regardless of the limitations of the sensor to measure the higher g's, this will work well, again, in the context of a grade 5 experiment. I recently found a YT video of what I would like to do, where it looks like they used essentially the same type of sensor with some success; https://www.youtube.com/watch?v=Gho6Z8DXm1A

This had me wondering if the microphone output (a piezo device I imagine) might serve as a proxy for measuring (in a relative sense) the really high accelerations.

[Jens Noriʇzsɔɥ]

The rate of the external sensor in the video is substantially higher with 1000 Hz (IIRC). Perhaps half of that could be sufficient, so, an alternative Android smartphone… please report back if you succeed.

Ad mic: good question… I would suspect some software impact that shadows any effect, worth a try though.

[Don H]

So I just completed some learning sessions with the crash testing activity as described above. I ended up using an old iPhone 7 to record acceleration. I customised a phyphox test to present a simple graph of just the y axis acceleration as that seem to present the least noisy signal, and more importantly, to display the max/min values. While the abs. maximum measurable acceleration is only ~78 m/s2, most students were able to iterate the design of their crash barriers to achieve acceleration values below this, and thus were able to get a sense of the effect of design changes on barrier performance and compare with their classmates. Hence, I found the app, which seemed to provide reasonably meaningful numeric values despite the limitations of the phone, worked wonderfully for this purpose. In fact, some students well surpassed the performance of my designs with values as low as ~20 m/s2. I look forward to finding or perhaps making other experiments using this app for this age level of students.

[Erik Josefsson]

So I just completed some learning sessions with the crash testing activity as described above.  I ended up using an old iPhone 7 to record acceleration.  I customised a phyphox test to present a simple graph of just the y axis acceleration as that seem to present the least noisy signal, and more importantly, to display the max/min values.  While the abs. maximum measurable acceleration is only ~78 m/s2, most students were able to iterate the design of their crash barriers to achieve acceleration values below this, and thus were able to get a sense of the effect of design changes on barrier performance and compare with their classmates.  Hence, I found the app, which seemed to provide reasonably meaningful numeric values despite the limitations of the phone, worked wonderfully for this purpose.  In fact, some students well surpassed the performance of my designs with values as low as ~20 m/s2.  I look forward to finding or perhaps making other experiments using this app for this age level of students.

Hi Don, thanks for sharing!
Do you have any ideas how to easily measure speed before impact in the same experiment?
Does the car make a detectable sound both when it stops accelerating? I'm thinking of the Acoustic Stopwatch.
I have tried magnets and lights to measure speed but with limited success in the classroom.
Best regards.
//Erik

[solid]

I have tried magnets and lights to measure speed but with limited success in the classroom.
Best regards.
//Erik

You could try Doppler effect...