Chronology Current Month Current Thread Current Date
[Year List] [Month List (current year)] [Date Index] [Thread Index] [Thread Prev] [Thread Next] [Date Prev] [Date Next]

Re: [Phys-L] Simply more data for you-all Was: Re: fidget spinner: more data, more analysis, timestamps, force-law plot



Here is what Bernard's data looks like, for his second data file today:

https://i.imgur.com/3bfH2EH.jpg

..and in close up, the first twelve samples of that file:

https://i.imgur.com/wfW3OTK.jpg

Here are my observations, and deductions:
1) John's note notwithstanding: this data seems of excellent quality - with no noise worth mentioning. However, what you CAN see is a discontinuity around the  240th  time. (In bc's first data set, I noted two discontinuities.)

2) Looking more closely at the first 12 values in the second data set, you can see the saw tooth effect of marking all three lobes (today) which was also visible in the two lobes of the previous spinner.

3) Both spinners that bc has used to capture data from, are stiff - so stiff that it is very likely the grease seals (which are intended to make frictional contact with BOTH inner and outer races BTW)  are dominating the results. People claim 4, 6, 8 minutes and more!


4) The discontinuities seem to arise from using Bryan Mumford's excellent clock/pendulum data acquisition system well outside its design envelope, so that it skips several lobes at its highest spin rate, stepping down to every lobe as the spinner slows.

Suggestions:
1) use only ONE flag per revolution - this both brings the Mumford timer closer to its design envelope,  and avoids the saw-tooth times which are interpretable as 'noise'

2)  If you wish us to see the aero drag and the viscous drag while greatly reducing the friction which drowns them out,  then David's suggestion notwithstanding, remove the grease shields and remove all lubrication. Even hot water and a detergent like 'Dove' will achieve this result, if followed by warm drying or spraying with rubbing alcohol which is a desiccant.

Finally - I want to say what a pleasant thing it is to see some experimentation reported on this list. If it is on an engineering topic, there is interesting physics involvement.

Brian W

On 9/21/2017 1:14 PM, bernard cleyet wrote:
On 2017/Sep/20, at 20:47, bernard cleyet <bernardcleyet@redshift.com> wrote:


I flipped it rather weakly, as I feared it would be too fast (time/revolution too short) for the MicroSet to collect W/O error. It’s the only data I took. I intend to take more tonite.

All data I’ve taken of pendula have been very noisy. I don’t think it’s artifacts. Remember the resolution is ~ a microsec. My analyzer (Kaleidagraph) has smoothing functions.

A much faster spin, I think, will result ina quad. term. And the MicroSet will average beat up to 256 beats.


bc later

New data; of the massiest three legged spinner; I spun so fast MicroSet doesn’t respond until slower.

Two modes: time one where leading edge to leading times recorded (so will be three times the “actual” rate, I think.), and time 1 skip. This one is designed to measure the period instead of the beat of a pendulum, I pray, the result is 3/2 times the “actual”.

Note the last column is the PG interrupted time. This Bryan included so one may find the amplitude of a free pendulum using the work energy theorem.


http://cleyet.org/spinner/bc's/spinner%20third%20from%20left/


bc easily confused.
_______________________________________________
Forum for Physics Educators
Phys-l@mail.phys-l.org
http://www.phys-l.org/mailman/listinfo/phys-l