Re: [Phys-L] timestamped events; was: Bayesian Inference in Half-Life measurement

[Brian Whatcott <betwys1@sbcglobal.net>]

 I spent some time looking at Paul Nord's November First dataset mentioned below.

This is what it looks like in its original state.
https://imgur.com/C7voSaa

  Paul's original data: 20 sec counts vs bucket#
You will notice the 8+ minute setup time, and near the end some unexpected spikes.   
Here is a close up of the early data buckets:   
https://imgur.com/8Ptry5R ;  
       ( first 100 buckets)   


You will notice the 8+ minute setup time, and near the end some unexpected spikes.    
Here is the tail end of the decay series:    
 https://imgur.com/Tq3wyxA ;   
( last 1400 buckets)   


I saw a sequence of spikes:    
https://imgur.com/htcxXRH ;  

 The time between spikes was (180X20) seconds = 1 hour   I speculate these may be due to power line glitches.
I selected data from the 500th second data to before the first spike at 15000X20 seconds,  three plus days later.
https://imgur.com/LILceiS  ;
 cumulative decay plot end slope 0.58/second
You can see that the final slope of the last ten minutes of the Cumulative plot showsa slope of possibly 35 counts per minute.
Finally I looked for any idiosynchracies in the edited decay curve, buta periodogram showed only the expected energy in the first few periods, i.e 40 sec, 80 sec etcetera
 https://imgur.com/oojwUPg ;  

periodogram of first ten buckets   



Re: [Phys-L] timestamped events; was: Bayesian Inference in Half-Life measurement
   
   - From: Paul Nord <Paul.Nord@valpo.edu>
   - Date: Mon, 1 Nov 2021 17:25:28 -0500
John,

A better data set can be found here:
https://github.com/paulnord/bayesian_analysis/blob/main/CopperDecay20secInterval.csv

It's the same experiment with the geiger tube. The data was collected with
a multi-channel scaler.
There are 16k time bins recorded at 20 second intervals.
The first several bins record only background.
At some point the sample is placed near the detector. It appears to have
been done about 10 seconds into a counting cycle.
The very next bin (500 seconds) should be purely data with the sample in
place.

I did move the detector closer to the sample and took some care to center
the sample better on the end of the tube. I believe that this is largely
responsible for the increased count rate.

There seem to be some interesting increases in the activity near the end of
the run. There appear to be increases in background activity that last a
few minutes with some rise and fall. This may be the sort of thing that
one would expect from solar activity. But it's possible that other
external factors would influence detector gain in some way. And they may
simply be random coincidences.

Paul