| Author |
 Topic  |
|
|
StevenL
Canada
75 Posts |
 Posted - 06/07/2023 : 09:55:21
|
Hello,
I recently watched this video:
https://www.youtube.com/watch?v=SxP30euw3-0
I would be interested in seeding my Arduino Uno pseudo random number generator with a reading from my GMC 500+.
Has anyone in these forums ever done such a thing? Can anyone offer any advice on how it could be done with a automated setup?
Thanks.
Steven Lightfoot |
Steven Lightfoot, Eng.
Ottawa, Ontario, Canada |
|
| Reply #1
ullix
Germany
1171 Posts |
Posted - 06/09/2023 : 22:26:48
|
And it is only about seeding at starting, not a frequent call?
Even then a low count rate may make this cumbersome. It might be easier to measure the time between 2 clicks. At background of CPS=0.3 this should be some 3000 ms average, with an exponential distribution.
I haven't tried this, but I think you can feed the audio-out from the counter to an interrupt-in of the Arduino. But before connecting make sure to have the right voltage levels! |
 |
|
| Reply #2
StevenL
Canada
75 Posts |
Posted - 06/11/2023 : 03:47:48
|
Ullix, excellent as always, thanks, I will ponder on this. I have so many little self-learning projects on the go at once, it takes me time to get to things.
One thing I noted in the video link I attached (if you saw it) is that the presenter seemed to make a fundamental error in claiming the random nature of his radiation source could be characterized by a normal distribution, as opposed to a Poisson (although at higher count levels they appear similar visually).
Steven Lightfoot |
Steven Lightfoot, Eng.
Ottawa, Ontario, Canada |
|
|
| Reply #3
ullix
Germany
1171 Posts |
Posted - 06/11/2023 : 23:53:18
|
I looked through part of the video, but missed this distribution section.
The radioactive decay being determined by a Poisson distribution is so essential for the understanding of that decay that claiming it is a Normal distribution is really a give-away for lack of understanding! So I wouldn't really trust anything this guy was telling.
It can be mathematically shown that a Poisson distribution becomes a Normal distribution for a large mean, EXCEPT that a Poisson is still valid ONLY for integer values, while a Normal is valid for any value.
The limit for visual recognition of Poisson vs Normal is about a mean of 10, like CPM=10 or CPS=10. And of course ONLY when you set the StdDev of the Normal dist to be SQRT(mean)!!!
This may initially be hard to grasp. So the latest GeigerLog 1.5.0pre41 offers some options to play with the two types of distribution, download here:
https://sourceforge.net/p/geigerlog/discussion/general/thread/beafedc7b1/?page=4#7010
|
|
|
| Reply #4
StevenL
Canada
75 Posts |
Posted - 06/12/2023 : 06:55:28
|
| This is SOOOO good. Many thanks. |
Steven Lightfoot, Eng.
Ottawa, Ontario, Canada |
|
|
| |
Topic |
|
|
|
Seeding a pseudo random number gen with QG data?
