Poisson Statistics
Purpose
 Measure the background radiation by using an NaI detector, and create a poisson distribution.
Equipment
SpecTech equipment (back)
 Computer
 Windows XP Professional Version 2002 Service Pack 3
 Intel Celeron CPU 2.26 Ghz
 1.96 GB of RAM
 NaI detector
 TN1222 Detector Base
 Tracor Northern
 Model # TN1222
 SN 880067
 Bicron Mod 2N2/2
 SN CK091
 +1200 Volts, 1300 Max
 Spectech Universal Computer Spectrometer, UCS 30
 Connected to computer with USB Cable
 Manufactured by Spectrum Techniques inc. Oak Ridge Tennessee 37830
 Input channel connected to detector Input channel
 POS High Voltage channel connected to detector HV channel
 Spectrum Techniques UCS30 Software
 UCS30 Version 2.10.7, Friday January 23, 2009  2:41 PM
 SN. 505, FW Rev. 6
 20062007 Spectrum Techniques, Inc.
Safety
 High voltage, so beware of shock.
 We measured the background radiation, so there is no danger of radioactive substances.
Setup
 Turned on the power supply.
 Opened the Spectrum Techniques UCS30 Software on the computer.
 In the Mode menu, set the software to MCS Internal.
 In the settings menu, select "Amp/HV/ADC"
 We set the High Voltage to 1000V.
 We also set the channels to 256.
 Then we set the dwell time to 80ms initially.
Procedure
 Press go.
 We changed the dwell time to 80, 100, 200, 400, 800, 1000ms and ran the experiment again.
 For each of these we plotted the occurrences verse the number of counts.
 Next we plotted the Poisson distribution for each dwell time and the Gaussian distribution for the 1000ms dwell time.
Data
Poisson Statistics XL Doc
Poisson distribution function
Calculations
Poisson function:

Gaussian function:

We used these two functions in our graphs to show how our data related to a Poisson distribution, and in the last graph of 1000ms, we also plotted the Gaussian distribution (yellow).
Error
 The error in our graphs is because we chose to follow the manual and use the lowest number of channels. We used only 256. If we had used more channel we would have gotten more measurements, and therefore a better result. Despite this, all of our graphs except the 800ms and 1000ms graphs match the Poisson distribution closely.
Citations/Acknowledgements
