PSD literature

http://wlab.yale.edu/sites/default/files/psdpresentation.pdf

http://www.slcj.uw.edu.pl/en/events/nwall/presentations/soderstrom.pdf

Scintillation - http://neutron.physics.ucsb.edu/docs/scintillation_presentation_info.pdf

http://cyclotron.tamu.edu/reu/2004%20Presentations/rebecca_files/v3_document.htm

Masters dissertation- http://digitalscholarship.unlv.edu/cgi/viewcontent.cgi?article=3136&context=thesesdissertations

http://is.muni.cz/th/51646/fi_r/teze.pdf

http://uu.diva-portal.org/smash/get/diva2:342845/FULLTEXT01.pdf

Dissertation
http://uu.diva-portal.org/smash/get/diva2:342845/FULLTEXT01.pdf

Old paper with formulas
http://www.osti.gov/scitech/servlets/purl/6266648-DgpRmT/

Useful Links for "Math" FPGA Design

Adding an adder as a peripheral PL block:
http://zedboard.org/content/creating-custom-peripheral


Building a matrix multiplier in Vviado/SDK
http://zedboard.org/content/matrix-multiplication-vivado

Things to Look-Up

Figure of Merit (FoM):


Integration Gate (as it pertains to FPGA):


Short Gate (interval):


Long Gate (interval):


Charge Comparison Algorithm:

What is the interaction in the detector to detect gamma rays?  high energy neutrons? low energy neutrons?


How does the PL accept the information from the processor? How fast can a matrix multiplier operate in the PL of an FPGA housing a 1 GHz processor?




What is computer architecture?

Pulse Shape Discrimination Algorithms

Algorithms from Digital approaches to field neutron spectrometry

Charge Comparison Algorithm

To find the discrimination parameter, integrate over the long interval, which spans the sum of all the samples, and integrate over the short interval, which is the tail of the pulse. (Note: The long integral is proportional to the total light output.)

The advantage of this algorithm is that it is said to be the simplest of the three and “it would be desirable to implement the Charge Comparison algorithm as embedded code in the FPGA…”

 

Pulse Gradient Analysis (PGA) Algorithm

To find the discrimination parameter, take a sampling pulse voltage at a specific time after the peak and then find the ratio of that sampling point to the peak amplitude. (Note: The ratio is larger for neutron event than a gamma event.)

Model Pulse Algorithm

First, gamma and neutron pulses are created. Then the unknown pulses are then characterized by the pulse that is most similar by using chi-squared. The discrimination parameter is the difference between chi-squared for the gamma model and chi-squared for the neutron model.

Pulse Shape Discrimination

Pulse shape discrimination:

When studying the two waves above (Gamma and Neutron), one can notice that they look very similar. The peak height is about the same and the beginning of the wave is quite similar as well. The tails of the waves, however, differ in the fact that the gamma wave approaches zero at a faster rate than the neutron wave.

NASA takes measurements by looking at the ratio between the area under the peak and the area under the tail of the curve. This ratio is called the pulse shape.

Then they graph the pulse shape vs. pulse height shown in the top center of the phot above. This allows one to see the differences between gamma rays and a neutron.

Note:

When studying gamma rays, one must observe the movement of the electrons because gamma rays cannot be directly observed. As shown in the top right corner of the photo above, the gamma rays will come in and “bump” an electron. By studying the paths of these electrons, one can easily tell where the gamma rays are coming in from.

Similar to the gamma rays, one cannot directly observe the path of a neutron. One could observe the path of a proton that has been “bumped” by a neutron instead. As shown by the diagram on the right center of the photo above, there can be more than one detector to find the path of the neutron by observing the different paths of the protons.