This project analyzes SiPM (Silicon Photomultiplier) photon count data from CoMPASS, a physics data acquisition system. We're looking for:

• Peaks in photon count curves, at different gain and pulse voltages.

• How those peaks shift, how far apart they are, and whether you can correlate signals across channels.

There are two main analysis types:

1. Single-Channel Analysis (focuses on one detector)

2. Coincidence Analysis (compares two detectors, looking at correlated events)

#️ What to Run — Step-by-Step

1. Put the data files in photon_counts_data. Clear the folder generated_peak_data_results that gets populated when running the main script for the other analysis scripts (slope, comparison, etc) to have just the most recent data.

2. Setup the parameters in the main script

plot-fit-peaks-SiPM-data.py

• data_dir: change the name of the name starting with a date in the data directory to look at. All data data-photon-counts-SiPM/, you need to specify which day's data you want to analyze

  • e.g 20250505_coic_correlation_time_vary isa foldername you can change in there to 20250428_more_light

• gain_voltages_to_plot to see only data associated with the desired gain voltages determine the number of plots and which plots are plotting for that specific gain voltage

  • e.g if my data for example has [65.7,65.8,65.9] in its list, then changing it to [65.7,65.8] plots only data associated with that tag and same for a single [65.7]
    • obviously, this breaks when you choose something that doesnt exist, probably an exception or error will hiw

• crop_off_start how many values you want to crop off the start of the data

  • e.g if you want to crop off the first 100 values because the data is shifted to the right, then change it to crop_off_start=100

• crop_off_end how many values you want to crop off the end of the data

  • e.g if you want to crop off the last 2000 values that have no data in them (produces data up to 40000 and only relevant data is ~1500 in size, then change it to crop_off_end=2000

• vertical_lines boolean True or False, puts a vertical line where the calculated peaks are, it sometimes makes it easier to read the plot. For example, if turned on then all the plots will look like this:

• 

  • I set only one value of the gain to look at currently and that is its plot

• The next set of parameters determine the fitting of the data (using interpolation of the data to get the curve but will double check)

• counts_threshold: threshold of counts that anything below 100 will not be considered a peak, easy to change around when you see lower peaks not being found by peak finder

• peak_spacing_threshold: threshold of distance between peaks,( e.g prevent multiple points on the same peak), mine is at 16, which means the peaks have to be at least 16 indices away from each other to be considered a peak

• sigma this is used to make the data more smooth (is the standard deviation of the gaussian fit, this is used to determine how wide the peak is, if you want to change it to be wider or narrower, you can change this value)

• pulse_color_map allows you to change the specific colors of each of the pulse height curves

• manual_peak_indices if no matter what you do to try and get all the peaks, some won't be found because they are on a vertical hill or something like that, so you can add in peaks by analyzing the graph and putting in that index so it gets properly counted as a peak

• clear the folder that gets populated when running the main script for the other analysis scripts (slope, comparison, etc) to have just the most recent data in generated_peak_data_results

• csv updates automatically just check the date it was generated if unsure

  1a). Description: This python file will plot all the gain voltage subplots associated with the channel. It plots the data, fits and smoothes the data, and generates the data required for further analysis in the 'single-channel-analysis' and other subdirectories.

Run the script.

Results files get generated after each run of the main script. The folder generated_peak_data_results gets populated with the data used for the other analysis scripts (slope, comparison, etc).

3. Run single-channel-analysis/plot_index_vs_peak_slope_spacing_table.py to generate the index vs peak data. This python file will plot the index vs peak data for each channel and save the plots in the folder "index_vs_peak_data_results". It will also save the data in a csv file in the same folder.

4. Run single-channel-analysis/plot_spacing/plot_spacing_between_peaks.py to generate the spacing vs peak data. This python file will plot the spacing vs peak data for each channel and save the plots in the folder "spacing_vs_peak_data_results". It will also save the data in a csv file in the same folder.

• Output:

  • Peak plots (for each channel)

  • CSV files with peak positions (in generated_peak_data_results/)

• What it does:

  • Reads in CoMPASS data

  • Smooths, crops, finds peaks

  • Plots curves and saves peak index data

• Use script plot_light_vs_dark_counts_pulse_height_vary_gain_voltage_loop_subdirectory.py

• Needs folders with names like _dark and pulse for comparison

Each folder is named like:

65_7_gain_1_1_pulse_300s

The script extracts:

• Gain: 65.7 V

• Pulse Height: 1.1 V

• Duration: 300s

For coincidence measurements:

peak4_and8_50ns_correlation_window_65_7_gain_1_6_pulse_60s_filtered

Extracts:

• Peaks: 4 (main) and 8 (comparison)

• Correlation window: 50 ns

• Whether it's filtered or raw

• Same info about gain/pulse

• Don't forget to delete old files in generated_peak_data_results/ before rerunning.

• Wrong folder path = no data found. Make sure data_dir is correct.

• Make sure each run folder inside data-photon-counts-SiPM/ is well-formed (naming conventions matter).

In single-channel-analysis:

* `plot_index_vs_peak_slope_spacing_table.py`

	* Reads peak data from `generated_peak_data_results`/

	* Plots peak index (x-axis) vs. peak number (y-axis)

	* Calculates slope of each curve (i.e., peak spacing trend)

* `plot_spacing_between_peaks.py`

	* Also uses peak data

	* Shows how the spacing between peaks varies with peak number

	* This helps see if spacing is regular, how it changes with light/gain

* Each creates:

	* Plots in its own folder (index_vs_peak_data_results/, etc.)

	* CSV files with calculated slopes or spacings

The data structure was choosen to be able to loop through subdirectories in the data folder because in CoMPass, they package a single run into a project to be able to so easily drop the runs in a specific **drop CoMPass run folders can be simply dropped into the data-photon-counts-SiPM folder. (you need to change the path or take files out of each folde) **

• check that the data is in the correct format to parace information out of them - - WARNING: COMPASS PATH DATA WAS PULLEN OUT DIRECTLY, SO YOU'D HAVE TO CHANGE THE PATH UNTIL I BUILD OUT SOMETHING TO PARCE THINGS OUT OF THOSE :

• WARNING: If there are files in /generated_peak_data_results from a previous run, delete them before running the script. This will ensure that the new data is not mixed with old data.

Gain-voltage-number-decimal-represents-decimal**_gain_**Pulse-voltage-height-decimal-is-underscore_pulse_Duration-of-the-run_s contains folders with the data

• • Example: indata-photon-counts-SiPM there is a folder named 20250428_more_light Therewill be subdirectories inside which contain the run data from CoMPass. Currently for me they are:
• • 65_7_gain_1_1_pulse_300s
• • 65_7_gain_1_0_pulse_300s
• • • This means that I took two runs that day where I was changing the pulse height. The 65_7 gets parced out to be 65.7 because it's in the back of the word gain
  • • The 1_1 in the directory name means I was using a pulse height voltage of 1.1 because it is in front of the word pulse
  • • The 300s means I was running the experiment for 300 seconds.
    • Some files have a dark written on the name, this was used in the dark measurement comparison and that gets paraced out in a very specific script plot_light_vs_dark_counts_pulse_height_vary_gain_voltage_loop_subdirectory in other scripts these types of files will not be seen

peakNUMBER-OF-PEAK-TO-COMPARE-TO_andNUMBER-OF-SECOND-PEAK-FOR-CORRELATION_correlation_window the next part is the same file format as the single channel, but it ends with words filtered,unfiltered, and raw

• This means we can extract additional information of the peak numbers, correlation window, and filtered or unfiltered (or raw) to be able to sort all data coming from those folders to be in the same place
• • Example: in data-photon-counts-SiPM there is a folder named 20250505_coic_correlation_time_varyand it might have a subdirectory that should look like this:peak4_and4_50ns_correlation_window_65_7_gain_1_6_pulse_60s_filtered
    • peak4_and8: parces out to be comparing peak number 4 and correlation with the other detector peak 8
      • 4 peak comparing to
      • 8 peak checking correlation between photons
      • 50ns_correlation_window is the correlation window I used to count when photons are considered simultaneous, denoted by being in front of those words correlation_window
      • We still get the other information using single channel analsis (voltage gain, pulse height, duration of experiment)
      • at the end of the name is filtered or unfiltered or raw is the type of data we are using. This allows us to make comparisons between filtered and unfiltered data
        • filtered means that the data was filtered to remove noise (by ComPass)
        • unfiltered means that the data was not filtered (by ComPass)
        • raw raw data from CoMPass

Output: crops, identifies peaks, and fits the curve

It produces an two images (one for CH0 and one for CH1) that looks like this:

In our initial analysis, we wanted to be able to distinguish between the peaks with no light, the "dark", and compare it to different pulse heights. There are peaks with no light because these are thermal photons that are being detected once the instrument turns on. We compared the lowest possible level of light to see how it compared to the dark.

• The initial analysis used 1.0V pulse height and 65.7V gain voltage. The lsb settings on each channel was different from the latest ones now.

• 

There is barely any light, but can see it the most when zooming into 65.7V plot above 2nd and 3rd peaks We also wanted to compare the dark with differnt light levels (higher pulse height) to see the difference.

First with a 1.3V pulse and further with a 1.6V pulse height

This comparison didn't turn out to be that helpful initially, but seeing the curve get cut off at the beginning , putting all those counts in the first bin where they should not be. So, we learned how critical it is to have set the right lsb settings for the channels in ComPass.

HOW TO AVOID: I typically move my lsb settings to put the curve way out to the right. There are a total of 4000 ADC channels so there is a lot of wiggle room to put the peaks to never have to think about this issue

For each set of scripts in each analysis folder, some scripts need to run before others because they generate more results for other scripts. So, starting by order

• Purpose: Analyzes the relationship between index and peak values.
• Description By analyzing the peak index versus peak number, we are looking fow how the position of the peak (the index) changes with each peak number. A constant slope would mean that each peak moves at the same rate, so increasing the peak number doesnt change the spacing
• Functionality:
  • Reads from .csv from plot-fit-peaks-SiPM-data.py
  • Generates a plot of index vs peak values.
  • Calculates the mean of the slope of different light levels for each gain voltage
  • Exports calculated slopes data to a CSV file It uses the means calculated in plot_index_vs_peak.py between all the lines representing different pulse height for each gain and channel. Outputs the mean of the light levels spacing as a function of gain voltage

There are no free parameters for the user to change in this script, those are set by running plot-fit-peaks-SiPM-data.py and the data this script uses comes from a folder generated_peak_data_results which is the combined peak data generated by above mentioned main script

And produces a plot for every gain and every channel (turn off interactivity if you dont want to annoying click exit on all of them)

And example of how the gain and channel being changed looks like:

• #TODO double check labels 1.3 showing up for both.

Description: This python file will plot the index vs peak data for each channel and save the plots in the folder 'index_vs_peak_data_results'. It will also save the data in a csv file in the same folder.

• Purpose: Examines the spacing between peaks and their corresponding values.
• Functionality:
  • plots the spacing between peaks using the slopes calculated in index_vs_peak.py
  • It uses the means calculated in plot_index_vs_peak.py between all the lines representing different pulse height for each gain and channel. Outputs the mean of the light levels spacing as a function of gain voltage
• Usage: Run this script to study the spacing between peaks in the data, using the method of extracting the mean slopes of light levels from .csv outputted by index_vs_peak.py

This script generates a plot that looks like this:

• Purpose: Plots position of first peak for each pulse height as a function of gain voltage
• Functionality:
  • Reads in csv file generated from plot-fit-peaks-SiPM-data.py which contains all the first peak position data in a separate .csv
  • Reads in first peak data per pulse height and plots it
  • It uses the means calculated in plot_index_vs_peak.py between all the lines representing different pulse height for each gain and channel. Outputs the mean of the light levels spacing as a function of gain voltage
• Usage: Run this script to study the spacing between peaks in the data, using the method of extracting the mean slopes of light levels from .csv outputted by index_vs_peak.py

This script generates a plot that looks like this:

3.) Now you can run any coincidence analysis. The first file to run before running the others is 'plot_coic_addback.py'. and then you can analyze the weighted mean plot and/or the correlation window plot

• data_dir = 'photon_counts_data' → directory of data files
• pulse_voltages_to_plot = [1.6] → only plots those pulse voltages
• gain_voltages_to_plot = [ 65.7] → only plots those gain voltages
• crop_off_start = 100 → clean noisy start
• crop_off_end = 3000 → clean noisy end
• vertical_lines = False → whether to have vertical lines
• counts_threshold = 100 → tune peak detection
• peak_spacing_threshold = 16 → tune peak detection
• sigma = 3.6 → smoothness of curve
• pulse_color_map = {1.0: 'black', 1.1: 'darkblue', 1.3: 'green',1.6: 'orange', 2.0: 'deeppink', 2.3: 'red',} →
• manual_peak_indices = {('CH0', 65.7, 1.6): [140, 178, 553, 590], ('CH1', 65.7, 1.6): [130, 177],} → force specific peaks to be counted if auto fails
• data_by_channel = {"CH0": defaultdict(lambda: defaultdict(list)), "CH1": defaultdict(lambda: defaultdict(list))}pulse_by_voltage = defaultdict(lambda: defaultdict(float)) →