Studies of the Modular COsmic Ray Detector (MCORD) using an automatic temperature control loop to maintain constant gain parameters of semiconductor SiPM photomultipliers

This paper evaluates and optimizes automatic temperature control strategies for the Modular Cosmic Ray Detector (MCORD) to ensure stable SiPM gain performance, presenting the most effective configuration alongside recent hardware and software updates.

The Gist

Imagine you have a very sensitive microphone (the SiPM) that is designed to hear the faintest whispers of cosmic rays hitting a plastic block (the scintillator). This microphone is so sensitive that it's like a high-strung violin string: if the room gets even a little bit warmer, the string stretches, the pitch changes, and the sound becomes distorted. If the room gets colder, the string tightens, and the pitch shifts the other way.

In the world of particle physics, this "pitch" is called gain. If the gain shifts, your detector can't tell the difference between a real cosmic ray and random noise, or it might misjudge the energy of the particle.

This paper is the story of how the scientists at the National Centre for Nuclear Research in Poland built an automatic thermostat for their cosmic ray detector to keep that "pitch" perfectly steady, no matter how the weather changes.

Here is the breakdown of their journey, explained simply:

1. The Problem: The "Fickle" Microphone

The detector they are building (called MCORD) uses Silicon Photomultipliers (SiPMs). These are amazing sensors, but they have a major flaw: they hate temperature changes.

• The Analogy: Imagine trying to tune a radio to a specific station. If the temperature in the room changes, the radio drifts off the station. You have to constantly twist the dial to get back to the music.
• The Reality: As the temperature goes up, the SiPM needs less voltage to work correctly. As it goes down, it needs more. Without help, the detector's readings would be all over the place.

2. The Solution: The "Smart Thermostat" (Temperature Loop)

The team built a software system called a Temperature Loop (TL). Think of this as a smart home thermostat that doesn't just turn the heat on or off, but constantly adjusts the voltage to the sensor to keep the "pitch" perfect.

• How it works: The system constantly checks the temperature near the sensor. If it sees the temperature rising, it automatically lowers the voltage. If it gets cold, it raises the voltage. It does this mathematically using a "correction factor."

3. The Experiment: Building a "Mini-Detector"

The actual detector is huge (the size of a desk), so they couldn't fit it inside a giant climate-controlled oven (a climate chamber) to test it.

• The Analogy: You can't test a new car engine in a tiny garage, so you build a perfect, scaled-down model of the engine to test in the lab.
• The Reality: They built an Equivalent Detector (ED). It's a tiny, 3D-printed box containing small plastic scintillators, the sensitive sensors, and a radioactive source (Na-22) that acts like a "fake cosmic ray" to test the system. They put this little box inside the climate chamber and cranked the temperature up and down to see if their thermostat software worked.

4. The "Compton Edge": The Perfect Tuning Fork

How do you know if the detector is tuned correctly? You can't just look at the numbers; you need a reference point.

• The Analogy: Imagine a musician playing a note. To know if they are in tune, they compare it to a tuning fork.
• The Reality: They used the Compton Edge. This is a specific, sharp "edge" in the energy spectrum of the radiation. It's like a distinct peak on a graph. If the detector is working right, this peak stays in the exact same spot on the graph, even if the temperature changes. If the peak slides left or right, the detector is out of tune.

5. The Tweaks: Finding the Perfect Settings

The team didn't just turn the thermostat on; they tried different settings to see what worked best. They treated the software like a recipe, testing different ingredients:

• The "Dead Band" (The Threshold): This is the rule: "Only adjust the voltage if the temperature changes by at least X degrees."

  • The Lesson: If you set this too low (e.g., 0.1°C), the system gets jittery, constantly making tiny adjustments like a nervous driver tapping the brakes. If you set it too high (e.g., 3°C), the system is too lazy, and the detector drifts out of tune.
  • The Winner: They found the "Goldilocks" zone: 0.5°C. This stops the jitter but keeps the detector perfectly tuned.

• The "Averaging Time": How long should the system wait to see if the temperature change is real or just a glitch?

  • The Lesson: They tried waiting 0.1 seconds, 1 second, and 10 seconds.
  • The Winner: It didn't matter much! The system was robust enough that it didn't need to overthink the data.

• The "Math Method": Should the computer use a simple average, a weighted average, or a complex geometric mean to calculate the temperature?

  • The Lesson: Again, it didn't matter. The simple math worked just as well as the complex math.

6. The Hardware Fix: Quieting the Static

Before the software tests, they had to fix a hardware problem. The electronics were so sensitive that they were picking up "static noise" (like radio static) when measuring tiny currents.

• The Fix: They added a few tiny capacitors (electronic components that smooth out electricity) to the circuit.
• The Result: The noise dropped by 10 times. It's like going from a room with a buzzing fluorescent light to a silent library. This made their measurements incredibly precise.

The Big Takeaway

The paper concludes that they successfully built a detector that can survive in the real world, where temperatures fluctuate. By using a smart software loop that adjusts the voltage based on temperature, they ensure that the detector's "ears" stay perfectly tuned.

In a nutshell: They took a finicky, temperature-sensitive sensor, built a tiny test version of it, figured out exactly how much the temperature changes its behavior, and wrote a smart computer program that acts like an automatic tuner, keeping the detector perfectly calibrated no matter how hot or cold it gets. This ensures that when they finally use it to catch cosmic rays, the data will be clean, accurate, and reliable.

Technical Summary

1. Problem Statement

The Modular Cosmic Ray Detector (MCORD) utilizes Silicon Photomultipliers (SiPMs) for light detection in plastic scintillators. A critical challenge for SiPM-based systems is their high sensitivity to ambient temperature variations.

• Gain Instability: SiPM gain is linearly dependent on the over-voltage, which is determined by the breakdown voltage ($V_{br}$). Since $V_{br}$ shifts with temperature (typically ~50 mV/°C), constant bias voltage leads to significant gain fluctuations in varying environments.
• Measurement Noise: Previous iterations of the MCORD Analog Front End (AFE) electronics exhibited significant noise (20–30 bits) at very low currents, affecting the reproducibility of breakdown voltage measurements and calibration.
• Lack of System-Level Calibration: While SiPM manufacturers provide a nominal temperature coefficient, the combined effect of the SiPM, scintillator, and AFE electronics on the total system gain had not been empirically verified for this specific detector configuration.

2. Methodology

The authors employed a comprehensive experimental and software development approach to address these issues:

• Experimental Setup (Equivalent Detector - ED):
  • Due to the size constraints of the full MCORD detector, a scaled-down "Equivalent Detector" (ED) was 3D-printed and housed in an aluminum enclosure to mimic the thermal and electrical properties of the full system.
  • The ED was placed inside a Binder MK53 climate chamber capable of precise temperature control (-40°C to +180°C).
  • Radiation Source: Na-22 sources (1 MBq) were used to generate 511 keV gamma rays. Coincidence detection between two scintillator tiles allowed for the monitoring of the Compton Edge in the energy spectrum as a proxy for detector gain stability.
• Hardware Modifications:
  • AFE Noise Reduction: Capacitors (1 µF and 10 µF) were added to the LTC6101HV operational amplifier circuit in the Low DropOut (LDO) regulator block. This increased the measurement time constant, reducing current fluctuation noise from ~20 bits to <1 bit.
• Software Development:
  • Rewrite: The AFE and HUB firmware were completely rewritten (C/STM32F0 for AFE, MicroPython for HUB) to support advanced control loops.
  • Temperature Loop (TL): A closed-loop feedback system was implemented. It measures the temperature near the SiPM, averages the data, and adjusts the bias voltage ($V_{bias}$) to maintain a constant gain.
  • Algorithms: Various averaging methods (arithmetic mean, weighted exponential moving average, etc.) and threshold criteria (dead-band) were tested to optimize loop stability.
• Calibration Strategy:
  • Step 1: Measured the SiPM breakdown voltage vs. temperature to determine the raw sensor coefficient.
  • Step 2: Measured the shift of the Compton Edge vs. temperature for the entire system (SiPM + Scintillator + AFE) to derive the effective system temperature coefficient.

3. Key Contributions

• System-Level Temperature Coefficient Determination: The study demonstrated that the manufacturer's nominal coefficient (52 mV/°C) is insufficient for the full system. Through Compton Edge analysis, the authors determined the effective system coefficient to be 62 mV/°C (a ~20% increase), accounting for thermal effects on the scintillator and electronics.
• AFE Hardware Correction: Identified and resolved a noise floor issue in the low-current regime by modifying the LDO circuit, significantly improving measurement repeatability for breakdown voltage determination.
• Robust Compton Edge Analysis: Developed a novel fitting method (using piecewise functions and NonlinearModelFit in Mathematica) to accurately determine the Compton Edge position even with low statistics (30–40 counts), enabling rapid testing cycles.
• Advanced Temperature Control Firmware: Implemented a flexible TL loop with configurable parameters (dead-band, averaging time, averaging algorithm, and voltage ramping) to prevent oscillations and ensure smooth voltage transitions.

4. Key Results

• Gain Stability: The Temperature Loop successfully maintained the Compton Edge position within measurement uncertainty during multi-day experiments where ambient temperature fluctuated between 15°C and 30°C.
• Threshold Sensitivity (Dead-Band):
  • A dead-band (threshold) is essential to prevent the loop from making continuous, unnecessary micro-adjustments.
  • Optimal Value: A threshold of ≤ 0.5°C was identified as optimal. At 0.5°C, the deviation in the Compton Edge was <5%. Increasing the threshold to 3°C caused deviations exceeding 10%, rendering the detector unstable.
• Averaging Parameters:
  • Time Window: The duration of temperature data collection (tested at 0.1s, 1s, 10s) had no significant impact on the final gain stability under normal laboratory/outdoor conditions.
  • Algorithm Type: The choice of averaging algorithm (arithmetic mean vs. weighted exponential) showed negligible differences in performance.
• Voltage Ramp Control: The implementation of a DAC ramp controller ensured that voltage changes occurred in small, configurable steps, preventing sudden shocks to the detector.

5. Significance

This work provides a critical blueprint for the deployment of SiPM-based detectors in variable thermal environments (e.g., cosmic ray monitoring, space applications, or field experiments).

• Reliability: It proves that automatic temperature compensation is not only feasible but necessary for maintaining detector calibration over long periods.
• Optimization: By identifying that the temperature threshold (dead-band) is the most critical parameter, while averaging time and method are secondary, the authors provide clear guidelines for configuring future detectors.
• Holistic Approach: The paper emphasizes that system-level calibration (including scintillators and electronics) is required, as relying solely on sensor datasheets leads to suboptimal performance.
• Foundation for MCORD: These findings, combined with the hardware noise reduction and software overhaul, finalize the MCORD detector's readiness for its intended applications in cosmic muon detection and veto systems.