CosmicWatch Desktop Muon Detector v3X

• CosmicWatch Desktop Muon Detector v3X is a compact, portable device that integrates pulse-height analysis, coincidence detection, and environmental logging to measure cosmic-ray muons and ionizing radiation.
• It features significant hardware and firmware upgrades including sub-millisecond dead time, a 12-bit ADC for pulse spectroscopy, and configurable data acquisition modes for precise measurements.
• The detector supports both educational and research applications by enabling detailed cosmic ray studies, high-altitude experiments, and time-of-flight analysis with a build cost under $100.

Searching arXiv for the specified detector paper and closely related CosmicWatch papers to ground the article in the cited literature. The CosmicWatch Desktop Muon Detector (v3X) is a compact, low-cost, and portable detector for ionizing radiation, including cosmic-ray muons, built around a plastic scintillator, a silicon photomultiplier (SiPM), custom signal-processing electronics, onboard data storage, an OLED display, environmental sensors, and USB connectivity. In the 2025 technical reference, the v3X is presented as an integrated hardware-and-firmware revision of earlier CosmicWatch detectors, with total component cost under \$100, build time suitable for high school students, and example measurements spanning sea-level flux, angular dependence, pulse-height spectroscopy, high-altitude balloon operation, time-of-flight, and environmental correlations (Axani et al., 16 Aug 2025).

1. Development lineage and design goals

The CosmicWatch line originated as a self-contained, hand-held cosmic ray muon detector intended for research applications and outreach, with material cost under \$100, USB power, recording either directly to a computer or to a microSD card, and Arduino- and Python-based software (Axani et al., 2018). That earlier formulation established the project’s defining constraints: portability, low cost, novice-accessible assembly, and a data path compatible with both standalone logging and host-computer analysis.

Within that lineage, the v3X is explicitly described as building on previous iterations while introducing “significant hardware and firmware improvements” that enhance sensitivity, usability, and data acquisition capability (Axani et al., 16 Aug 2025). The stated deltas relative to v1/v2 are concrete: sub-ms dead time of $400\,\mu{\rm s}/{\rm event}$ instead of $50\,{\rm ms}/{\rm event}$ in v2, a 12-bit ADC with $2.5\,{\rm V}$ reference for pulse-height analysis, on-board live histogramming on the OLED, configurable coincidence/serial/SD modes, environmental logging of temperature, pressure, acceleration, and gyroscope channels, buzzer feedback, and a hardware coincidence port with accidental coincidence rate $\simeq 4.5\times10^{-5}\,{\rm Hz}$ versus $4.5\times10^{-4}\,{\rm Hz}$ (Axani et al., 16 Aug 2025).

A common simplification is to treat CosmicWatch solely as an educational counter. The documented v3X feature set suggests a broader characterization: it remains accessible in cost and assembly, but its architecture is also structured for introductory laboratory and field measurements in particle and astroparticle physics, especially where pulse-height information, coincidence logic, and environmental metadata are useful (Axani et al., 16 Aug 2025).

2. Detector medium, photosensor, and front-end electronics

The active medium is a $5\,{\rm cm}\times5\,{\rm cm}\times1\,{\rm cm}$ slab of doped polystyrene containing $1\%$ PPO and $0.03\%$ POPOP by weight. The scintillator photon yield is given as $\sim10\,000$ photons/MeV, with emission peak at $420\,{\rm nm}$ and absorption cut-off $50\,{\rm ms}/{\rm event}$0. One face is optically coupled to the SiPM, while the remaining faces internally reflect or are foil-wrapped; the assembly is wrapped with aluminum foil reflector plus 3–4 layers of black tape for light tightness (Axani et al., 16 Aug 2025).

The photosensor is an onsemi MicroFC-60035 C-Series SiPM in a $50\,{\rm ms}/{\rm event}$1 package with $50\,{\rm ms}/{\rm event}$2 active area. The SiPM operates at $50\,{\rm ms}/{\rm event}$3 DC, produced from a $50\,{\rm ms}/{\rm event}$4 system supply by a MAX5026 DC–DC booster. At $50\,{\rm ms}/{\rm event}$5 above breakdown, the photon-detection efficiency is reported as $50\,{\rm ms}/{\rm event}$6 at $50\,{\rm ms}/{\rm event}$7, and the gain as $50\,{\rm ms}/{\rm event}$8 per photoelectron. Optical coupling is provided by PMT optical gel or a $50\,{\rm ms}/{\rm event}$9 silicone pad, and the device is mounted on a custom 2-layer SiPM PCB secured by two 0-80 standoffs and connected to the main PCB through a 6-pin connector (Axani et al., 16 Aug 2025).

The analog chain follows a four-stage structure: SiPM output, AC coupling, a non-inverting TPH2502 preamplifier with gain $2.5\,{\rm V}$0, and then a split into a peak detector and a comparator. The peak detector consists of a Schottky diode, capacitor, and buffer feeding the ADC. The comparator threshold is set through MCU PWM filtered by an RC network and used for event triggering. The full 2-layer PCB schematic integrates SiPM bias, analog shaping, comparator, a Raspberry Pi Pico MCU, microSD, OLED, BMP280 and MPU-6050 sensors, an RJ45 coincidence port, and external connectors (Axani et al., 16 Aug 2025).

Subsystem	Implementation	Function
Scintillator	$2.5\,{\rm V}$1 doped polystyrene	Ionization-to-light conversion
Photosensor	onsemi MicroFC-60035 C-Series SiPM	Photon detection and gain
Bias supply	$2.5\,{\rm V}$2 via MAX5026	SiPM overvoltage operation
Preamplifier	TPH2502, gain $2.5\,{\rm V}$3	Pulse amplification
Trigger/readout split	Peak detector + comparator	ADC pulse height and MCU trigger

The design objective is not calorimetry in the high-resolution sense. Rather, the combination of plastic scintillator, SiPM PDE near the scintillation emission peak, and a peak-detector-based ADC path supports minimum-ionizing-particle detection with pulse-height discrimination and low-rate spectroscopy-like analyses (Axani et al., 16 Aug 2025).

3. Firmware architecture and data acquisition

The v3X controller is a Raspberry Pi Pico using dual ARM Cortex-M0+ cores at $2.5\,{\rm V}$4 and a 12-bit ADC. The firmware is explicitly partitioned by function. Core 0 handles real-time acquisition: it polls the comparator output for rising edges, triggers a $2.5\,{\rm V}$5 coincidence window implemented as 32 polls at $2.5\,{\rm V}$6, reads the peak-detector ADC, timestamps the event, flags coincidence, shunts the peak detector capacitor through a MOSFET to reset it, emits a data packet over USB serial, and writes the event into a 4 k-event ring buffer. Core 1 handles auxiliary tasks: OLED updates, BMP280 and MPU-6050 reads, event transfer from the primary to a secondary $2.5\,{\rm V}$7 buffer, batch writes to microSD every 20 events or 10 s, and management of configure.txt parameters including threshold, detector name, LEDs, buzzer, and WiFi (Axani et al., 16 Aug 2025).

The stated maximum sustained event rate is $2.5\,{\rm V}$8 with serial plus SD logging, and $2.5\,{\rm V}$9 in SD-only mode (Axani et al., 16 Aug 2025). This rate capability is materially distinct from earlier CosmicWatch implementations. In the 2018 detector description, acquisition was built around an Arduino Nano, a 10-bit ADC, ASCII serial output at 115200 baud, and dead time governed by the peak-detector discharge loop at roughly $\simeq 4.5\times10^{-5}\,{\rm Hz}$0 per event (Axani et al., 2018). The v3X architecture therefore shifts the platform from simple thresholded event counting toward a more explicitly buffered and concurrent DAQ model.

The operational modes also matter experimentally. The device supports configurable coincidence, serial, and SD modes, on-board live histogramming on the OLED, and environmental logging. A plausible implication is that the instrument is designed not merely to register event counts but to preserve enough context—pulse height, timing, coincidence state, and local conditions—to support offline rate corrections, threshold studies, and cross-detector synchronization (Axani et al., 16 Aug 2025).

4. Signal model, thresholding, and analysis framework

The v3X documentation makes the detector model explicit in terms of deposited energy, scintillation yield, SiPM PDE, and avalanche gain. With $\simeq 4.5\times10^{-5}\,{\rm Hz}$1 the energy deposition in MeV, $\simeq 4.5\times10^{-5}\,{\rm Hz}$2 photons/MeV the scintillation yield, and $\simeq 4.5\times10^{-5}\,{\rm Hz}$3, the number of detected avalanches is

$\simeq 4.5\times10^{-5}\,{\rm Hz}$4

Using SiPM gain $\simeq 4.5\times10^{-5}\,{\rm Hz}$5 and electron charge $\simeq 4.5\times10^{-5}\,{\rm Hz}$6, the event charge and a representative pulse-current form are written as

$\simeq 4.5\times10^{-5}\,{\rm Hz}$7

These relations connect deposited energy to the peak-detector observable used by the ADC (Axani et al., 16 Aug 2025).

For coincidence geometry, the angular acceptance is expressed in terms of the overlap of two $\simeq 4.5\times10^{-5}\,{\rm Hz}$8 slabs separated by $\simeq 4.5\times10^{-5}\,{\rm Hz}$9:

$4.5\times10^{-4}\,{\rm Hz}$0

with the count rate

$4.5\times10^{-4}\,{\rm Hz}$1

The accidental coincidence rate is given by

$4.5\times10^{-4}\,{\rm Hz}$2

and the threshold set in configure.txt is parameterized as

$4.5\times10^{-4}\,{\rm Hz}$3

A linear fit between ADC and $4.5\times10^{-4}\,{\rm Hz}$4 is used for calibration and inverted to obtain SiPM-equivalent pulse height (Axani et al., 16 Aug 2025).

Pulse-height analysis is central to the v3X revision. The SiPM pulse-height spectrum is described by a Landau/Moyal form,

$4.5\times10^{-4}\,{\rm Hz}$5

and in the example data a fit over $4.5\times10^{-4}\,{\rm Hz}$6–$4.5\times10^{-4}\,{\rm Hz}$7 yields $4.5\times10^{-4}\,{\rm Hz}$8, interpreted as $4.5\times10^{-4}\,{\rm Hz}$9 energy loss, with $5\,{\rm cm}\times5\,{\rm cm}\times1\,{\rm cm}$0 (Axani et al., 16 Aug 2025). This links the front-end pulse-height variable to the expected minimum-ionizing energy deposition in a $5\,{\rm cm}\times5\,{\rm cm}\times1\,{\rm cm}$1 plastic slab.

Related CosmicWatch work reinforces that the device should not be understood as a pure threshold counter. A 2019 source-measurement study re-purposed portable CosmicWatch scintillator detectors for $5\,{\rm cm}\times5\,{\rm cm}\times1\,{\rm cm}$2-source activity extraction and source discrimination by examining the voltage spectrum output, after atmospheric-muon calibration and Geant4 cross-checks (Masias et al., 2019). That use case depends on the same underlying fact emphasized by the v3X guide: the detector retains amplitude information rather than collapsing all events into a binary trigger.

5. Measured performance and representative experiments

The v3X technical guide reports a two-detector “full-sky” measurement in configuration $5\,{\rm cm}\times5\,{\rm cm}\times1\,{\rm cm}$3, with $5\,{\rm cm}\times5\,{\rm cm}\times1\,{\rm cm}$4, integrated over $5\,{\rm cm}\times5\,{\rm cm}\times1\,{\rm cm}$5. The total trigger count was $5\,{\rm cm}\times5\,{\rm cm}\times1\,{\rm cm}$6, corresponding to $5\,{\rm cm}\times5\,{\rm cm}\times1\,{\rm cm}$7. The coincident trigger count was $5\,{\rm cm}\times5\,{\rm cm}\times1\,{\rm cm}$8, giving $5\,{\rm cm}\times5\,{\rm cm}\times1\,{\rm cm}$9. The dead time was $1\%$0 of live time, or $1\%$1 (Axani et al., 16 Aug 2025). These values place the detector in the expected range for sea-level muon observations with coincidence selection.

Angular measurements using configuration $1\%$2 were taken for $1\%$3 over 24 h per point, and the coincidence rate was reported to follow $1\%$4 plus a flat accidental background with $1\%$5 (Axani et al., 16 Aug 2025). This is consistent with the broader CosmicWatch literature, where angular-distribution measurements with two-detector coincidence were also fit to $1\%$6 (Masias et al., 2019).

The detector has also been exercised in non-laboratory environments. In a high-altitude balloon flight to $1\%$7, the rate peaked at $1\%$8 at $1\%$9, identified as the Regener–Pfotzer maximum. The altitude–pressure relation was modeled using

$0.03\%$0

with standard-atmosphere constants (Axani et al., 16 Aug 2025). For time-of-flight, two detectors separated vertically by $0.03\%$1 were read out with BNC plus an oscilloscope at $0.03\%$2; from $0.03\%$3 events, the distribution of $0.03\%$4 gave $0.03\%$5 using a Gaussian $0.03\%$6 (Axani et al., 16 Aug 2025).

Environmental channels are not auxiliary decoration but part of the documented measurement program. Barometer and temperature were logged against time, and the rate was found to anti-correlate with pressure with slope $0.03\%$7 (Axani et al., 16 Aug 2025). A closely related CosmicWatch physics exposition reports a barometric coefficient $0.03\%$8, placing the v3X result within the same phenomenological regime (Axani, 2019).

6. Construction, operation, and application domains

The v3X bill of materials totals approximately \$100 and includes the $0.03\%$9 plastic scintillator slab, the onsemi MicroFC-60035 SiPM, MAX5026 DC–DC booster, TPH2502 dual op-amp, Raspberry Pi Pico, BMP280, MPU-6050, precision passives, microSD socket and card, $\sim10\,000$0 SPI OLED, BNC and RJ45 connectors, 2-layer PCB fabrication, aluminum enclosure and faceplates, optical gel or silicone pad, mechanical hardware, and USB cable and wiring (Axani et al., 16 Aug 2025). The required tools are a fine-tip soldering iron, solder paste or wire, multimeter, optional PCB stencils, optional hot air station, wire cutters, tweezers, and a computer USB port. Estimated build time is 3–5 hours by a novice at the high-school or undergraduate level (Axani et al., 16 Aug 2025).

Assembly proceeds through PCB fabrication and inspection, SiPM-board population, main-board population, mechanical integration of the SiPM PCB onto the main PCB, optical coupling of scintillator to SiPM followed by foil and tape wrapping, insertion into the aluminum enclosure, USB power-up and firmware flashing onto the Raspberry Pi Pico, microSD insertion and configure.txt customization, and functional verification through OLED boot information and pulse testing with an LED or $\sim10\,000$1-source using import_data.py or the GUI (Axani et al., 16 Aug 2025). The sequence is notable because it combines conventional detector assembly steps—light-tightness, optical coupling, bias and analog validation—with embedded-system configuration.

The application list in the v3X guide is broad but concrete. Educational exercises include sea-level muon flux measurement compared to $\sim10\,000$2, angular-distribution measurement using the $\sim10\,000$3 law with two-detector coincidence, barometric pressure dependence, high-altitude balloon cosmic-ray profiling, and muon time-of-flight or velocity studies. Research and citizen-science applications listed are distributed muon tomography arrays for soil density mapping, the global radiation monitoring network CREDO, use as a calibration tool for large neutrino detectors, classroom and science-fair kits, and integration with GPS, WiFi, real-time clocks, and magnetometers for advanced networks (Axani et al., 16 Aug 2025).

One recurrent misconception is that a low-cost detector of this scale is restricted to atmospheric muon counting. The published record is more specific. The 2025 v3X guide centers on muon and ionizing-radiation detection with coincidence, pulse-height, and environmental logging (Axani et al., 16 Aug 2025); the 2019 repurposing study shows that CosmicWatch detectors can also be used for radioactive-source measurements and limited source identification through voltage-spectrum analysis after muon-based calibration (Masias et al., 2019). A plausible implication is that the detector’s most significant property is not simply portability, but the coexistence of a physically interpretable scintillator–SiPM response with sufficiently structured DAQ to support multiple small-scale radiation measurements under a common hardware platform.