A Modular Datalogger and Slow-Control Platform for Physics Experiments with Time-Series Telemetry and Web Dashboards

This paper presents a modular NIM 2U slow-control system for nuclear physics experiments that integrates a base controller with up to three extension boards to monitor diverse environmental and facility parameters via high-precision 16-bit ADCs, while utilizing a software chain to publish time-series telemetry to Graphite and visualize data through Grafana web dashboards.

The Gist

Imagine you are running a very delicate, high-stakes science experiment in a laboratory. It's like trying to bake the perfect soufflé, but instead of an oven, you have a giant particle accelerator, and instead of eggs, you have subatomic particles.

If the room gets too hot, the soufflé collapses. If the pressure drops, the recipe fails. If the electrical current wavers, the whole thing is ruined. To prevent this, scientists need a "slow-control system"—a vigilant watchdog that constantly checks the temperature, pressure, humidity, and electrical currents, 24/7.

This paper introduces a new, super-smart watchdog designed specifically for nuclear physics experiments. Here is how it works, explained simply:

1. The "Lego" Box (The Hardware)

Think of this system as a modular Lego box that fits into a standard rack in a lab.

• The Base: It comes with a main board that has 8 slots specifically designed to read temperature sensors (like the ones in your thermostat, but much more precise).
• The Extensions: Just like adding extra Lego bricks, you can snap up to three extra boards onto the main one.
  • One board can read standard industrial signals (like pressure gauges that use 4–20mA currents).
  • Another can measure tiny electrical currents (like a super-sensitive scale for electrons).
  • Another can just add more temperature sensors.

The Magic Trick: No matter which board you plug in, they all speak the same language. They all use the exact same high-precision "ruler" (a 16-bit digital converter) to measure things. This means if you compare a temperature reading from the base board with a pressure reading from an extension board, you know they are equally accurate and trustworthy. It's like using the same ruler to measure both the length of a table and the width of a book, so you never get confused by different units.

2. The "Brain" (The Computer)

Inside this box is a small, rugged computer (a BeagleBone Black). Think of it as the conductor of an orchestra.

• It wakes up the sensors.
• It asks them, "What's the temperature?" or "What's the pressure?"
• It takes the answers, converts them into digital numbers, and gets ready to send them out.

3. The "Postman" (The Software & Internet)

Once the computer has the data, it doesn't just sit there. It acts like a super-efficient postman.

• It sends the data over the internet (Ethernet) to a giant digital notebook called Graphite. This notebook stores every single measurement ever taken, creating a perfect history of the experiment.
• Then, it shows this data on a Grafana dashboard. This is like a live TV screen in the control room. Instead of looking at raw numbers, the scientists see colorful graphs, lines, and charts. If the temperature starts to rise too fast, the graph turns red, and an alarm might go off.

4. Why is this a Big Deal?

Before this system, scientists often built their own "one-off" monitoring tools. It was like every scientist building their own custom car engine from scratch. They worked, but they were hard to fix, hard to understand, and didn't always talk to each other.

This new system is like a standardized, reliable car engine that anyone can buy and install.

• It's Modular: You only build what you need.
• It's Consistent: Every measurement is done with the same high-quality "ruler."
• It's Historical: It saves every moment of data, so if something goes wrong today, scientists can look back at the graphs from last week to figure out why.

The Bottom Line

This paper describes a "Swiss Army Knife" for physics labs. It's a compact, plug-and-play box that keeps a constant, super-accurate eye on the environment of a nuclear experiment. By turning messy, analog signals into clean, digital history, it helps scientists focus on the big discoveries rather than worrying about whether their equipment is overheating.

Technical Summary

1. Problem Statement

Nuclear physics experiments require robust, reliable monitoring of auxiliary parameters (temperature, humidity, vacuum pressure, currents) to ensure apparatus stability and data quality. Existing solutions often suffer from:

• Ad-hoc designs: Custom-built systems that are poorly documented and difficult to maintain.
• Metrological inconsistency: Heterogeneous channels often use different ADCs and references, making cross-channel comparison and calibration difficult.
• Lack of historical context: Many systems lack integrated time-series storage, hindering long-term diagnostics and the correlation of anomalies with operational shifts.
• Integration complexity: Difficulty in standardizing interfaces for diverse sensor types (RTDs, industrial 4–20 mA/0–10 V, and low-level currents) within a single compact unit.

2. Methodology and System Architecture

The authors present a modular, NIM 2U slow-control system designed to unify acquisition, control, and telemetry. The system is built around a "metrologically uniform" philosophy, ensuring consistent performance across all channel types.

Hardware Architecture

• Form Factor: Standard NIM 2U module housing a base controller board and up to three plug-in extension boards.
• Base Controller:
  • Features 8 native channels for PT100/PT1000 RTD probes.
  • Hosts the BeagleBone Black (BBB) embedded controller for acquisition logic, bus management, and network communication.
  • Provides power conditioning (generating ±5V, +3.3V, +5V rails) and distributes a common reference voltage.
• Extension Boards:
  • RTD Extension: Adds 8 more PT100/PT1000 channels.
  • Industrial Signal Extension: Supports 8 channels of 4–20 mA (2/3/4-wire) and 0–10 V inputs with manual per-channel selection.
  • Current Reader: Dedicated to low-current measurements (500 nA – 100 µA) with automatic gain-range selection via I2C.
• Uniform Conversion Chain:
  • All boards utilize the AD7689 16-bit SAR ADC.
  • A common external 2.5 V reference (REF192) is distributed to all boards, ensuring a uniform Least Significant Bit (LSB) of 38.15 µV regardless of the sensor type.
  • Analog and digital grounds are strictly separated to minimize noise coupling.
• Connectivity: Front-panel SMA connectors provide shielded, robust connections. Internal flat cables and a backplane distribute power and SPI/I2C buses.

Software and Telemetry Stack

• Acquisition Layer: The BBB firmware handles SPI readouts, channel configuration, and metric generation.
• Transport Layer: Data is published via Ethernet using the Carbon protocol to a Graphite time-series database. Timestamps are synchronized via NTP/SNTP.
• Visualization Layer: Grafana provides web-based dashboards for real-time monitoring, historical analysis, and alerting.
• Naming Convention: A hierarchical, dot-separated metric naming scheme ensures semantic stability across experiments, allowing dashboards to be reused with minimal reconfiguration.

3. Key Contributions

• Metrological Uniformity: The systematic use of a single ADC model and a shared voltage reference across heterogeneous channels eliminates scale mismatches, simplifying calibration and inter-channel correlation.
• Modularity: The NIM 2U form factor with hot-swappable extension boards allows rapid adaptation to specific experimental needs without redesigning the core system.
• Integrated Telemetry: The seamless integration of embedded control with Graphite/Grafana provides a "laboratory-ready" pipeline for continuous historical recording, moving beyond simple alarm triggers to deep operational analysis.
• Versatile Input Support: The platform natively supports RTDs, standard industrial process signals, and low-level diagnostic currents within a single chassis.

4. Performance and Results

The paper presents preliminary quantitative characterizations:

• RTD Chain (PT100/PT1000):
  • Precision: Short-term precision of 0.0166% (approx. 4.16 m°C) relative to a 25°C reference.
  • Stability: Long-term drift of 5.68 ppm/day.
  • Inter-channel Coherence: In ambient conditions, channels overlap within 0.3–0.4°C deviation.
• Current Chain (500 nA – 100 µA):
  • Linearity: Linear fit yields $I_{read} = 0.99506 I_{set} + 10^{-7}$.
  • Accuracy: Gain error of -0.494% and an offset of 100 nA. Maximum absolute error at full scale (120 µA) is ~0.411%.
• Industrial Chains (4–20 mA / 0–10 V):
  • Functional validation confirmed correct scale interpretation, absence of saturation, and the ability to track slow variations and transients (e.g., vacuum pump cycles).
  • Successfully interfaced with vacuum gauges and infrared thermometers.
• System Diagnostics: Internal power rails are monitored as time-series data, enabling the distinction between sensor drift and power supply anomalies.

5. Significance

This work addresses a critical gap in experimental physics infrastructure by providing a standardized, reproducible, and maintainable slow-control solution.

• Operational Efficiency: The web-based interface reduces the need for direct access to the embedded controller during shifts, while the time-series backend simplifies the identification of anomalous conditions.
• Scalability: The system can scale by running multiple modules in parallel, all feeding a central Graphite backend, provided unique metric namespaces are used.
• Future-Proofing: The open platform design allows for the development of new extension boards as experimental requirements evolve, preventing the obsolescence of the core control unit.

The system has already been deployed in the 8Be–X17 experiment at the Legnaro National Laboratories, demonstrating robustness and practical utility in a real-world physics environment.