Lab 8 Part 2: Sensor Calibration

Learn sensor calibration techniques and characteristic curve analysis

Learning Goals

  • Create scripts and functions in Blockly
  • Learn and practice:
    • How to calibrate a sensor
    • How to find the characteristic curve of a sensor

Software and Hardware

  • BlocklyProp Solo
  • Activity Board and Parallax USB programming cable
  • PAR sensor
  • Calibrated PAR sensor - Quantum meter

What is PAR?

Photosynthetically active radiation (PAR), which is solar radiation ranging from 400 to 700 nm, is the most important energy source for crops, used by them for photosynthesis. A low or high PAR intensity can deteriorate photosynthesis; hence, it may become a stress factor.

Background

Sensor calibration is required to ensure accurate measurements by the sensor. Therefore, it is necessary to minimize the sensor's error, i.e., to minimize the difference between the measured value by the sensor and the actual value of the quantity being measured.

The purpose of calibration is to find or update the characteristic curve, which defines the relationship between the quantity being measured and the sensor's output. Calibration must be performed in a timely manner to prevent performance degradation and ensure accurate measurements.

Multi-point Curve Fitting

The objective is to calibrate the PAR sensor, which returns raw numbers. The goal is to find a characteristic curve that maps the light intensity in µmol/(m²s) to the sensor's output.

Calibration Steps:

  1. Take measurements with the PAR sensor at various light intensities (low, medium, high)
  2. Repeat measurements with a calibrated reference instrument under the same conditions
  3. Record measurements in a table
  4. Plot sensor measurements versus actual light intensity
  5. Fit a linear model to obtain the characteristic curve
Calibrated and Uncalibrated Sensors Quantum Meter

Excel Data Analysis

Creating Scatter Plots

In Excel: Select the two columns > Insert tab > Charts > Scatter.

Excel Scatter Plot Creation

Adding Trendlines

Right click on one of the points > add trendline > select linear model > Display Equation and R-squared value on Chart

Adding Trendline Trendline Options

Two-Point Calibration

Use the first and last row of your data (extreme values) to find a calibration model using the two-point calibration equation:

corrected value = (raw value - RawLow) × (ReferenceRange/RawRange) + ReferenceLow

Since RawLow and ReferenceLow are both zero, the equation simplifies to:

corrected value = (raw value) × (ReferenceRange/RawRange)

What to Submit

  1. Measurements Table: Your data entered into the calibration table (no points)
  2. Scatter Plot: Plot showing characteristic curve equation (20 points)
  3. Inverse Function: Calculate inverse of characteristic curve (10 points)
  4. Calibration Tables: Complete both calibration method tables (20 points)
  5. RMSE Analysis: Calculate RMSE for each technique and determine which is better (10 points)
  6. Sensor Sensitivity: Determine sensitivity and its unit (10 points)
  7. Sensor Offset: Identify if sensor has bias/offset (10 points)
  8. Maximum Range: Calculate maximum measurable light intensity (10 points)
  9. Calibrated Plot: Plot calibrated vs. actual values with trendline analysis (10 points)

Submit responses as Word/PDF file