pH Sensor Calibration using Linear Regression
📌 The tutorial focuses on using the linear regression method for calibrating sensors, specifically demonstrating it with a pH sensor.
⚙️ Necessary prerequisites include a calibrator (pH buffers of 4, 6.86/7, and 9) and a program to display the sensor's ADC (Analog-to-Digital Converter) output on an LCD.
📈 Data collection involves cleaning the sensor between dips and recording the stable ADC values corresponding to known pH buffer values (e.g., pH 4.01 yielding ADC 222).

Data Analysis and Formula Derivation in Excel
📊 The collected data (known pH values vs. raw ADC values) must be imported into Microsoft Excel.
🔗 After plotting the data, select the 'Scatter' chart type and add a 'Trendline', choosing the 'Linear' method, as the sensor response is expected to be linear for this method to apply.
📝 Activate options to 'Display Equation' and 'Display R-squared' on the chart; the resulting linear equation, in the form $y = mx + c$, is crucial for the next step.

Implementing the Calibration Formula in Arduino
💻 The linear equation derived from Excel ($y = mx + c$, where $y$ is the ADC reading and $x$ is the calibrated pH) must be rearranged to solve for pH ($x$) using the measured ADC value.
🔢 The rearranged formula is manually entered into the Arduino code, typically by defining a `float` variable for the calculated pH, ensuring variables handling decimal values (like ADC and pH) are declared as `float`.
🧪 The variable holding the raw ADC output from the sensor needs to be substituted into the rearranged equation to convert the raw data into meaningful pH units displayed on the LCD.

Key Points & Insights
➡️ Sensor cleanliness is vital: Always clean the sensor with tissue paper between dipping it into different pH buffer solutions to prevent cross-contamination and inaccurate readings.
➡️ Stability requirement: Wait patiently until the ADC reading stabilizes at a single point when dipped in a solution; if the value fluctuates, choose the value that appears most frequently to ensure calculation accuracy.
➡️ Model selection: If the sensor response is non-linear, choose alternative trendline methods in Excel, such as Exponential, Logarithmic, or Polynomial, matching the visual curve of the plotted data.
➡️ Error tolerance: Post-calibration testing showed small errors (e.g., 0.07 at pH 4, 0.05 at pH 7, and 0.01 at pH 9), which are considered acceptable for typical pH measurements.

📸 Video summarized with SummaryTube.com on Jan 11, 2026, 14:05 UTC