ADC Offset Calibration

Question

Problem: ADC Offset Calibration. Solve this hands-on electronics task on EWskills.

Accepted Answer

#include 

class ADCSensor {
private:
    // Fixed calibration offset hidden from the user
    const int offset_val; 

public:
    // Initialize offset to 100
    ADCSensor() : offset_val(100) {}

    // Process raw data into calibrated application data
    int getCalibratedSample(int raw) const {
        // Apply offset
        int calibrated = raw - offset_val;

        // Saturate to 0 (Clamp negative results)
        // This handles cases where noise drops the reading below the offset
        if (calibrated < 0) {
            return 0;
        }
        return calibrated;
    }
};

int main() {
    int N;
    if (!(std::cin >> N)) return 0;

    ADCSensor sensor;

    for (int i = 0; i < N; ++i) {
        int raw;
        std::cin >> raw;

        // Use the encapsulated interface to get clean data
        std::cout << "Sample: " << sensor.getCalibratedSample(raw) << std::endl;
    }

    return 0;
}

Explanation & Logic Summary:

This problem addresses Data Transformation and Validity.

The Context: Raw hardware data is rarely "application-ready." It usually requires bias correction, scaling, or filtering.
The Flaw: The template code performed the math raw - offset inside main(). This exposes the offset value to the user and fails to handle the edge case where noise pulls the reading below the offset (resulting in negative numbers).
The Fix: The getCalibratedSample method encapsulates the correction algorithm. It centralizes the subtraction and adds the critical check if (calibrated < 0) return 0;.
The Benefit: The application layer now deals only with "clean" physics values (0 or positive), oblivious to the hardware calibration details.

Firmware Relevance & Real-World Context:

Sensor Fusion: In IMUs (accelerometers/gyroscopes), raw registers must be biased and scaled before use. Doing this ad-hoc in main leads to inconsistent data.
Unipolar Constraints: Many physical quantities (light intensity, pressure, distance) cannot be negative. Firmware must clamp noise-induced negative values to zero to prevent logic errors downstream.
Read-Only Config: Calibration constants (like offset_val) are often stored in OTP (One-Time Programmable) memory or Flash and should be const or private to prevent accidental modification during runtime.

168. ADC Offset Calibration