Energy Consumption Monitoring in Switched-Mode Power Supplies Using ESP32 and Hall Effect Sensors

The Challenge of Measuring Power in High-Frequency Supplies

In modern electronic engineering, thermal management and energy efficiency are no longer optional perks but fundamental design requirements. Switched-mode power supplies, technically known as SMPS, power virtually all modern electronic devices, ranging from small phone chargers to robust industrial servers. They operate by switching transistors on and off at frequencies ranging from tens to hundreds of kilohertz, ensuring reduced physical dimensions and high efficiency. In practice, this means energy is not wastefully dissipated in heavy step-down resistors but intelligently modulated. However, this high frequency and the dynamic nature of the load make accurate energy consumption monitoring a fascinating technical challenge full of pitfalls for the designer.

When we need to audit the actual consumption of equipment powered by a switched-mode supply, traditional instrumentation fails or becomes unsafe. Conventional plug-in power meters offer a macro, unintegrated view, while standard multimeters cannot capture rapid current variations caused by load transients. The solution involves integrating non-intrusive or isolated current sensors directly into the power supply circuit, connected to a low-cost, high-performance microcontroller like the ESP32. This approach allows us to collect voltage, current, and power data in real time, transmitting everything to a cloud infrastructure or local server for behavioral analysis and operational anomaly detection.

Understanding the Hall Effect Principle in Practice

To measure electrical currents without physically breaking the energy path or risking damage to the microcontroller, we use the Hall effect. First described in the 19th century, this physical phenomenon occurs when an electric current passes through a conductor immersed in a transverse magnetic field, generating a perpendicular voltage proportional to that field's intensity. In practice, the ACS712 sensor—one of the most popular in hobbyist electronics and industrial prototyping—features an internal copper conductor through which the measured current flows. A tiny silicon sensor chip placed directly above this conductor captures the magnetic field generated by moving electrons and translates that magnetic force into a readable analog voltage signal.

The great asset of this technology is galvanic isolation. In practice, this means the high-power circuit (handling mains electricity or high supply currents) remains physically separated and insulated from the low-power circuit (where the ESP32 operates at 3.3 volts). If a catastrophic short circuit or severe surge occurs on the load line, the sensor's dielectric barrier prevents overvoltage from destroying the microcontroller or endangering the operator. However, the designer must deal with an important trade-off: Hall effect sensors are susceptible to external magnetic interference, demanding extra care with high-current trace routing and the proximity of transformers or electric motors on the printed circuit board.

Hardware Architecture: Connecting the ESP32 to the Analog World

The ESP32 microcontroller is widely known for its integrated Wi-Fi and Bluetooth connectivity, but its analog-to-digital converter, the ADC, has known linearity limitations, especially at extreme ranges near zero and the supply rail. To mitigate this behavior and ensure reliable current readings in switched-mode supplies, the signal conditioning circuit requires surgical attention. The Hall effect sensor output typically provides an analog voltage centered on a fixed reference (for example, 2.5 volts for a 5-volt supply), oscillating up or down depending on the direction and magnitude of the monitored current.

Since the ESP32 operates on 3.3-volt logic at its input pins, directly connecting a sensor's 5-volt output without a proper resistive voltage divider or a voltage follower operational amplifier will result in immediate saturation and potential GPIO burn-out. On the bench, we use 1% precision resistors to perform voltage level shifting, ensuring the total excursion of the sensor signal fits comfortably within the ESP32's 0 to 3.3-volt measurement range. Additionally, adding a 100-nF ceramic decoupling capacitor close to the power and signal pins helps filter out high-frequency noise generated by the monitored power supply's own switching action.

Sampling Processing and Effective Power Calculation

Collecting voltage and current data in isolation is not enough to calculate the real power consumed by a switched-mode supply, especially if the load is inductive or exhibits harmonic distortions typical of cheap switched supplies without power factor correction (PFC). The ESP32 features two 32-bit Xtensa processing cores, allowing us to dedicate one core exclusively to high-speed analog signal sampling, while the second core manages network connectivity and the user interface. In practice, this ensures the acquisition loop is never stalled by TCP/IP packet delays or unstable Wi-Fi reconnections.

To calculate real power in watts, the firmware computes the RMS (Root Mean Square) value for both voltage and current. Sampling must be rapid—ideally collecting hundreds of points per AC mains cycle (50Hz or 60Hz)—to faithfully capture the distorted waveform entering the switched supply. By multiplying instant voltage and current samples point by point and averaging this product over a complete period, we obtain the actual active power consumed. The conceptual formula implemented in code calculates the root mean square of raw readings minus the resting offset, converting raw digital output into real physical units using calibration constants previously obtained with a standard bench multimeter.

Practical Firmware Implementation for Data Acquisition

Developing code for the ESP32 requires using FreeRTOS real-time operating system tasks to ensure deterministic sampling. Below is a functional C++ snippet using the Arduino IDE, configuring analog reading and basic RMS current calculation in a dedicated task.

#include <Arduino.h>const int SENSOR_PIN = 34;const float VCC = 3.3;const int ADC_RESOLUTION = 4095;const float SENSITIVITY = 0.185; // Sensitivity for 5A ACS712 (185 mV/A)const float ZERO_V_OFFSET = 2.5; // Theoretical rest voltage adjusted by dividervoid setup() {  Serial.begin(115200);  pinMode(SENSOR_PIN, INPUT);  analogSetAttenuation(ADC_11db); // Set attenuation for 0-3.3V rangem}void loop() {  long sumSquares = 0;  int samples = 1000;  for (int i = 0; i < samples; i++) {    int rawValue = analogRead(SENSOR_PIN);    float voltage = (rawValue * VCC) / ADC_RESOLUTION;    float current = (voltage - (ZERO_V_OFFSET * 3.3 / 5.0)) / SENSITIVITY;    sumSquares += (current * current);    delayMicroseconds(100);  }  float rmsCurrent = sqrt(sumSquares / samples);  Serial.print("RMS Current: ");  Serial.print(rmsCurrent, 3);  Serial.println(" A");  delay(1000);}

This code sets up the correct attenuation for the ESP32 analog-to-digital converter and runs a fast sampling loop to calculate the root mean square of the current. In practice, the obtained value may show slight variations due to thermal fluctuations in the sensor, requiring digital smoothing routines or moving average filters to stabilize the reading displayed on the control panel.

Noise Mitigation and Bench Calibration Challenges

One of the biggest obstacles faced by engineers implementing Hall effect measurement systems near switched-mode supplies is electromagnetic immunity. Switched supplies generate intense electromagnetic fields due to rapid current switching in their inductors and transformers. If the sensor or signal wires are placed too close to these inductive components, the system will suffer from induced noise, resulting in fluctuating current readings even when the load is turned off. In practice, the solution involves using shielded cables for the analog connection, keeping the wires twisted, and physically separating the ESP32 board from the power supply block.

Another critical point is thermal drift. The ACS712 sensor and signal conditioning circuit components experience slight shifts in electrical characteristics as the ambient or internal enclosure temperature rises during continuous operation. To overcome this, the firmware must implement an automatic offset calibration routine whenever the system boots up with no load connected. This routine measures the quiescent voltage level for a few seconds, stores the value as a dynamic reference, and subtracts the deviation from subsequent readings, ensuring long-term high precision without manual intervention.

Final Considerations

Monitoring energy consumption in switched-mode power supplies using Hall effect sensors and ESP32 microcontrollers represents an elegant, low-cost, and highly reliable solution for engineers and advanced enthusiasts alike. By combining the sensor's safe galvanic isolation with the ESP32's dual-core processing power and connectivity, we build a robust telemetry ecosystem capable of exposing operational inefficiencies and preventing premature failures in critical equipment. With careful assembly, close attention to analog signal conditioning, and proper electromagnetic noise filtering, laboratory-grade data can be achieved using accessible, fully cloud-integrated hardware.