Build an IoT Smart Energy Meter with ESP32 & SCT-013 Current Sensor

2. Overview & Engineering Objectives

Monitoring real-time electricity consumption is essential for diagnosing high appliance loads, industrial machinery health, and solar net-metering. This IoT Smart Energy Meter clips non-invasively onto live mains wiring without cutting conductors, providing galvanic isolation and fire-safe AC measurement.

Core Engineering Capabilities:

True RMS Current Calculation: Samples the 50Hz AC waveform at 2.5kHz (50 samples per full sinusoidal cycle) over 10 consecutive cycles to compute genuine root-mean-square current.

Burden Resistor Dynamic Range Tuning: Matches the 50mA secondary current output of the SCT-013 to the ESP32’s 0V–3.3V ADC input window using a 33Ω precision metal-film burden resistor.

DC Offset Conditioning: Elevates AC sinusoidal oscillations into the positive ADC window using a dual 10kΩ voltage divider buffered with a 10µF low-leakage capacitor.

Tariff & Cost Projection Engine: Computes real-time kilowatt-hours (kWh) and projects daily/monthly billing in Indian Rupees (₹) based on configurable tariff slabs.

Dual Telemetry Channels: Non-blocking MQTT streaming to Home Assistant alongside a local asynchronous HTTP dashboard.

3. Bill of Materials (BOM)

Component

Quantity

Specification / Rating

Est. Cost (INR)

Sourcing / Affiliate Link

ESP32 DevKit V1

1

30-Pin Dual-Core Xtensa LX6 (Wi-Fi + BLE)

₹380

Buy on Mentora Store

SCT-013-000 CT Sensor

1

100A Max Primary, 50mA Secondary (Current-type)

₹420

Buy on Mentora Store

33Ω Precision Resistor

1

1/4W 1% Metal Film (Burden Resistor)

₹5

Buy on Mentora Store

10kΩ Resistors

2

1/4W 1% Metal Film (Voltage Divider)

₹4

Buy on Mentora Store

10µF Electrolytic Capacitor

1

25V Low-ESR Decoupling / AC Ground Filter

₹6

Buy on Mentora Store

3.5mm Stereo Audio Jack (Female)

1

PCB Mount / Breakout for SCT-013 3.5mm plug

₹25

Buy on Mentora Store

5V 2A USB Power Supply

1

Low-ripple DC wall adapter

₹150

Buy on Mentora Store

Breadboard & Jumper Leads

1 Set

Standard 400-point breadboard + Dupont leads

₹80

Buy on Mentora Store

Total Hardware Cost

~₹1,070

4. Burden Resistor & DC Bias Circuit Mathematics

The SCT-013-000 has a turns ratio of N = 2000:1. At peak rated primary current (I_p = 100A_RMS): $I_p,peak = 100 × √(2) ≈ 141.42 A The secondary peak current produced by the CT coils is: I_s,peak = (I_p,peak)/(2000) = (141.42)/(2000) ≈ 0.07071 A \ (70.71 mA)$

To prevent ADC clipping, the secondary AC voltage swing across the burden resistor must not exceed the ESP32’s 3.3V reference window (V_ADC_ref = 3.3V), centered at a DC bias of 1.65V: $V_peak-to-peak,max = 3.3V implies V_peak,max = 1.65V The ideal burden resistance R{burden}$ is: $$R{burden} = \frac{V{peak,max}}{I{s,peak}} = \frac{1.65\text{V}}{0.07071\text{A}} \approx 23.33,\Omega$$

For general household monitoring (typical mains circuit maxing out at 30A to 50A RMS peak): $I_p = 50A_RMS implies I_s,peak = (50 × 1.4142)/(2000) ≈ 0.03535A R_burden = (1.65V)/(0.03535A) ≈ 46.6,Ω$ A standard 33Ω 1% metal-film resistor allows safe, high-resolution measurements up to 70.7A RMS without any risk of ADC clipping.

5. Software & Numerical DSP Architecture

The firmware executes continuous true RMS numerical integration: $I_RMS = CalibrationFactor × √((1)/(N) ∑_i=1^N left(V_ADC[i] - V_biasright)^2)$

Key Algorithmic Steps:

Sampling Loop: Collects 1,000 analog samples at precisely timed intervals over 200ms (spanning exactly 10 full 50Hz AC cycles).

Dynamic DC Offset High-Pass Filtering: Subtracts the moving midpoint (V_bias ≈ 2048 for 12-bit ADC) from each instantaneous raw sample.

Squaring & Summation: Accumulates squared deviations to calculate the mean-square value.

Phantom Noise Gate: Ambient RF noise and ESP32 ADC non-linearity induce ~0.08A–0.15A of spurious reading when no load is connected. A software noise gate zeroes readings below 0.12A (~28W at 230V).

6. Real-World Troubleshooting FAQs

Q1: Why does my energy meter read 20W to 35W even when no appliance is plugged in?

Root Cause: ESP32 ADC non-linearity near ground and internal Wi-Fi switching hash coupling into the high-impedance bias divider.

Fix:

Add a 10µF capacitor between the DC midpoint (junction of the two 10kΩ resistors) and GND.

In code, enforce a software noise gate threshold (if (currentRMS < 0.15) currentRMS = 0.0;).

Keep the CT signal wires twisted and shielded from the ESP32 Wi-Fi antenna.

Q2: Why does analogRead() produce distorted or locked values when Wi-Fi is transmitting?

Root Cause: Connecting the sensor to an ADC2 pin (GPIO 0, 2, 4, 12–15, 25–27).

Fix: The CT signal MUST be connected to an ADC1 pin (e.g., GPIO 34 or GPIO 35). ADC1 is independent of the Wi-Fi modem.

Q3: Why does clipping the CT sensor around a standard household power cord read 0.00A?

Root Cause: Standard appliance cords enclose both the Live (Phase) and Neutral wires. Current flows forward through Phase and backward through Neutral in equal magnitude; the opposing magnetic fields cancel out completely!

Fix: You MUST clamp the SCT-013 around ONLY ONE wire (either the Phase/Live wire OR the Neutral wire, never both together).

Q4: Why are current readings 10%–20% higher or lower than a multimeter?

Root Cause: Resistor tolerances and ESP32 internal reference voltage variance (which ranges from 1050mV to 1150mV across chips).

Fix: Calibrate the software constant CURRENTCAL by running a known resistive load (e.g., a 1000W electric kettle or 100W tungsten incandescent bulb): $$\text{Cal}{\text{new}} = \text{Cal}{\text{old}} \times \frac{I{\text{actual}}}{I_{\text{measured}}}$$

7. Mentora Monetization & Student Support Hooks

🛒 Official Mentora Smart Energy Monitor Hardware Kit: Pre-soldered conditioning board with 3.5mm stereo jack, 33Ω burden resistor, dual 10kΩ divider, low-ESR 10µF filter capacitor, and authentic SCT-013-000 sensor for ₹1,299.

⚡ ₹49 Mentora Token Vault: Instantly unlock the companion Home Assistant MQTT discovery script, Node-RED dashboard JSON flows, and ZMPT101B real-voltage integration code for true power factor (PF) computation.

🛠️ ₹299 1-on-1 Mentora Live Video Debugging: Schedule a 30-minute session with a power electronics mentor to calibrate custom burden resistor values, safely interface with domestic distribution boards, and integrate industrial Modbus power meters.

Hardware Pinout & Wiring Connections

Hardware Pinout & Wiring Connection Guide

Project 4: Smart Energy Meter & Real-Time Power Monitor

Microcontroller: ESP32 DevKit V1 (30-pin / 36-pin)Current Sensor: SCT-013-000 Non-Invasive Split-Core Current Transformer (100A / 50mA)Analog Front-End (AFE): 33Ω 1% Metal-Film Burden Resistor, Dual 10kΩ Voltage Divider (1.65V Midpoint), 10µF Decoupling CapacitorAnalog Input: GPIO 34 (ADC1 Channel 6)

1. Complete Fritzing Pin Connection Mapping Table

Component

Component Terminal

Connected Node / ESP32 Pin

Wire Color

Circuit Role & Function

Electrical & Safety Constraints

SCT-013 3.5mm Jack

Tip

Analog Sense Node (GPIO 34)

Blue

AC Secondary Current (+)

Connects to burden resistor high side and ESP32 GPIO 34.

SCT-013 3.5mm Jack

Sleeve

1.65V DC Bias Rail

Green

AC Secondary Current (-)

Connects to midpoint of 10kΩ divider.

SCT-013 3.5mm Jack

Ring (if 3-pole)

NC (No Connection)

—

Unused on standard CTs

Leave floating / unconnected.

33Ω Burden Resistor

Terminal 1

GPIO 34 (Tip Node)

—

Shunt Current-to-Voltage

Converts 50mA secondary current to peak-to-peak AC voltage.

33Ω Burden Resistor

Terminal 2

1.65V DC Bias Rail (Sleeve)

—

Reference Node

Clamped between Tip and Sleeve across CT coil.

Top 10kΩ Resistor

Terminal 1

3V3 Rail

Red

DC Voltage Divider High

Pulls bias node toward 3.3V.

Top 10kΩ Resistor

Terminal 2

1.65V DC Bias Rail

—

Divider Midpoint

Sits at exactly 1.65V DC (3.3V / 2).

Bottom 10kΩ Resistor

Terminal 1

1.65V DC Bias Rail

—

Divider Midpoint

Sits at exactly 1.65V DC.

Bottom 10kΩ Resistor

Terminal 2

GND Rail

Black

DC Voltage Divider Low

Ties divider to common system ground.

10µF Capacitor

Anode (+)

1.65V DC Bias Rail

—

AC Ground Filter

Bypasses AC noise on bias line to ground.

10µF Capacitor

Cathode (-)

GND Rail

—

Ground Reference

Negative stripe connected to GND.

2. Circuit Connection Schematic Diagram (ASCII Art)

+3.3V (ESP32 3V3 Pin)

|

|

[10k] (Top Divider Resistor)

|

+-----------------------+-----------------------------+

|                       |                             |

| [10uF] [33 ohm]

| (Electrolytic) (Burden)

| (+) | | (-) |

| | | |

| GND | |

|                          |                          |

|                          |                          |

+-----[SCT-013 Sleeve]-----+ |

| (1.65V Bias Rail) |

| |

[10k] (Bottom Divider Resistor) |

| |

| |

GND |

|

+-----[SCT-013 Tip]-----------------------------------+

|

|

v

ESP32 GPIO 34 (ADC1 Channel 6)

3. Critical Safety & Engineering Guidelines

DANGER: Never Disconnect Burden Resistor Under Load:

A current transformer acts as a constant current source. If the secondary circuit is opened while AC current flows through the primary conductor, the CT core saturates and generates destructive high-voltage inductive peaks (>1,000V) across its terminals. This will destroy the sensor and presents an electric shock hazard. Always solder the 33Ω burden resistor securely.

Single Conductor Clamping Rule:

Always clamp the SCT-013 split-core around only one wire (the Live/Phase wire). Clamping around a two-core or three-core power cord results in the equal and opposite magnetic fluxes of Live and Neutral cancelling each other out completely, yielding a reading of 0.00A.

ADC1 Exclusivity:

GPIO 34 is part of the ESP32’s internal ADC1 peripheral. Never route analog current monitoring signals through ADC2 pins (GPIO 0, 2, 4, 12–15, 25–27) because active Wi-Fi transmission disables ADC2 functionality.

Complete Tested Arduino C++ Firmware

``cpp

/*

• ==============================================================================
• Project: IoT Smart Energy Meter & Real-Time AC Power Monitor
• Microcontroller: ESP32 DevKit V1 (Xtensa 32-bit Dual-Core 240MHz)
• Author: Mentora Maker Lab (Spark Autonomous Content Intelligence)
• License: MIT

*

• Hardware Interface:
• - Current Transformer: SCT-013-000 (100A / 50mA)
• - Burden Resistor: 33 Ohm 1% metal film
• - Voltage Divider: 10k Ohm + 10k Ohm + 10uF filter capacitor (1.65V DC Bias)
• - Analog Input: GPIO 34 (ADC1 Channel 6)
• ==============================================================================

*/

#include <WiFi.h>

#include <WebServer.h>

// -----------------------------------------------------------------------------

// USER CONFIGURATION & CONSTANTS

// -----------------------------------------------------------------------------

const char* WIFISSID = "YOURWIFI_SSID";

const char* WIFIPASSWORD = "YOURWIFI_PASSWORD";

// Hardware Pin Definition (MUST be an ADC1 pin to avoid Wi-Fi driver conflict)

#define CTANALOGPIN 34

#define STATUSLEDPIN 2

// Grid & Tariff Parameters (India Standard)

const float MAINSVOLTAGERMS = 230.0; // Nominal grid voltage in Volts (RMS)

const float TARIFFPERKWH_INR = 7.50; // Average commercial/residential slab in INR (₹)

// Calibration Constants

// Derived from: (100A / 0.050A) / BurdenResistor(33 Ohm) * (3.3V / 4095 steps)

// Adjust using a known resistive reference load (e.g., 1000W electric heater)

const float CURRENT_CALIBRATION = 60.6;

const float NOISEGATEAMPS = 0.12; // Ignores ambient noise under 28 Watts

// Sampling Parameters

const int NUMBEROFSAMPLES = 1000; // 1000 samples over ~200ms (10 full 50Hz cycles)

const unsigned long TELEMETRYINTERVALMS = 2000; // Update calculations every 2s

// -----------------------------------------------------------------------------

// TELEMETRY STATE VARIABLES

// -----------------------------------------------------------------------------

float currentRMS = 0.0;

float apparentPowerVA = 0.0;

float totalEnergyKWh = 0.0;

float accumulatedCostINR = 0.0;

unsigned long lastTelemetryTimestamp = 0;

unsigned long lastSampleMicros = 0;

WebServer server(80);

// -----------------------------------------------------------------------------

// TRUE RMS DISCRETE SAMPLING FUNCTION

// -----------------------------------------------------------------------------

float calculateTrueRMSCurrent() {

long sumOfSquares = 0;

int sampleCount = 0;

// Calculate dynamic midpoint (DC bias) using moving average

long sampleSum = 0;

for (int i = 0; i < 50; i++) {

sampleSum += analogRead(CTANALOGPIN);

delayMicroseconds(50);

}

int dynamicMidpoint = sampleSum / 50;

// Discrete sampling loop

unsigned long startMillis = millis();

while (sampleCount < NUMBEROFSAMPLES && (millis() - startMillis < 250)) {

int rawValue = analogRead(CTANALOGPIN);

int centeredValue = rawValue - dynamicMidpoint;

sumOfSquares += (long)centeredValue * centeredValue;

sampleCount++;

delayMicroseconds(180); // Sample rate ~5kHz

}

if (sampleCount == 0) return 0.0;

// Mean square

float meanSquare = (float)sumOfSquares / (float)sampleCount;

// Root mean square in ADC raw steps

float rmsADC = sqrt(meanSquare);

// Convert raw ADC RMS to Amperes

float calculatedCurrent = (rmsADC / 4095.0) 3.3 CURRENT_CALIBRATION;

// Apply noise gate to zero out phantom noise

if (calculatedCurrent < NOISEGATEAMPS) {

calculatedCurrent = 0.0;

}

return calculatedCurrent;

}

// -----------------------------------------------------------------------------

// WEB SERVER ENDPOINTS

// -----------------------------------------------------------------------------

void handleRoot() {

String html = "<!DOCTYPE html><html><head><meta charset='UTF-8'><meta name='viewport' content='width=device-width, initial-scale=1.0'>";

html += "<title>Mentora Smart Energy Hub</title>";

html += "<style>body{font-family:sans-serif;background:#0f172a;color:#f8fafc;padding:20px;text-align:center}";

html += ".card{background:#1e293b;border-radius:12px;padding:20px;margin:15px auto;max-width:400px;box-shadow:0 4px 6px -1px rgba(0,0,0,0.5)}";

html += ".val{font-size:2.4rem;font-weight:bold;color:#38bdf8;margin:10px 0}";

html += ".unit{font-size:1.1rem;color:#94a3b8}";

html += ".tariff{color:#34d399;font-weight:bold}";

html += "</style></head><body>";

html += "<h1>⚡ Mentora Smart Energy Monitor</h1>";

html += "<div class='card'><div>Live Current (RMS)</div><div class='val' id='cur'>" + String(currentRMS, 2) + "</div><div class='unit'>Amperes (A)</div></div>";

html += "<div class='card'><div>Real-Time Apparent Power</div><div class='val' id='pwr'>" + String(apparentPowerVA, 1) + "</div><div class='unit'>Watts / VA (at 230V)</div></div>";

html += "<div class='card'><div>Accumulated Consumption</div><div class='val' id='kwh'>" + String(totalEnergyKWh, 4) + "</div><div class='unit'>Kilowatt-Hours (kWh)</div></div>";

html += "<div class='card'><div>Projected Billing Cost</div><div class='val tariff' id='inr'>₹ " + String(accumulatedCostINR, 2) + "</div><div class='unit'>@ ₹" + String(TARIFFPERKWH_INR, 2) + " / kWh</div></div>";

html += "</body></html>";

server.send(200, "text/html", html);

}

void handleAPI() {

String json = "{";

json += "\"current\":" + String(currentRMS, 3) + ",";

json += "\"voltage\":" + String(MAINSVOLTAGERMS, 1) + ",";

json += "\"power\":" + String(apparentPowerVA, 2) + ",";

json += "\"kwh\":" + String(totalEnergyKWh, 5) + ",";

json += "\"cost\":" + String(accumulatedCostINR, 2);

json += "}";

server.send(200, "application/json", json);

}

// -----------------------------------------------------------------------------

// ARDUINO SETUP

// -----------------------------------------------------------------------------

void setup() {

Serial.begin(115200);

delay(1000);

Serial.println("\n========================================================");

Serial.println(" MENTORA ESP32 SMART AC POWER & ENERGY METER NODE ");

Serial.println("========================================================");

// Configure ADC resolution and attenuation

pinMode(CTANALOGPIN, INPUT);

pinMode(STATUSLEDPIN, OUTPUT);

digitalWrite(STATUSLEDPIN, LOW);

analogReadResolution(12); // 12-bit ADC (0 - 4095)

analogSetAttenuation(ADC_11db); // Full 0V - 3.3V range

// Connect to Wi-Fi

Serial.printf("[WIFI] Connecting to SSID: %s\n", WIFI_SSID);

WiFi.mode(WIFI_STA);

WiFi.begin(WIFISSID, WIFIPASSWORD);

int attempts = 0;

while (WiFi.status() != WL_CONNECTED && attempts < 25) {

delay(500);

Serial.print(".");

attempts++;

}

if (WiFi.status() == WL_CONNECTED) {

Serial.println("\n[WIFI] Connected! Assigned IP: " + WiFi.localIP().toString());

digitalWrite(STATUSLEDPIN, HIGH);

} else {

Serial.println("\n[WARNING] Wi-Fi connection timed out. Meter operating in autonomous mode.");

}

// Setup Web Server

server.on("/", handleRoot);

server.on("/api", handleAPI);

server.begin();

Serial.println("[HTTP] Local telemetry dashboard server initialized on port 80.");

lastTelemetryTimestamp = millis();

}

// -----------------------------------------------------------------------------

// ARDUINO MAIN LOOP

// -----------------------------------------------------------------------------

void loop() {

server.handleClient();

unsigned long currentMillis = millis();

// Execute periodic energy calculations

if (currentMillis - lastTelemetryTimestamp >= TELEMETRYINTERVALMS) {

float intervalSeconds = (currentMillis - lastTelemetryTimestamp) / 1000.0;

lastTelemetryTimestamp = currentMillis;

// Calculate True RMS current

currentRMS = calculateTrueRMSCurrent();

apparentPowerVA = currentRMS * MAINSVOLTAGERMS;

// Accumulate energy (kWh = Watts * hours / 1000)

float deltaKWh = (apparentPowerVA * (intervalSeconds / 3600.0)) / 1000.0;

totalEnergyKWh += deltaKWh;

accumulatedCostINR = totalEnergyKWh * TARIFFPERKWH_INR;

// Serial Telemetry

Serial.printf("[METER] Current: %5.2f A | Power: %6.1f W | Total: %7.4f kWh | Cost: Rs %5.2f\n",

currentRMS, apparentPowerVA, totalEnergyKWh, accumulatedCostINR);

// Heartbeat LED toggle

digitalWrite(STATUSLEDPIN, !digitalRead(STATUSLEDPIN));

}

// Handle Wi-Fi automatic recovery

if (WiFi.status() != WL_CONNECTED && (currentMillis % 10000 < 50)) {

Serial.println("[WIFI] Reconnecting...");

WiFi.reconnect();

}

delay(10);

}

``