How a Current Transformer Donut Measures AC Power Without Breaking the Circuit

The One-Sentence Definition and Core Principle

A current transformer donut is a toroidal or split-core magnetic sensor that clamps around a single AC conductor to induce a proportionally smaller, safely measurable secondary current without breaking the circuit. In a real installation, this component changes everything: it steps down dangerous, mains-level currents (like 100A at 240V) into milliamp or millivolt signals that are completely safe for 3.3V microcontroller ADCs or standard panel meters, all while maintaining strict galvanic isolation between the high-voltage mains and your low-voltage logic.

Think of it like a traffic camera mounted over a highway overpass. It doesn't physically touch the cars (the primary current), it doesn't slow them down (zero insertion impedance), but it counts exactly how many pass by per second based on their magnetic "shadow." The primary wire acts as a single-turn primary winding, while the donut's ferrite or iron core concentrates the alternating magnetic field into a secondary winding wrapped thousands of times around the core.

The Math: Turns Ratios and Burden Resistors

To use a current transformer (CT) donut with a microcontroller, you must convert the induced secondary current into a measurable voltage. You do this using a burden resistor. Let's run a worked numeric example using the most common hobbyist CT, the YHDC SCT-013-000.

1. Calculate Secondary Current: If your primary load is drawing 100A, the secondary current is 100A / 2000 = 0.05A (50mA).
2. Determine Target Voltage: An ESP32 ADC reads 0V to 3.3V. Because AC swings positive and negative, we must bias the signal to the midpoint (1.65V). This leaves a maximum peak swing of 1.65V in either direction.
3. Convert Peak to RMS: The RMS voltage of our maximum swing is 1.65V / √2 = 1.166V RMS.
4. Calculate Burden Resistor: Using Ohm's Law (R = V / I), we get 1.166V / 0.05A = 23.32Ω.

In practice, you select the next lowest standard resistor value to avoid clipping the ADC at maximum load. A 22Ω 1/4W resistor is the perfect choice here, yielding a max RMS voltage of 1.10V and keeping your ADC safely within its linear range. For a deeper look at the underlying magnetics and phase shift errors introduced by the burden resistor, the OpenEnergyMonitor CT theory guide provides excellent bench-level validation data.

Where You Meet This in Practice

You will find current transformer donut arrays inside commercial home energy monitors like the Emporia Vue or Sense, where a dozen split-core CTs clamp onto individual branch circuits in a subpanel. In the DIY and maker space, they are the backbone of solar inverter monitoring, smart breaker projects, and automated welder logging.

However, choosing the right sensor topology is critical. Here is how the CT donut stacks up against alternatives you might find on a parts list:

Bench Scenario: Wiring an SCT-013 to an ESP32 (And the Mistake That Fries Your ADC)

Theory is clean; the workbench is messy. Here is a real-world scenario walkthrough of measuring a 15A space heater with an ESP32 DevKit v1, including a common failure mode that destroys microcontrollers.

The Setup

We clamp the SCT-013-000 over the hot wire of a 120V space heater. The secondary wires connect to a breadboard. We place a 22Ω burden resistor across the CT leads. To bias the signal for the ESP32's 0-3.3V ADC, we build a voltage divider using two 10kΩ resistors from 3.3V to GND, creating a 1.65V virtual ground, and buffer it with a 10µF capacitor. The CT signal is AC-coupled or referenced to this 1.65V midpoint before feeding into GPIO 34.

The Numbers

The space heater draws 15A RMS. The CT steps this down to 7.5mA RMS. Across the 22Ω burden resistor, this generates 0.165V RMS. The peak voltage is 0.233V. Referenced to our 1.65V bias, the ESP32 ADC sees a clean sine wave swinging from 1.417V to 1.883V.

The Outcome

Using the EmonLib library, the ESP32 samples the wave at 2kHz, calculates the RMS of the ADC readings, and accurately logs 15.02A to an MQTT broker. The galvanic isolation means a fault in the heater wiring cannot reach the microcontroller.

What Went Wrong (The War Story)

A reader previously attempted this exact build but grabbed an SCT-013-000 from their parts bin and forgot the external burden resistor and DC bias circuit. They plugged the raw CT leads directly into the ESP32 GPIO, assuming it would "just read the voltage" like the SCT-013-030 model.

Common Confusions and Pitfalls

When sourcing or specifying these sensors, builders frequently confuse the current transformer donut with other current measurement topologies. Understanding the differences prevents catastrophic bench mistakes.

• Confusion with Hall-Effect Sensors (ACS712): Hall-effect ICs measure the magnetic field directly using a semiconductor element, meaning they can measure both AC and DC current. A CT donut relies strictly on Faraday's Law of Induction, which requires a changing magnetic field (dΦ/dt). If you pass 12V DC from a battery bank through a CT donut, the secondary output will be exactly zero. For DC battery monitoring, you must use a Hall-effect sensor or a shunt resistor.
• Confusion with Rogowski Coils: Rogowski coils look like flexible CT donuts, but they are air-cored and measure the rate of change of current (di/dt). They require an active analog integrator circuit to reconstruct the actual current waveform. Standard split-core CT donuts have ferrite cores, measure absolute current directly, and only require a passive burden resistor.
• The Unburdened Solid Core Danger: While split-core "donuts" are designed to be opened and closed, solid toroidal CTs (often found in industrial panel meters) must never be opened or left unburdened while primary current is flowing. As noted in All About Circuits' guide on CT safety, an open-circuited CT under load can generate thousands of volts across the secondary terminals, posing a lethal shock hazard and instantly destroying the core's insulation.

Frequently Asked Questions