When using a 12-bit or 16-bit AD converter (like the TI ADS1115 or an ESP32 internal SAR ADC) paired with a current transformer to convert a measured 20A AC current reading into real power, the direct converted answer is 2400W on a standard 120V single-phase system (assuming a 1.0 power factor). However, that exact same 20A ADC reading shifts to 4600W on a 230V single-phase system, and 7205W on a 208V 3-phase system. The foundational formula used with these values substituted is: P = V × I × PF (e.g., 120V × 20A × 1.0 = 2400W). For 3-phase, the formula expands to P = √3 × V_LL × I × PF (1.732 × 208V × 20A × 1.0 = 7205W).
The Core Assumption: What Fixes the ADC Output?
An AD converter does not measure Watts; it measures instantaneous voltage. When you clamp an SCT-013 current transformer around a mains wire, the CT outputs a tiny AC voltage (typically 1V RMS at 30A) proportional to the current. Your AD converter digitizes this analog voltage into raw counts (e.g., 0 to 4095 for a 12-bit ADC).
To convert those raw counts into a meaningful power reading, your firmware must lock in three assumptions:
- Voltage (V): The nominal grid voltage (120V, 230V, etc.).
- Phase Configuration: Single-phase vs. 3-phase (which introduces the √3 multiplier).
- Power Factor (PF): The ratio of real power to apparent power, dictated by the load's inductance or capacitance.
If you hardcode 120V and a 1.0 PF into your microcontroller's math, your AD converter will output accurate Watt readings only for purely resistive loads (like space heaters or incandescent bulbs) on a North American split-phase grid. Change the grid, change the load, or change the phase, and your hardcoded conversion factor becomes a liability.
Neighboring Values: ADC Count to Watts Table (±20% Range)
The table below maps a ±20% current range (16A to 24A) around our 20A baseline. This assumes a 1.0 PF and shows how the physical power shifts dramatically across different global grid standards, even though the AD converter's raw current reading remains identical.
| Measured Current (A) | 120V 1-Phase (W) | 230V 1-Phase (W) | 208V 3-Phase (W) |
|---|---|---|---|
| 16A (-20%) | 1,920W | 3,680W | 5,764W |
| 18A (-10%) | 2,160W | 4,140W | 6,485W |
| 20A (Baseline) | 2,400W | 4,600W | 7,205W |
| 22A (+10%) | 2,640W | 5,060W | 7,926W |
| 24A (+20%) | 2,880W | 5,520W | 8,646W |
How the Answer Shifts: 120V vs 230V vs 3-Phase Systems
When designing a PCB or writing firmware for an energy monitor, you must scale your AD converter's LSB (Least Significant Bit) weight based on the target grid.
120V Single-Phase (North America / Japan)
At 120V, a 30A CT sensor represents a maximum of 3600W. If you are using a 3.3V reference on a 12-bit AD converter, your resolution is roughly 0.8mV per count. Because the voltage is low, the current required to deliver high power is high, meaning your CT sensor will output a robust, easily readable analog signal with a high signal-to-noise ratio.
230V Single-Phase (Europe / UK / AU)
At 230V, that same 30A CT sensor now represents 6900W. The AD converter reads the exact same current waveform, but the physical power has nearly doubled. If your firmware is hardcoded for 120V, you will under-report the power consumption of a European HVAC system by roughly 48%. You must update the voltage multiplier in your RMS calculation loop.
208V / 400V 3-Phase (Industrial / Commercial)
Three-phase systems require measuring the line-to-line voltage (V_LL) and applying the √3 (1.732) constant. Furthermore, a standard single-channel AD converter cannot capture total 3-phase power by measuring just one leg. You either need a multi-channel AD converter (like the 4-channel MCP3204) to sample all three phases simultaneously, or you must assume a perfectly balanced load and multiply the single-leg reading by 3. According to All About Circuits, assuming balanced loads on aging industrial motors often leads to 10-15% measurement errors due to phase asymmetry.
When the Conversion is Meaningless: The Power Factor Trap
The conversion from AD converter raw counts to Watts becomes entirely meaningless when the Power Factor (PF) is unknown and the load is inductive.
If you use a single-channel AD converter to measure the current of a well pump or a refrigerator compressor, and you multiply that current by 230V, you are calculating Apparent Power (VA), not Real Power (W). Inductive motors introduce a phase shift between the voltage and current waveforms. If the motor has a PF of 0.65, your simple V × I math will overstate the actual real-world power consumption (and the resulting heat/energy cost) by 35%.
FAQ: AD Converters in AC Power Measurement
How many bits does an AD converter need for 230V AC measurement?
For basic hobbyist energy monitoring (like an ESP32 smart plug), a 12-bit ADC is sufficient, giving you 4096 discrete steps. However, for 230V systems where you need to accurately resolve low standby currents (e.g., a 5W LED bulb drawing ~20mA) alongside high peak currents (e.g., a 3000W oven), you need a wider dynamic range. In professional 230V metrology, 16-bit sigma-delta AD converters (like the TI ADS1115) or 24-bit dedicated energy metering ICs are standard to prevent quantization noise from swallowing low-current readings.
Why does my AD converter read zero watts on a 3-phase motor?
If your AD converter is sampling at a rate lower than the Nyquist frequency of your AC grid, or if your sampling isn't synchronized to the AC zero-crossing, your RMS calculation will collapse. For a 60Hz grid, you need a minimum sampling rate of 1 kSPS (kilosamples per second) to accurately reconstruct the waveform and calculate the true RMS value. If you are using a slow, multiplexed ADC to read three phases sequentially, the phase delay between channel reads will corrupt the 3-phase power calculation.
Can a standard AD converter measure power factor directly?
No. An AD converter is strictly a voltage-digitizing component; it has no concept of time, phase, or power. It simply outputs a binary number representing the analog voltage at the exact microsecond it was polled. Power factor is a mathematical relationship (the cosine of the phase angle between voltage and current). To derive PF, your microcontroller must use the AD converter to capture both waveforms, calculate the time delay (in microseconds) between their respective zero-crossings, and convert that time delta into an angle based on the grid frequency (50Hz or 60Hz). For a deeper look at ADC timing constraints, refer to the TI ADS1115 Datasheet regarding conversion times and I2C bus latency.
