Metering AC power and energy put us to new challenges when compared to DC.

Lets first assume a resistive load where current and voltage are always in phase. If we sample with a frequency lower than double line frequency (eg. 100 Hz), we’ll get aliasing noise. For energy metering this noise doesn’t pose a problem because it is eliminated over time with many measurements. This is true for harmonic, non-distorted (sinusoidal) signals. For instant power measurement, the aliasing noise can reduce our bandwidth (i.e. time resolution) by a magnitude. We’ll need long smoothing filters to remove the aliasing.

In case can be out-of-phase with voltage (reactive load, capacitive or inductive), we need higher sampling freq. MSP430I2040 samples with 8 KHz, which is 4 KHz Nyquist bandwidth. This covers the 66th harmonic for 60 Hz AC. See the datasheet page 12 for metering formula (RMS, active, reactive, apparent, PF, fundamentals, THD)

Chips

• HLW8012 (non-isolated, active power, I rms, U rms)
  • 0.5% RMS accuracy
  • +-43mV diff range for burden shunt resistor
  • measures with voltage divider
  • pwm output
• MSP430I240,
  • eval board: EVM430-I2040S
• ADE7953
  • 1 voltage channel, 2 current channels
  • highly flexible chip with x22 PGA
  • peak detection (TODO programmable alert pin?)
• M90E26
  • ditronix IPEM ESP32 energy monitor board [schematics]
  • similar to ADE7953, cheaper

cs5460a

• cheap china chip with a voltage and current channel. Measures RMS. SDI interface.
• arduino break out board exists
• 25 ppm/°C

PZEM-004T V3

Isolation

PS2501

Current Transformers

ZMCT103C. THT, 1:1000, 10 A, 50ohm. There are arduino modules with LM358.