Knowledge Base

SPice 10204: Accuracy and calibration considerations

The output of the current transformer on the SPice 10204 is processed by onboard circuitry to produce a DC voltage that is measured by the controller’s analog to digital converter. The processing consists of:

  • An AC pre-amplifier that amplifies the small signal from the CT to a useable level.
  • A precision rectifier that uses an operational amplifier and diodes to produce a fullwave rectified representation of the ac current waveform.
  • A dc averaging circuit (filter) that converts the fullwave signal into a smooth DC.
  • A calibration potentiometer (quite different to the user accessible general purpose potentiometers).

During factory calibration we apply 4.902A rms at 50Hz to the CT and adjust the calibration potentiometer for 250 counts. We do it at just under full scale to guard against over-ranging the analog input on the controller.

There are several sources of potential inaccuracy in this chain:

  1. The accuracy of our 4.902A test current.
  2. Non-linearity and offset errors in the precision rectifier at currents significantly less than full scale.
  3. The calibration accuracy of the host controller board.

Overall, the above factors may add up to about ±1% to 2% (of full scale) with currents over 5% of full scale. We don’t and won’t quantify these, because they pale into significance in relation to the one factor we, or you, cannot control namely …

Effect of non-sinusoidal current

The SPice 10204 uses a simple averaging measurement. It is calibrated for a sinewave. For any waveform other than sinusoidal it will be inaccurate. Non-sinusoidal waveforms can be due to something as simple as harmonic distortion in the mains voltage caused by non-linear loads such as fluorescent lighting. To make matters worse, many mains loads draw very non-sinusoidal currents even if the voltage is a perfect sinewave (fluorescent lighting being a classical example).

This, by the way, is no different to a typical multimeter. Some up-market multimeters boast true rms AC measurement, but most use pretty much the same kind of average reading circuitry as the SPice 10204.

The result of all this is that any program you write needs to either be adjusted for whatever non-sinusoidal waveform you encounter, or should use a strategy that establishes a practical norm and then works from there. Any such strategy should include a margin of error for changes in the mains distortion.

Effect of frequency

The SPice10204 is designed and suitable only for 50Hz and 60Hz currents. Onboard lowpass filtering will impair accuracy at higher frequencies.