### Accuracy estimation

Dear all,

I am new on this subject and I would like to ask you the correctness of an attempt made in order to improve the accuracy of the current measurement.

I am using the SCT-013-000 clamp connected with a Micaz+MDA300 sensor board. It is an hardware similar to the Arduino, and running with the TinyOS. I run some test and I have a doubt about the nature of calculation has to be made in order to obtain the power value.

Specifically, the question is about the reference voltage. The MDA300 sensor board has an output of stabilized 2.5 Volts which I use as input  for the resistor divider.  The reference voltage on the divider is never as it should be, then the 0 of the sinusoidal curve I obtain is never matching exactly with the theoretical value of 2,5/2 = 1,25 Volts.

In order to get the value of the power first I virtually move to zero the curve and then calculate the Vrms which I multiply for the number of  secondary turns in the CT (2000)and  with the predictable Vrms of the supplier (230) and then divided all by the burden resistor value.

The question is how to fix the zero.

Is it correct to use the Average value of the taken measurement, Instead of the theoretical 1,25 volts value?

Thank you and regards.

Franco Di Persio

### Re: Accuracy estimation

Franco,

The average value should give you the "average" value of the wave. If the wave is truly symmetrical, and you are averaging over an exact integer number of cycles, or you are averaging over so many cycles that the error from including part of a cycle does not matter, then that will be the number you want.

In emonLib, we use a software high pass filter to remove the d.c. component and the output of the filter is then centred about zero.

Neither method will give the correct answer if the wave is not symmetrical, and I do not think there is a method that will always work correctly in that case.