Hi,

We are using the ADE7880 metering IC for our energy meter project. We need to display the power parameters for each individual phase and System Energy parameters.

For Active power, i am reading the xWATT registers and multiplying the read ADC count with 16 (2^4) and then multiplying with the scaling factor. In this case the error observed is with in our expected accuracy.

For Reactive power, i am configuring the harmonic engine and reading only the FVAR register. The read ADC count is multiplied with 16 (2^4) and then multiplied with the scaling factor. In this case, the error observed is 4 times more. I tried to change the multiplication factor from 16 (2^4) to 4 and then the error observed was with in the limit. Is this an expected behavior and the datasheet needs to be updated or am i missing something?

Almost the same behavior is observed with energy calculations. The error observed with Active Energy computation is within the limit while the error observed with Reactive Energy was too high (almost 200%). I had to divide the read ADC count (xFVARHR) with 4 to bring down the error percentage assuming the above theory to be correct.

Request you to please provide your suggestion in this regard at the earliest.

thanks,

mazahar

Hi mazahar

I do not quite understand why you are multiplying by 2^4. Ideally the value you read out of these registers must be multiplied with the conversion constant already obtained, to get the specific real-world value and that should be enough. Both the active and reactive registers have the same implementation internally and hence there shouldn't be much of a difference between them. Are you comparing the result in and out of the harmonic engine, because in that case, you might have to wait for a particular amount of settling time(750ms) while using harmonic engine before you read the value.

Also while calculating the conversion constant, make an effort to fix as accurate a value as possible so that the variation in that doesn't introduce errors as well. Use ideal voltage and current source to determine the expected power value and then compare it with the waveform register reading to accurately get the conversion constant. Waiting for a longer time period before making a measurement will improve the accuracy.

Regards

hmani

Regards

hmani