Hi!

I have a problem with ADE7880 and the accuracy of the active power.

I do the calibration as mentioned in AN-1171

My calibration steps as follows:

1. gain matching

2. Phase calibration

3. Watt hours per LSB

4. xPGAIN

5. RMS current and voltage per LSB

6. Watt and VA per LSB

I calibrate with nominal value for U = 110V, angle 0° and I = 1A, angle 300° and the power factor of 0,5

The measurements for calibration values bring good results (U = 110V, 0°;I = 1A,300°;PF = 0,5):

109.997 U L1-N

1.000 I L1

55.063 P L1

110.076 S L1

0.505 PF L1

But for example (active power with an error of 1.8%):

U = 110V, 0°;I = 1A,240°;PF = 0,5:

109.991 U L1-N

1.000 I L1

-53.997 P L1

110.082 S L1

0.495 L1

another worse example concerning the power factor!

U = 110V, 0°;I = 1A,0°;PF = 1:

109.991 U L1-N

1.000 I L1

109.028 P L1

110.063 S L1

-0.999 PF L1

Has anybody an idea to get more accuracy?

Regards

Heimo

Hi Heimo

Based on the data that you have provided, I tried calculating the error observed. The errors observed is within the expected range and far lower than your claim. Please find below:

Note: I have considered all values in the table as absolute (without sign), for simplicity.

I can see that you have calculated ~1.8% error by assuming that -53.997 must have been close to - 55.063. But the power factor isn't exactly +0.5 and -0.5 in each of these cases. If it were, then you could estimate errors that way. In order to determine if the chip is accurately calibrated, I would suggest estimating the input signal conditions accurately and comparing the measured result with the expected value.

One other point that I wish to add is that, if possible, kindly perform phase calibration first, followed by gain calibration and then finally perform offset calibration. This would yield more accurate results.

Regards

hmani