Post Go back to editing

half current after some reading

Category: Datasheet/Specs
Product Number: ADE7953

We are calibrating a board using the ADE7953 as a power meter.

Initially, with the default values, we observed a low error in the current measurements. For example, with a reference value of 25 mA, the reading was 0.019 A.

After calibrating the current gain register, we successfully obtained the correct reading of 25 mA.

We then proceeded to calibrate the voltage, power factor, and power gain registers, which worked as expected.

However, when adjusting the power offset register, we encountered an issue: the current readings suddenly dropped to half their actual value.

Even after reinitializing the ADE7953 to its default register settings, the readings remained at exactly half the expected value.

Notably, if we adjust the current gain register, we are able to obtain the correct current value.

This phenomenon has occurred with four different boards in similar ways. We have not been able to determine the cause or find a procedure to restore the device to its "factory settings."

We would appreciate any guidance on resolving this issue.

Best regards,

  • What is the full-scale Current of the meter?

    Gain calibration and phase calibration are usually done around 10% to 20% of I max. Adjust the gain so you get the expected RMS at the test voltage and current. Pgain should not be needed. 

    After you do gain and phase set your current to 25 ma and use watt offset register to adj the low current. 

    Note in the calibration application note the equation calculating CFDEN there is an error. This error requires you to use pgain. With the proper value used in the calculation you should not need pgain. 

    Using CF_FULLSCALE = 207.488Khz instead of 206.9Khz will calculate CF to the expected value without needed pgain register.

    AN-1118 (Rev. B) pg 16 download the doc so you get page numbers. Here is says 210Khz. Use 207.488Khz

    in the datasheet cf fullscale is  206.9Khz  pg33 ADE7953 (Rev. C)

    If you want to verify this number set CFDEN to 1

    Write 0x0 to Igain and Vgain and 0x7FFFFF to Awattos, WATT register will be 0x7FFFFF / 2^9 = 16384lsb this value will be constant. If you measure cf freq 
    cf with cfden=1 should be close 699.137Hz, depends oncrystal freq. (pmax / 16384)* 699.137Hz = 207488hz fullscale CF

    Assumes 3.579545 Mhz if using a different crystal freq this number will be different.

    Dave