ADE9078 response/latency

I want to verify my understanding of the datasheet since this will be a critical point for us.

If I am reading the datasheet correctly I could get a 4kHz update of real power if I disable the low pass filter. To be clear it is only real power I need that update for, reactive and apparent power I don't need that quickly and could even do without if necessary.

In our application we actually only need a 10mS update rate but would like 5mS so 4kHz works nicely.
  • 0
    •  Analog Employees 
    on Oct 7, 2021 2:36 PM

    The real power updates at 4Ksps and is filtered. Not sure why you would disable low pass filter but you can. Why do you think you need to disable the lowpass? 

    Why is reading the power with the lpf on every 10ms not good enough? There will be a lot of ripple without the lpf. 


  • Perhaps I'm misreading but the datasheet (table 15) says that the settling time approaches a second if the low pass filter is on. That's clearly unacceptable for a 10mS response.

    Hmm, re-reading that, it looks like even with the filter off the response will be on the order of 10's of mS. That will be an issue if that's the case.


  • 0
    •  Analog Employees 
    on Oct 7, 2021 3:27 PM in reply to Robert000000000000000000000

    There is also a high pass filter and this has some delay. 

    For fastest response reading the 4Ksps samples after the sinc and decimation is best but there will be no high pass filter.  Dc offsets will start adding errors. 


  • What would the response be in that case?

    I'm not interested in energy for this case only power so offsets won't add from sample to sample. I can likely manage offsets if they are not too large. Can you point me to where they are defined in the datasheet?

    I'm beginning to think that maybe I'd be further ahead using direct A/D to get the real power but I don't want to re-invent an existing wheel. Or maybe one of the other monitor ICs is a better fit?

  • 0
    •  Analog Employees 
    on Oct 7, 2021 7:09 PM in reply to Robert000000000000000000000

    DC offsets do impact power. Any dc offset in each channel with multiply each other.

    Maybe you can tell me more about your application it might help me understand your limitations. 

    The output of the sinc filter is 16Khz after decimation is it 4Ksps the bandwidth is 1.6Khz so it should settle fast enough if you do the power calculation in the micro 

    Not using the highpass filter will cause power error over temp as the dc offsets are temperature dependant. using the highpass is recommended for billing and to meet datasheet specs for watt var va. 

    All the ADE metering parts will have similar issues with dc offset without highpass and settling if the highpass filter is used. 

    To get the fasted response you might need and ADC. Why is 10ms so important? 1/2 cycle at 50Hz ? Is this response time for a protective relay or something?