Hi Joe.
I set up a quick experiment today to check the resolution of the ADE7763 ZX output. I used a function generator to input a 60Hz signal onto the ADE7753 EVB. I triggered on the ZX falling edge and set up an oscilloscope with infinite persistence around the next falling edge. Within a few seconds, I could see 60us of variation around the falling edge after the trigger. Seeing 60us variation on the falling ZX edge with a 1/60Hz = 16.667ms period implies that if you were looking reading to reading, you could measure 1/(16.667ms+60us) = 59.785Hz ... so there is ~0.2Hz jitter on this signal (+/-0.1Hz).
I could see the line period changing nicely from 59.8 to 60.3Hz with mean period values shown below, measured on the scope:
Line f Mean Period (ms)
59.8 16.36
59.9 16.34
60 16.31
60.1 16.28
60.2 16.26
60.3 16.23
My input source could only provide xx.xHz resolution, not xx.xxxHz so I could not check if the output changed with respect to 1mHz changes in the input.
I think the 60us of variation makes sense because there is a CLKIN/4/32 ~=28kHz clock used at some point in this circuit and 1/28kHz = 35us.
The rising edge of ZX had similar behavior.
I tried averaging the ZX output point on the scope over 8 readings and the variation on the scope seemed to go in half, for ~ +/-50mHz instead of +/-100mHz.
For this reason I don't think 1mHz resolution is possible with ADE7763 ZX circuit.
One other thing to consider is that ADE7763 ZX circuit does not include a high pass filter. If the temperature changes, the modulator dc offset will change which can influence this circuit. Most likely this would be in terms of the duty cycle. The lack of HPF also limits the range of input voltages which can be used with this ZX circuit.
Best Regards
Meghan