I have read enough to build a simulated circuit and it seems to do what I need: https://everycircuit.com/circuit/6423598925086720
Did I miss anything?
(purpose below)
Purpose: Building an ESP8266 based controller for a 100W to 200W pump. Controller will include a current transformer to monitor the state of the pump. Pump is plugged into a regular 120V wall socket. ESP relay switches on the socket. It is desirable to have more resolution in the < 5A range. But someone can plugin anything into the socket and draw the maximum from the 15A breaker. This circuit is intended to protect the ESP8266 ADC which can take 1V max. Also that ADC has only 1024 values so mapping the entire 15mA range may not provide very interesting results.
The simulation does not have a current transformer, so I used a "voltage controlled current source". This is just to simulate CT producing its' mA. CT "zmct103c" is 1000:1 and rated for 5A. I tested it to be linear enough until 6 or 7A. Above 8A it curves and looks like will not reach 15mA.
Secondary question: anything wrong with measuring half the wave? I suspect that -0.5V to 0.5V biased to 0-1V is not any better than 0 - 1V with half the cycle missing. The software can take care of reading for max V over some number of samples and mapping the max to what I can measure with a KillAWatt (I only have mA meters :( ).
Thank you.
Edit: removed one of the zener diodes, it seems unnecessary
Added a 1uF capacitor to make sure the ADC never encounters even the tiniest negative voltage. The peak signal got lower a bit, but easy to tweak with the divider.
Another question occurred: is there any downside to this arrangement vs this What is the purpose of the capacitor in this current transformer to arduino interfacing circuit? where the ac signal is given a DC offset to make it positive?
Thanks again
