Accurate current correction/calibration for the PZEM-004T AC measurement module

A follow-up to the previous video on the Peacefair PZEM-004T AC measurement module.
In that one, my attempt to make a current calibration adjustment failed.
But here I find a way to accurately calibrate the current measurement of the module in software.
My previous video about the adapter PCB:-
• Adapter PCB for Peacef...
Video by TheHWcave that inspired the above:-
• Modifying the Peacefai...

[@PhongNguyen-rs5jf]

Thankyouuu

[@t1d100]

Brilliant

[@BjornV78]

Hi, did you ever tried to use 2 diodes back to back in series with the burden resistor ?
That way, below the diode voltage drop, the burden resistor is not in use, only when the burdenvoltage exceed the diode voltage drop, the burden resistor is in use.

[@bertoid]

No, I haven't tried that. I can only assume that it would really mess things up, maybe even damage the controller chip.
Remember that the burden resistor is only half an Ohm, so the voltage across it is tiny - only 6mV with 2.5A current flowing. The diodes would add 600mv to that...

[@laverdadapesta100pre]

Hello, thanks for that great video, I would just like to ask you if it is possible to access the data of the function you measured.

[@bertoid]

Thanks!
I'm not 100% sure what you mean, but assuming you want the measurements I made, I've uploaded the test runs data to:-
uploadnow.io/f/fbbC10b
It's a RAR archive with 8 test run files in CSV format, each with done with a different load.
So the max currents (at 240V) in each file vary from 265mA to 6.7A.
The format of the data is:-
HPV - Voltage measured by the HP3456A
PZV - Voltage measured by the PZEM-004T
HPI - Current calculated by the HP3456A from the voltage across the shunt resistor
PZI - Current measured by the PZEM-004T with 5 turns through the current transformer (so 5 times the HPCurrent)
BUT! I'm not happy with the data at low currents.
I should have used a much higher shunt resistance so that the voltage developed across it at low currents is above the noise floor.
The data files could all be joined together, sorted (and smoothed) to get the full range with many more data points.
Hope that helps...