DCM230 reading no current with 20A

Hi all!
I am building an invasive monitoring system for my PV with two DCM230 for the strings and a SDM630 for the AC. Fixed MC4 connectors for input and outputs to the inverter.
Shunts are 20A for 75mV.
In order to verify no problems in MC4 and evaluating the thermal heating of this configuration I use a 20A current regulating dc supply. I supply the positive first string input and the return is the second string negative input.
The negative first string input is connected to the second string positive input. Each string output normally connected to the inverter inputs are shorted.
For a thermal heating estimation with an infrared camera everything seems OK with a +10K on the MC4 connectors and a +20K on the shunts.
My surprise is that the measured current at the display of each DCM230 is desperately 0.000A. even if I rise the current flowing in the system at 40A!!
With 20A the measure with a multimeter at the DCM shunt input is in the range of the nominal 75mV.
Is there any guy here that use a DCM230 successfully?
I have already seen that there is a kind of SW limitation that display null voltage below a few volts in order I think to not integrate energy when almost no voltage.
Anyone has an idea how to check the current measure at almost no voltage?
Thanks for your help!

Looking at the wiring diagram Eastron Electronic Co., Ltd. there is nothing that tells me, for a test, you must measure the voltage across the “load” (your P.V. panels). You can easily use a separate d.c. bench power supply connected to terminals 1 & 4, leaving the shunt connected to 5 & 6.

You might need to link either terminals 4-5 or terminals 4-6.

Of course, “power” will be V14 × I56, whereas the true power in your short circuit is exactly zero. :smile:

So finally everything work well. The DCM230 has only a non documented feature that force voltage and current to zero when there is small voltage measured, in order to not integrate energy if you have an offset in the measurements. 12 volts is enough to bypass this and measure the current.
And if you use minimalmodbus to read the DCM you need to use byteorder=0 to read the values correctly.