Here's a recent reading from my two Suncontrol boards
The first board has a solar panel attached which provides part of the energy to the main circuit. Whenever that happens I expect (and accept) slightly lower efficiency of the board. Battery plus solar provide 690 mW and the output consumes 506 mW, the rest is consumed by the SunControl board.
The second board ("servo") currently has its solar panel disconnected. However the math doesn't add up. The output consumes more wattage (76.5 mW) than the battery provides (74.6mW). I'm at a loss as to what could cause that. I may swap the boards to see if the INA3221 is to blame on the second board, but maybe you have other ideas?
The Suncontrol board in question has three loads - a servo that comes on only every five minutes for a few seconds, and shouldn't really count; the QPM board that controls the various solar panels; and the I2C bus for the entire suite of sensors etc. (I had made sure to disconnect I2C power pin from the Raspberry Pi, and to verify that the SunControl boards do not provide power to the I2C bus unless asked).
The circuit shares a common ground with the main circuit, and there are lots of bits and bobs connecting them together (the QPM, for one). But I can't really see where the extra energy would sneak in from the first circuit into the second circuit, and I can't measure it when the first circuit is off, for obvious reasons.
Maybe it's a question of load magnitude? Maybe the INA3221 works better under heavier load?