BQ40Z60: Full Charge Capacity in mWH FullChgCap is completely incorrect (almost by a factor of four) after running IT learning cycle

I am working with Steven on this project and we are both using identical (but separate) hardware.

1: I think that 1 actually is a reasonable value for the CC gain, because our design uses a 1 mOhm sense resistor.  After a current calibration, I ended up with a value of 1.004.

2: We both were able to complete learning cycles, ending with Update Status=6.

3: The capacities are only off when in mWh capacity reporting mode.  When I set CapM=0, the mAH full charge capacity and remaining capacity values are plausible, and seem accurate through discharge and charge.  The inaccurate values only occur when CapM=1.

4:  BQstudio seems to be reading the raw hex values of the two mWH capacity registers incorrectly, at least according to the "raw" values that show up when you click on the value in BQstudio. See below for Full Charge Capacity:

The raw value 0x49CB is actually 18891, not 57838

And that happens to be the full capacity in cWH, rather than mWH!

You can see here that the raw and decimal values for one of the other Full Charge Energy registers is also 0x49CB, but in the case of “Flt Full Chg E”, 0x49CB gets accurately converted to 18891 cWH:

5: I disconnected the EV2300 and used a Raspberry Pi to read the full and remaining capacities using SMBus.  They appear to be the decimal versions of the hex values that show up as the “Raw” values in BQstudio. For example, the Full Capacity register (0x10) read as 19014 over SMBus on the Raspberry Pi.  In Bqstudio, it reads as 59068 in decimal, 0x4A46 in raw hex. 0x4A46 is actually 19014 ! In other words, it seems that this problem is a BQstudio bug and might not not affect the readings that the BQ sends to the host, as long as the host knows to expect cWH instead of mWH.  Do you agree with this assessment?