Bug in def shuntvoltage

[jean-claudeF]

I discovered a bug in the library:

@property
    def shunt_voltage(self):
        """The shunt voltage (between V+ and V-) in Volts (so +-.327V)"""    # no, +-80mV!
        value = _to_signed(self._read_register(_REG_SHUNTVOLTAGE))
        # The least signficant bit is 10uV which is 0.00001 volts
        # NO!! It is 2.5uV
        # return value * 0.00001
        return value * 0.0000025

Chris defined the shunt voltage LSB as 10uV, but the datasheet says 2.5uV (page 24)
Unfortunately my results are still wrong after this correction (by a factor 5 for current and power)

[lioff]

For me the ina226 readings are roughly 15% too high compared with a multimeter (at 200mA / 5V). Did you find any other problems or do you still get wrong results?