Oops, a three-phase meter with only two phases connected? Could that unbalanced load cause trouble too?
As long as all the voltage coils are energised the meter will be very accurate.
If one voltage coil is de-energised the voltage flux from the missing phase, its braking effect at light loads, may make the meter go a little faster in the order of around 1 to 1½ %.
This only applies to genuine 3Ø4Wire meters.
e.g. 3x230/400 Volts or 3x120/208 Volts with a Neutral.
Some 3Ø meters have the BØ PT coil wired with 3 wires in (not 100% sure) Scott type connection between the PT coils, so that these can also be used on 3Ø without Neutral as well e.g. 3Ø3Wire or 3Ø4Wire ccts.
In my days in meter calibration I have proved it on the test bench with a 3 phase meter running at 2Ø or even 1Ø. As long the other PT coils are live, the meter will be very accurate to within its class.
In New Zealand when with tariffchanges in the early 1990's, some older pumpsheds were converted from 3Ø to 1Ø it was not always warranted to fit a new 1Ø meter. ( Usage in some of these pumps is sometimes less than 100 kWh / Year). We left the 3 Ø meter on site and put PT jumpers in to the other 2 phases to keep the coils energized and the meter accurate. A lot of these meters were the Sangamo S 301.2 type, 66 2/3 rev kWh,
3 x 10 / 60 Amps, clockdial type