Right/Left balance very different

If GARMIN VECTOR 3 (not 3S) was calibrated, it's ALWAYS MORE ACCURATE than any trainer because Vector 3 collects data directly from each pedal (Neo collects power data only from cassete moving and calculate approximate left/right balance from torque moment, so this method is always less accurate). Also spider based powermeters (Quarq Dzero, for example) are more accurate than trainer sensors because it closer to legs and there are minimum transmission losses.
A capacitive left and right detection sensor is useful only for cadence measuring. It's the only thing that provides accurate data on Neo 2/2T trainers (Neo T2800 has virtual cadence).

the Vector has true balance (left balance = only your left leg ), the Neo2T has pseudo balance only (left balance = left push phase + right pull , right balance = right push + left pull, like spider based powermeters: Quarq and Power2Max are working the same, so dont panic, both values are good just the calculation method is different )

I also see vastly different balance between the Neo 2T (0.0.31) and Garmin Vector 3 pedals. The results are so different that I wouldn't call them good or useful yet. It's clear that the Neo never will be as accurate but I hope they can improve this in future firmware updates.

See https://analyze.dcrainmaker.com/#/public/59ab9711-50ab-4c9b-7087-2e8f3aa24c3e (4157235386 are the Vector 3 pedals, 4157235489 is the Neo) for an example with 30s smoothing. This comparison also shows that firmware 0.0.31 over-reports power at low intensity and large gears in ERG mode. Still an improvement over 0.0.27, though. 

how do You know that Neo overreports the power and not the V3s have issue and underreport it as time goes by? what kind of gear did You use over the test, a fix one or variable gearing?  Or are these steps a gear down and as You geared down at rear the power difference started to increase?

Because I did a torque test (https://support.garmin.com/en-US/?faq=AVZeQisZvi3oBYZUUZdPV6) on each pedal

huh, do You have 2 user names? (jk and dahs1e?) :) 

How heavy weight did You use? 

No I'm dash1e and stop! I've used 5, 10 and 15 kg for test

okie ( i made these calibration with 20kg, heavier is better! ) , and what was the result of this static calib? didnt need modify the scale factor?

and Do You have power difference also as jk has between 2T and V3? 

back to the balance question, use the V3 balance only, and ignore the Neo2's one, it is only estimated and never will be accurate 

 Ok but... right/left balance I have to ignore it, ant+ for speed/cadence sensor I have to disable it. Neo 2T it's a 1200 euro device and I have to accept to ignore some of its features? I can't believe it, we've bought the best and we want the best!