Re-assess barometer weight and logic for altitude fusion

[digitalentity]

When accelerometer weight is reduced by #4681 or manually, barometer influence is not enough to maintain altitude consistently and reliably.

TODO:

• Re-evaluate barometer weight and find a better default
• Re-evaluate barometer fusion into altitude (possibly extract climb rate from the raw data and correct velocity from that data separately).

[shota3527]

@digitalentity
Is #4681 a good fix for the problem?
It is not reducing the accelerometer weight but set the measured acceleration to 0. Which means velocity/position estimation by accelerometer is also 0. Then value from accelerometer and baro is fighting with earth other.