the value '1000' is not really magic, and it was resulting in sending
a single 1ms pulse on all IO channels for a brief instant on
startup. With some servos this led to the control surfaces being at
extremes (and straining) on startup. It may also contribute to making
ESC calibration harder on multi-rotors

In the worst case it could cause a IC motor with reverse throttle to
go full throttle on startup

A logic analyser shows the problem very clearly. Changing the
default_value fields to 0 from 1000 fixes the issue and no pulses are
generated until an explicit value suitable for the airframe is
provided via one of the many methods PWM output values can be
generated

Conflicts:
	src/drivers/boards/px4fmu-v4/px4fmu_pwm_servo.c
	src/lib/ecl

- after jumped takeoff set previous vel setpoint such
that thrust setpoint is continuous

Revert "for multicopter landings make sure that the copter moves"

This reverts commit ad1058d3742bbfa9cbd16648aa2925fa1e618a55.

multirotors shouldn't use tailsitter recovery code.

- Changed from 8 spaces indent back to 4 which I find appropriate for
  Python.
- Fixed linelength to 80 chars.