Recently during the RTL phase of a quadplane test flight my aircraft was commanded to speed up past its
TRIM_ARSPD_CM of 18 m/s to reach 21.1 m/s. With reference to my previous post, this caused the aircraft to overshoot the landing zone and crash into a tree - this problem has been mitigated with the addition of
Q_TRANS_DECEL in the latest Plane release to dynamically change the transition point based on expected stopping distance.
However, this still doesn't clear up why the aircraft was asked to exceed
TRIM_ARSPD_CM in the first place. On inspection of previous logs, in previous flights with this and other aircraft,
TECS.spdem on RTL was always between
ARSPD_FBW_MAX, leaning towards being closer to
ARSPD_FBW_MAX, even though the aircraft is descending.
Is this expected behavior? What I naturally expect during RTL is that the aircraft keeps to
TRIM_ARSPD_CM or less during descent.
Crash into tree: https://drive.google.com/open?id=10WGdI3Fhge8TQIYt4FOmO5MCYLBvk2kw
(TRIM_ARSPD_CM = 18 m/s, ARSPD_FBW_MAX = 24 m/s, RTL TECS.spdem = 21.21 m/s)
Previous flight before crash: https://drive.google.com/open?id=1rEAGXu6oFHJiVN3PU0MS75k6Z_7QdnrY
(TRIM_ARSPD_CM = 18 m/s, ARSPD_FBW_MAX = 30 m/s, RTL TECS.spdem = 21.68 m/s)
Flight on another day: https://drive.google.com/open?id=0B_bbiwjdkLT2ek9UU21Ma0RDclk (TRIM_ARSPD_CM = 18 m/s, ARSPD_FBW_MAX = 19 m/s, RTL TECS.spdem = 18.6445 m/s)