The Pitch controller of type pi-simple-controller had two constants
defined: Kp=-0.05 and Ti=-0.10.
This made no sense as a pi-simple-controller knows only the two
parameters Kp and Ki, not Ti, and thus the controller was actually
running as a purely proportional controller with Kp=-0.05 and Ki=0.0.
However, when following an ILS glideslope, the missing integration
term lets the aircraft drift below the glideslope.
Keeping the values Kp=-0.05 and Ki=-0.10 (changed from Ti to Ki)
make the AP able to follow an ILS glideslope, but it introduces
serious pitch oscillations at altitude hold at airspeeds over 220kts.
The values Kp=-0.05 and Ki=-0.001 appear to make the whole thing
kind of work for both altitude hold and following a glideslope.
More fine (and probably coarse) tuning needed after the FDM rework.
The Pitch controller of type pi-simple-controller had two constants defined: Kp=-0.05 and Ti=-0.10.
This made no sense as a pi-simple-controller knows only the two parameters Kp and Ki, not Ti, and thus the controller was actually running as a purely proportional controller with Kp=-0.05 and Ki=0.0.
However, when following an ILS glideslope, the missing integration term lets the aircraft drift below the glideslope.
Keeping the values Kp=-0.05 and Ki=-0.10 (changed from Ti to Ki) make the AP able to follow an ILS glideslope, but it introduces serious pitch oscillations at altitude hold at airspeeds over 220kts.
The values Kp=-0.05 and Ki=-0.001 appear to make the whole thing kind of work for both altitude hold and following a glideslope.
More fine (and probably coarse) tuning needed after the FDM rework.