Tuning term clarification/deep dive

ArduCopter Copter 4.6
1

Dear Dev Team, i have the following questions:

  1. Correlation Thresholds: What correlation coefficient indicates “good enough” tracking for FF estimation when FF=0?

  2. PID Component Ratios: In a well-tuned copter with FF, what percentage of output should each term contribute?

  3. D-Term Noise: What ratio of hover-RMS to maneuver-RMS indicates D-term is mostly noise?

  4. Autotune Targets: What metrics does autotune optimize for internally? (Correlation? Overshoot? Settling time?)

  5. Yaw Tuning: Should yaw FF typically be lower than roll/pitch FF? By how much?

  6. ATC_RATE_FF_ENAB: Can someone from the dev team clarify the exact signal flow when this is enabled vs disabled?

  7. Version Differences: Are there significant PID implementation changes between 4.5.x and 4.6.x that affect analysis?