D setpoint effect - Plasmatree/PID-Analyzer GitHub Wiki

D-setpoint weight adjusts your D-term according to the change of rate of rc-input (rate='velocity'; change of rate='acceleration'). Higher faster change of RC-Command = temporarily lower D The result is an intentional overreaction to your input. The quick overswing can compensate for the natural lag of the system by moving faster than requested for a short time. The quad basically "catches up".

BUT the reaction will not only be faster but also rougher. Experiment and find the sweet spot for your style of flight!

The influence on the system response on pitch is shown below (mind that the response now depends on the style of input you give, even with the same setting the result can vary. And it will only work in acro mode): The loop input/gyro plots show exemplary details of the log where you can directly see how the difference in response relates to in flight movements.

• Without setpoint, there is significant lag but otherwise the gyro looks like a copy of the input
• With the recommended setpoint the lag diminishes, the quad catches up without noticeable overshoot during normal maneuvers
• With highest possible setpoint, the you get over pronunciation and overshoot at every small edge

My recommendation is to first tune your quad without d-setpoint-weight (0). If you got a nice edge without overshoot, add the d-setpoint-weight to make it crisp! Use the zoom or navigation cross to zoom into the plots and see if and where it overshoots.

Keep in mind that the overshoot from too low D is different to overshoot from D-setpoint weight! The latter is related to rc-input and not to loop-input (it doesn't act on stuff like wind or propwash).

Mathematically I guess the ideal setting would be if the area below 1 (during the rise) is equal to the area above 1 (during overshoot). It should indicate that it catches up perfectly. But that's something I'll have to crosscheck.