Thank you for the details.
Well, actually this behavior can't be really solved by any PID parameter. What you can observe is caused very likely by adaptive geometry correction. This is feature that is trying actually to resolve pitch-up phenomenon when flying fast and changing abrupt collective pitch changes so that the model stays level.
And it is applied to both aileron and elevator, but is calculated separately in respect of what must be compensated.
When flying fast some helicopters require this correction in order to fly straight. But in some cases it could work in the wrong way.
These are:
a) swashplate center is off either because it was not leveled with cyclic leveler and/or when servo arms are at their mechanic center, there is non-zero angle between main blades. This issue (and other) can be eliminated by this guide: http://manual.spirit-system.com/index.p ... tive_Pitch
b) CoG is off, so that helicopter is pushed to some side. Very often it is elevator axis. The more CoG is off, the more the compensation can be apparent. But when flying fast, this imprecision is causing troubles.
c) Servo Travel Correction is used incorrectly
d) Geometry 6° is not configured well (very likely too high)
To check that it is really the case, you can just decrease your Geometry 6° value by maybe 20 - 30 points. Some difference should be visible, but it will have impact also on the flight characteristics. Possible it could eliminate the problem you can see.
The reason why you might not observe the issue with other units is, because there are only few FBL controllers that support adaptive geometry correction. With this there are basically zero pilots with pitch up issues in FFF, while with other controllers you have to either set such compensation statically (if supported) or you can just tune PID to make an issue invisible which is degrading overall performance. When you have to mask an issue, it mean some compromise must be made. But this is not way we want to go.
It is just interesting that a difference is visible between updates. Do you remember which version introduced any improvement? There was no change in any area like this for very long time. Big difference can be visible just be doing another configuration that has just different values. So for example closer it is to good values, the better results you can achieve.
|
Unfortunately I can't recall anybody else with similar problem. And they are flying with very same units.