Heading hold on Yaw Issues

[PatrikE]

There's problem with rudder on Planes.
The rudder behaves like Heading hold is active.
Planes is normally turned with ailerons & elevator. (Bank & Yank)
The Rudder will then countersteer to maintain the heading!
The heading is reset every time rudder stick is used.

Flying in ACRO and using the rudder to make "Clean" turns works fine.
The I term is set zero when you use the rudder stick so there's no problem if you use rudder during turns.
On copters this is not a problem because Yaw is used more actively for flying.(Or actually it makes it's harder to keep nice sweeping turns..)

The behavior is tied to YAW I.
If i zero Yaw I The rudder behaves in normal GyroMode.
With stock I=0.045 it will try to hold heading.

This will also affect NavCode for FixedWing.
The result will be the plane can fly sideways!
Rudder and Ailerons in opposite directions in combination with elevator is also the way to normally initiate a Spin.

My conclusion is..
Current Yaw I implementation is not healthy for Airplanes

Fly safe without Yaw I
/PatrikE

[shikra]

Is this noticeable since going to 2.3

I must admit with the copter I don't like 2.3 yaw when flying FPV, but do prefer if flying LOS. Find the yaw bounceback on tricopters not so nice in the goggles so have hacked it alittle to be more like old style.
Wish it was an option to use old style....

[Plüschi]

Would it be useful to put a "errorGyroI[YAW] *= 0.99;" in each cycle to have a decreasing I term?
Me too i wasnt very happy with 2.3 yaw.

[shikra]

Funny I never saw anyone mention not liking 2.3 yaw then you post and I get a couple of conversations the same last night. I love pitch/roll pid.

It only really annoys me on the tricopters and flying FPV and centering sticks after after yaw input . The springback is a not nice feeling under the goggles. Things like that just don't show up in normal LOS flying. I don't notice so much it much on the quad. Tricopter servos are much more effective I guess

I like to try your suggestion too.
I added some code so that when when the gyro direction changes after stick change it set accumulated gyro I and the various D delta values to zero. Effectively stopping the bounceback.

[PatrikE]

On a copter it works good for hovering and normal LOS flying since you correct the heading with Yaw while you fly.
But on planes it becomes wrong..
If you bank left and supports with elevator the plane will start to turn.
But the Yaw will try to maintain the initial heading and stear right!
This will result in something looking like KnifeEdge flying.

For planes something else is needed.
If it works like that when flying in Angle mode to fly straight line then it's ok.
But not in Acro Mode it should act like elevator and ailerons Just compensate for sudden changes.

[AlouetteIII]

This new advance in 2.3 yaw is very nice for acro boards, some very clever coding by AlexinParis - bravo ! Great for Noobs or for PARIS AIR people who have not using a GPS/MAG sensor - however its getting a lot of negative feedback from PARIS v5r3 GoPro film maker flyers who fly 2.2 then upgrade to 2.3 specifically doing aerial video with a MAG - both with MAG on or OFF - even in ACRO the aircraft now tends to yaw smoothly then "jump" a few degrees or it tries to return to a previously held heading (when the stick is centred to 1500) even when MAG is OFF - ruining the film makers effort to yaw very smoothly.

@AlexinParis - > serious Request for 2.4 = have the ability to use old yaw code 2.2 ; or to choose new 2.3 yaw code in config.h

[Deet]

I posted about this earlier this year

It doesn't matter if MAG is selected to not, the yaw hunts as it it is trying to maintain a fixed heading.

In fast forward flight you don't notice the issue, but any kind of hover, or trying to yaw in slow flight, the board seems to fight you