When Ardupilot auto trims the ailerons I end up with both of them deflected up, one more than the other. Ive set differential by setting min & max. I think that’s the correct way? All directions are correct and the plane flies in manual and auto modes. Im not using the DSPOILER parameter as I think its for complex wings?
Its not ideal having the both up a bit, limiting Cl max.
Any suggestions?
I dont use auto flaps or any sort of low speed flaps, its auto trim. Its supposed to figure out if a bit of aileron is required to keep level and move the trim center a bit but I think my way of setting differential is upsetting it. I dont know any other way to set it.
Yes probably.
I dont want any reflex, I just want the ailerons centered and if ardupilot wants a bit of aileron trim, I expect one to go up and the other down a bit.
1 Like
Still having this issue. Both ailerons have auto trimmed up about 8 deg again.
In a simple wing should I be setting the two aileron servos as aileron or flaperons? I dont want (at this stage) to do anything clever like droop or reflex but I think there is quite a bit of adverse yaw (in manual) so would prefer some differential.
I can see I could probably solve this issue by just setting 2 outputs as aileron and then set trim and endpoints.
So what is the auto trim attempting to achieve by putting them both up? I have not set a channel to flap (208).
I thought the problem was perhaps because I set the aileron servos as flaperon. I now have them set as simple aileron and the auto trim is still putting them both up. that cant be right? Im thinking the logic that trims them is not taking into account one is reversed.
I hope a dev can take a look at this.
Do you have differential ailerons?
It is a known issue that for differential ailerons they will be lifted and made more symmetric in deflection.
I have a bit more up than down on the aileron servo limits if that’s what you mean?
I dont think there is a specific differential aileron setting for simple wing? I know there is some complex settings for gliders/spoilers etc but I dont have that set.
Even with differential set by end limits I cant see how they both trim upwards.
Yes, differential ailerons are typically done with endpoints (at least on UAVs). Autotrim uses a primitive mechanism of adjusting SERVOx_TRIM by one every few seconds when trimming conditions are met. If one trim reaches the limit calculated using symmetric deflection assumption only the other works leading to raised ailerons.
Hardware report is useful for looking a trim changes.
It shows param changes over time. Your ailerons are on servo 1 and 2. We see SERVO1_TRIM start out at 1509 and end up at 1588, so a increase of 79. SERVO2_TRIM starts at 1493 and ends up at 1455 a decrease of 38. So one goes up and the other down. SERVO1_REVERSED is 1 and SERVO2_REVERSED is 1 also so your right, we would expect them to go the same way.
However, we never trim outside of the the middle 20% of the range. Taking servo 1 as a example,
SERVO1_MAX is 1970 and SERVO1_MIN is 1255. So the mid point is 1613, the range is 715. So the trim range is the mid point ± 10% of the range. 1613 + 71, 1613 - 71 1542 to 1684.
For servo 2 to the trim range is 1319 to 1463.
The problem in this case is that we will trim the servo towards the trim range but never away. So slight variations in required trim nudge it towards the trim range because it can only trim in one direction.
If we look again at the param changes over time we see this:
To start with the trim only increases then once it is in the trim range it goes up and down a bit.
Servo 2 is the opposite, it only decreases to start with until in the trim range.
This stuff is a bit tricky, the trim range is there so if something does go wrong we don’t endup trimming to a extreme. We could just refuse to change trim at all if the starting point is outside the trim range.
Ultimately, for now anyway, there are just some setups where auto trim does not work well.
One method we have talked about is rather than trimming individual servos directly we have new parameters for roll and pitch trim, that would mean that it does not mess up your servo trim for differential.
3 Likes
Peter, thanks for taking the time to explain that. Very interesting. I will probably just remove the differential. I think in everything except Manual the FC will take care of adverse yaw anyway. Ill do a video looking at the tail and test the adverse yaw in manual and FBWA.