Greetings,I finally had a chance to test the FMA, and ArduPilot. It was relatively windy, so not the best conditions for a big foam plane (Twinstar), but I don't get to choose the weather :-)The FMA works quite well at keeping the plane level in the x/y axis. I had the gain set to max, which I think may be part of the turning issue. Flying the plane manually, via rudder, with the FMA turned on requires quite a bit of rudder to turn. It's an ugly, funny looking turn, but it does come around eventually :-)With the plane flying away from me, FMA on, I engage the AP, and hear the motor change speeds. Unfortunately, in a number of tests, I saw no sign of the plane trying to turn in any direction.I walked around with the plane a bit, no doubt creating a topic of conversation for the rest of the people at the field who had no idea what I was doing. When walking away, the rudder turns maybe 1/4" to the right. If I turn around and walk back toward "home", it straightens up.Is that all it does? Should it ever turn left, and should it turn more than about a third of the available throw? Can the throw be increased?Thanks,Rusty (just becoming dangerous)
You need to be a member of diydrones to add comments!
I tested the Ardupilot with the latest code in RTL walking around with the airplane in my hands and the rudder was all over the place - definately not turning me toward the lauch point. Ok, I said, obviously the rudder is reversed.
I go back home, load the NE bearing test code and also change the yaw servo throw to 0 (normal) instead of 1 (reverse).
The second test was almost successfull... what do I mean? The rudder was turning towards the left when I was walking East or South, in order to "turn me" to the 45 degree bearing and then was becoming straight when I was on course. BUT when I was off course on the other side - walking North, Northwest, West it was not turning right and I am almost sure it was turning left... what is going on?
Are these normal reactions? Does Ardupilot turn only in one direction? (only left or only right if reversed)?
Additionally, walking and looking continuously at the rudder for quite some time it looks that there is only two positions: left and straight - no in between. Also true? My max/min PID yaw gains are at +/-30 per your latest advice.
Yes, I saw the same problem with ArduPilot in high wind today. This is caused by PID gains (they control rudder throw) that are too low (we tested them in windless conditions.). I've increased them from 15 to 30 and made that the default in the standard code. See this post for more.
Replies
I tested the Ardupilot with the latest code in RTL walking around with the airplane in my hands and the rudder was all over the place - definately not turning me toward the lauch point. Ok, I said, obviously the rudder is reversed.
I go back home, load the NE bearing test code and also change the yaw servo throw to 0 (normal) instead of 1 (reverse).
The second test was almost successfull... what do I mean? The rudder was turning towards the left when I was walking East or South, in order to "turn me" to the 45 degree bearing and then was becoming straight when I was on course. BUT when I was off course on the other side - walking North, Northwest, West it was not turning right and I am almost sure it was turning left... what is going on?
Are these normal reactions? Does Ardupilot turn only in one direction? (only left or only right if reversed)?
Additionally, walking and looking continuously at the rudder for quite some time it looks that there is only two positions: left and straight - no in between. Also true? My max/min PID yaw gains are at +/-30 per your latest advice.
Dimitris
Thanks! That's exactly what I was hoping for, so I'll give the higher gain a try. BTW, funny about the upwind throttle dilemma :-)
Sgt Ric,
The plane is a TwinStar, though I hope to be making my own design in the near future.
Cheers,
Rusty (CNC router in the garage, CNC foam cutter in the mail)