WARNING to TradHeli users:
Do not attempt to fly AC3.3 until you hear otherwise from me. I have not even bench tested it yet, let alone test flown it. There are quite a lot of changes that have gone in that could be crash-o-matic for helicopters.
AC3.3 release for Tradheli is likely to be delayed.
This post will serve as the official release progress thread for TradHeli.
Warning: DO NOT FLY RC6! It has a critical bug that will likely end in a crash.
Warning for RC5:
RATE_YAW_FILT_HZ param will change defaults from 5Hz to 20Hz to be more suitable for helicopters. If you have never changed this parameter from 5 before, then it will automatically update the value to 20 without warning. This much faster filter will probably require reduction in your P and D terms, and the yaw may oscillate badly. Be prepared to retune these. As a safety measure, probably cut them in half for your first take-off.
Warning #2 for RC5:
I accidentally broke the Landing Detector. It will probably never trigger. You should land in Stabilize mode, or be prepared to shut down the motor as soon as it touches the ground in Alt_Hold, Loiter, Auto, etc.
Replies
One really important thing with AC3.3 that I wanted to show you guys is the new Rate logging message that has been added. It's super useful for heli guys, particularly with doing our Rate FF tuning. The problem with helis has always been that the PID terms are not capable of controlling the helicopter properly, because they cause oscillation before they get to the point of giving control authority. But the FF params are difficult to tune, because they can't really be done in a simple hover, and really require pretty strong manoever, like pushing the heli 45° side to side. Meanwhile, you have to watch for the helicopter to respond strongly, but without wobbling. You're actually looking for a rate oscillation *within* the movement.
The Rate logging helps a lot with this. Here's is a great example from my Trex 500. The first two plots are the Roll, and then Pitch, Rate Desired versus achieved rate. You can see that the achieved rate (green) does not quite match the desired rate. It's not bad, but not good enough. I probably need to add a bit more Rate FF on these.
The third graph is the yaw, and this is the one that's really interesting. It shows the heli is actually overachieving the desired rate. This is probably because the Rate Yaw FF is too high, and the PID which is the error correction portion, doesn't have enough authority to fix it. So I will try reducing the FF a little bit.
Thany You,
that is very useful for me. By the way, I couldn't find 'Rate Yaw kI' in 'Tune' using MP 1.3.25 (kP=6, kI=16?, kD=26).
Hmmm... I thought it used to be there, but it's gone. I'm not sure what happened.
Sorryy bout that. I don't think it's a big problem to do that one manually, it shouldn't be that sensitive.
A few flights to tune out the vibrations on the TRex 550 with 3.3rc3
With the vibes, the land detector only worked a third of the time. When it did. Take of in loiter was good, but would swing about 45 degrees with each takeoff. The pixhawk can compensate the tail when taking off in stabilize, so it's probably a vibe induced heading issue while on the ground in loiter getting ready to lift off.
Short auto mission worked well take off until land Except didn't disarm (probably due to the vibes/landing detector issue).
Not really unfortunately. I only started flying my helis again recently. It's looking like 3.3 will be mostly passthrough for heli. This summer I should be able to get much more done. But before I do, I want to make sure to pull in yours and fhedberg's PR's. In fact, I've been holding off in touching the heli stuff until I get fhedberg's stuff in. (yours is pretty self-contained).
I would like to try to squeeze in a few things before AC3.3 goes stable. I've got a thing to make the Rate FeedForward LPF a parameter. That's actually leftover from AC3.2, it never got pulled in then. I would also like to add a Leak Min parameter. What it would do, is the Rate I-terms would only leak down to a certain point, then would hold. So you could set it to say 500, which means that of the typical 4500 I-Max, it would leak down to 500, then stop leaking. The idea is that, 500 is probably not enough to cause a ground tip-over, but would help a lot in hover. Right now, I find that the helis suffer a little bit because they don't have full I-term in hover.You can actually see it happen if you pay attention when your heli changes from Full-I to Leaky-I as it slows below 3 m/s.
I think I can squeeze this into AC3.3 before it goes out, but we'll see.
I'm actually going to be trying to figure out if there's a way for me to do mini-releases for helicopters. I don't want to have to wait for next fall for you guys to get any released improvements. If I have to do put out some compiled binaries, where we have AC3.3 + Heli Improvements, I may do that.
Mini-releases would be nice. Think that's part of the solution to speed up development and improve the APM community.
This has been quite successful for the Linux kernel where a lot of the sub-system maintainers release their own branches.
Yeah... there's an internal debate about this. Some people want a much faster release cycle. Some people think that we can't support major releases on more than a 6 month cycle.
Hopefully we find some middle ground between me publishing binaries for side-loading, and Fall 2015 for new helicopter improvements.
-
29
-
30
-
31
-
32
-
33
of 34 Next