Hi Tridge,
After loading 2.75, I took the Skywalker for a run. ( I was going to say spin but not good for airplane) I am still having trouble tuning on the 3rd flight but one thing is never did before until after the load of 2.75 was get confused when it reached home under RTL mode. This time it fly home and decided to go left (set to turn right) then it decided it should go right, then abruptly said nope, left, right, left, well you get the idea. Any ideas on what could casue this?
The thing the Skywalker did from day one that was different from what my other plane did, it entered the home loiter left and turned around to then go right. Never figured out that one. Maybe the issue is related to the above.
Another separate issue wrt tuning is the speed of Auto turns, they are quite abrupt. I have tired lowing the Time Constant and the P on roll. Maybe I haven’t done it enough. Which would you give priority to? I don’t want to change to many variables and not know which worked and which did not.
Last, thanks for the help and effort. The “manual” has improved a lot. My input would be to add some more self help in the tuning guides. For example, “if you have this issue, you should do this” and show the graphs. I still am not sure what a very good tuning graph looks like vs a good enough one.
Regards,
Steven
Hi Steven,
For less abrupt turns the parameter to tune is NAVL1_PERIOD. Raise it a bit until you are happy with the turn rate.
For the direction changes on reaching home, can you post a tlog?
Cheers, Tridge
[color=#004000]Split off from release thread.[/color]
[quote=“tridge”]Hi Steven,
For less abrupt turns the parameter to tune is NAVL1_PERIOD. Raise it a bit until you are happy with the turn rate.
For the direction changes on reaching home, can you post a tlog?
Cheers, Tridge[/quote]
Hi Tridge,
Here is the tlog. The RTL sequence that was so bizarre starts at 44.90%.
If you could also take a look at the Auto sequence for any general tuning issues, I would sure appreciate it.
I assume I would make minimum increments of 1.0 to NAVL1_PERIOD?
Thank you,
Steven
Hi Tridge, You may not have noticed but I uploaded my tlog per your request. Please have a look when you have time.
Regards,
Steven
Hi Steven,
The 2nd RTL in your log shows a failure of the loiter path capture code. Path capture is where the code switches from flying towards the waypoint to when it starts circling the waypoint. In your 2nd RTL it approached the waypoint from the wrong side which meant it was going counterclockwise around the home location. As the RTL was set to clockwise it then wanted to turn around to fly in the right direction, but when it did that it left the patch capture radius and went back to waypoint approach mode. Thus the “indecision” in the navigation.
I suspect if you lower NAVL1_PERIOD from 22 down to about 18 this will happen less often, but I will talk to Paul Riseborough about whether we can fix it properly.
I’ve also added a patch to git master to make the problem easier to reproduce in the simulator. You can now change WP_LOITER_RAD while in a loiter. If you change the sign while loitering the plane will try to turn around and circle the other way. If the NAVL1_PERIOD is too high then the behaviour you saw will happen.
Cheers, Tridge
Hi Tridge,
Thanks for the analysis. I had previously raised the NAVL1_PERIOD based on your advice. My WP turns were very abrupt. So now I have a conundrum. I never had this problem on my other plane, an EZStar. Of course it’s a 3 channel and smaller, so that may make a difference. Is there anything else I can do in combination to effect a combined success on my two issues?
Also, today when flying, I had the same anomaly wrt RTL. But I also had a rather severe roll oscillation upon switching to RTL and as the plane entered the erratic RTL mode mentioned previously. I had raised the NAVL1_PERIOD from 22 to 23, so that may have been the cause?
Regards,
Steven
Hi Steven,
ahh, I’d forgotten that, sorry.
Really the L1 code should be able to cope a bit better with this situation. I think I’ll wait till Paul has had a chance to look at the logs before recommending any other changes.
You mentioned it was a Skywalker, but what Skywalker model is it?
Cheers, Tridge
Sorry, I should have been more specific. It is a SW 168cm. Basic easy going SW. Only mods are re-enforcements to the wings. CG, location of APM is pretty standard. I have been trying to tune this plane for 2 weeks. My EZ star was EZ. The SW, which should be easier (ailerons) has been a bear.
Regards,
Steven
that is strange, a SkyWalker 168 is normally easy to tune. I flew one for a long time (I retired it about 18 months ago).
One thing that would be worth checking is your rudder direction. Are you sure it moves the right way? You have it set for “right rudder means higher PWM values” at the moment. Is that correct? You probably have it right, but I’m mentioning it as it has been a cause of bad tuning for other people.
I also notice you have set RLL2SRV_RMAX and PTCH2SRV_RMAX_UP. I’d clear those as they will tend to de-tune your plane a bit.
Overall though in AUTO your plane seems to be flying very nicely, with good path tracking. The only real odditity is the RTL, and I think that is indeed a L1 issue that we should fix in the code.
Cheers, Tridge
Hi Tridge,
Again, I set the RLL2SRV_RMAX at 60 per prior advice from you. This was the first attempt to slow down the abrupt turns. I will zero out both of these “max” values and lower the T CONST to 18. Will try this morning.
Thanks,
Steven