This is a discussion re the ArduCopter-3.1 AutoTune feature. A place for people to report successes and failures, ask questions and get answers.
The AutoTune wiki page is here.
You need to be a member of diydrones to add comments!
This reply was deleted.
Replies
Attached is AC3.2 Autotune log
2014-12-22 15-43-29.log.zip
Hi Artem,
Looking at your autotune log I can see what is happening. I am not sure why but your copter is responding slightly differently than usual. This may be your particular combination of motor and prop or it could be the fact it is a tri.
In any case looking at your autotune I think your roll and pitch rate D terms should be approximately 0.007. Your roll and pitch rate P and I terms should be approximately 0.12. Then with a Stab P term of 10 on roll and pitch you should be fine.
Let me know how you go. I will think about how we might be able to detect this type of response in the future.
Thanks
Leonardt, thank you! I will try this numbers as soon as I'll get a chance. My motors are 4010 390kv and props are 15x55, same was happening with 4108 380kv and 4016 400kv. All pretty standard motor/prop combo. May I ask you how you were able to determine these PID numbers? Thx!
Artem,
I looked at each test that Autotune did to see when it started to miss the correct interpretation of the rate response curve. I am looking for the point where the D term causes the rate to bounce back. In your case it is doing this much earlier and with a lower magnitude than usual. This then results in the bounce being missed. Once I find these points I look at the pids at this time and then make some adjustments based on the maximum rate achieved.
Something I would appreciate you trying is doing autotune with your INS_MPU6K_FILTER set to 10. You may find this fixes your problems. If you are willing to try this I would really appreciate it!
leonardthall,
I have finally found time to do autotune at 10hz mpu6k filter.
logs attached.
2015-01-07 15-15-19.bin
leonardthall,
I have attempted to fly with mpu6k filter at 10hz, however copter started oscillating badly on the PIDs you mentioned earlier and I it was too early in the morning and very cold so my brains were "frozen" along with the fingers and I didn't think of resetting PIDs to defaults at that time. Anyway will be attaching log of that short moment when I took of with MPU6k filter at 10hz and attempt autotune at 10hz mpu6k next year :)
It appears that partially the problem with shaky video was with the T2D gains, after I adjusted them it maintains smooth video even with the mentioned PIDs, Strangely I am flying that gimbal with AC3.1.5 autotune PIDs since September w/o shakes.
2014-12-31 07-48-07.log
Ok, I will run autotune at mpu6k @10hz, just finished flying 26min with the PIDs you suggested, and YES those are really good PIDs except they make motors make a really high pitch whinning sound constantly, also, when looking at the footage I could see some really bad low frequency vibration, not jello though, at 60fps is not able to focus.... To make sure this is not the props.motors issue I changed PIDs back to old manual PIDs I mentioned up couple of posts, all vibrations are gone, and no funny noises, but the copter is much less responsive than with yours... I will try increasing D a bit tomorrow as well as try autotune at 10hz MPU6k
I hope that I am on topic with this. I have version 3.2 loaded into a Tarot Pro hex extended for 15" props. I am doing missions and have run into 3 problems. When I create a mission with Takeoff as the first command and follow it with several waypoints, some with time delay and some without, the hex will not proceed to the next command. I arm in stabilize with the throttle down then switch to auto and advance the throttle. It takes off perfectly then just sits still at the takeoff altitude I listed in the command. If I remove the takeoff command and start with a waypoint, it will execute perfectly, going to all waypoints. Similarly, if I have a loiter_turns in the mission, the hex executes the turns ok but skips the following waypoints in the mission and seems to skip to the last command which is "return to Launch". The other problem is with the mission planning page. The rows and columns for the mission show up totally collapsed when I begin. It is only when I enter commands that the columns open up but only enough to show the width as never more than the number of characters that I have entered in the numerical cells. So for example, I cannot see the word delay in the header if I only have 1 second typed in but if I type 1.000 then I can see 5 letters of the word delay. Is this a bug in mission planner or do I need to make a setting change for how the display works?
I am very happy to be this far along and appreciate the help I have received both on this forum and by the excellent support staff in Mexico.
Bernie
Bernie Liskov ,
Sorry, this is not the right place to ask those questions. I would suggest the 3.2 thread or open a help thread on the arducopter forum.
regarding AC3.1.5 autotune PIDs....... I cannot find the setting file I saved, I think I've overwritten it with the AC3.2 setting file....will have to reflash 3.1.5 and redo the autotune.... should have time on the 25th, will post then.
-
13
-
14
-
15
-
16
-
17
of 39 Next