Large drone for agricultural processing, pid and all

Please watch the video I can’t understand what kind of oscillations are in the Loiter mode

i am think! 25 kg min / max 35 kg =0.714

0.714 * ATC_ACCEL_P_MAX (23761)= 16965
0.714 * ATC_ACCEL_R_MAX (26520)= 18935
0.714 * ATC_ACCEL_Y_MAX (6845)= 4887

Yes that is the correct use of those parameters. They can be rounded to nearest hundreds or thousands, individual digits and decimal places are not critical.
Keep a record of the good autotuned values. You can even have a .param file (which is just text) with the parameters you use for NO payload, and a .param file for WITH payload. This way you can reliably swap between the two different configurations without making mistakes.

I’ll check your log and video, I hadnt been able to get back to them yet.

AutoTune: Yaw(E) Rate: P:0.716, I:0.071, D:0.0000
AutoTune: Yaw(E) Angle P:4.688, Max Accel:6845

AutoTune: Pitch Rate: P:0.140, I:0.140, D:0.0186
AutoTune: Pitch Angle P:5.985, Max Accel:23761

AutoTune: Roll Rate: P:0.145, I:0.145, D:0.0220
AutoTune: Roll Angle P:7.879, Max Accel:26520

That looks like a D term oscillation, and in the log it’s affecting the position too, which matters in Loiter. For the short time you are in AltHold the oscillation completely goes away.
I would start by lowering the D terms produced by Autotune, since they are extremely high, and increase the logging:


These should be safe and wont cause instability.
I’m not sure why Autotune would have produced such high D terms, but it’s no big deal for us to set them to something effective.

And temporarily set these position controller values - they are just half of default so easy to put back in place later:


When you can, do another test flight with good comparable sections of AltHold and Loiter. We might be able to put back those position controller values to normal if we think the attitude control is more stable.

Great! thank you very much.
I would also like to change these parameters a little. I think they are too big

ATC_ANG_PIT_P befor 5,9 after 5,6
ATC_ANG_RLL_P befor 7.8 after 6,5

p/s I have long wanted to ask about changing the parameters with a toggle switch on the control equipment, if this is possible. Although if the load is consumed gradually, then this will be Incorrect

I would probably leave those how they are for now - the extra logging will show if they are overactive or not, and currently there was no sign of high P term oscillations - it was all small fast oscillations associated with D term.

Yes, I hesitated between these and other parameters, but I think that you are saying a very wise thing and I will listen to you

Hi all! Today I carried out tests and flew well, except for the althold mode, there were strange fluctuations in it after moving the sticks diagonally. but after a few hours I charged the batteries and launched the drone without allowing it to cool down to street temperature, took off and it turned over… now I stop flying until the propellers arrive. I think this is due to the fact that there is no temperature calibration.

00000020.BIN - Google Drive this is one flight
and 00000021.BIN - Google Drive crash.

Sorry to hear of that crash.

In the log 20, everything looked great, attitude control is quite good. Good to see you using Stabilise mode too.

In log 21
I dont (yet) suspect the temperature calibration, but more so sitting armed and spinning motors for almost 10 seconds before take off. At this stage I cant see the actual cause though.
I want to look at this more, so I’ll let you know what I find.

Big copters are sometimes more prone to tipping over as they try to move horizontally with a shifting GPS position while landing gear is still touching the ground, or accidental stick input. I didnt see either of those.

Change these:


The RTL ones are not essential, but make RTL
a) not always climb to 15meters if close by
b) always rise at least 2 meters as a visual indicator

@Ilya_Smolnikov Recently I’ve used two different 25+kg arducopter drones that were commercially manufactured, and both suggested doing at least the first take-off of the day in Alt-hold mode and then switch to loiter once in the air. This was just to address the issue that @xfacta raised. You might want to consider such a procedure.

Thank you, you are a very cool person and an expert, please tell me. Maybe there are some courses for learning Ardupilot, I read the documentation on the site, but I can’t put everything together in one pile.

It takes a while to get a grip on all the Ardupilot theory and intricacies. Most people just stick with the bits they need to know for their aircraft (or rover). It can be a huge exercise because Ardupilot is so all-purpose, has over a 1000 parameters and can operate almost any sort of craft.

The big copters, say with 30 inch props, get very sensitive and can be a little tricky to tune - and dangerous! Small copters are so much more forgiving, and cheap when something breaks.

1 Like

This is the problem with ardupilot, it is intended for developers, for engineers, not for ordinary people. To properly set up an ordupilot, you need to have knowledge of: - programming - computing - aerodynamics - electronics and have a lot of free time. and one more thing: - a great desire to figure it all out, plus you need a well-pumped brain that will combine all this into one. have a lot of money to buy new spare parts. and the absence of any specific structured course in order to gradually realize all this only worsens the situation. Yes, I saw an article where it is described How to set up a Drone step by step and methodically, but there are a lot of things there that are not clear, and no one explains how to do it on your fingers and what will happen if you move away from some values, how the Drone will behave if I decrease and increase any parameter. Another disadvantage is the lack of a friendly, human-readable interface. I know a huge number of people who use inav and after switching to ardu they simply go crazy because they do not understand what needs to be done. I don’t know if anyone is looking into this issue, but if this continues, ardu will lose its users.

I am looking into the issue, and my company offers courses and provides documentation for free.

But I rely on you, the users, to explain to me what is not clear in the free documentation.

If you do not explain exactly what is not clear we can not provide clear documentation.
And complaining that the documentation is too long is not allowed, for obvious reasons.

1 Like

I’m not saying anything against ardupilot, it’s the best aircraft control system I’ve used, I’ve tried both Ainav and Bat of Light, I’ve tried dji, but ardupilot allows you to make your aircraft the most reliable and the highest quality, and Thank you very much, a big bow to you for your efforts, but there are some things that I cannot explain, for example, how to make a figure eight using a script to determine the quality and settings of the compass, I will ask questions as much as possible by reading your instructions inside and out

What exactly is not clear in the section 4 of How to methodically tune (almost) any multicopter using ArduCopter 4.4.x which point is not clear?

Why edit anything In these files that include functions that need to be loaded if they already contain included functions that activate the necessary set of parameters for recording what exactly I might need there. I looked into these parameters and saw that every parameter is needed there and everyone needs to be included. What else should I edit there?

In many cases you do not need to edit anything, the file already configures the system the way you need. But you still should confirm the changes.

If you have a different RC receiver you need to edit the RC receiver file.

If you have a different ESC, or need different ESC settings, for an octo for example you need to edit the ESC file.

The same for GNSS and the other files.

The mag calibration is specific to your vehicle you can not use the mag calibration from my vehicle in yours. But the setup can be reused, unless you require 3 IMUs instead of 2.

If that is complicated, we do offer paid consulting to customise your system to your needs.

Let’s mentally imagine two different drones with compasses.
I took the parameters from 15_Magnetometer_Calibration, and I think that all these parameters are necessary for any drone, and the question is, why do I need to disable something here?

You are correct, that file needs no editt, even if the vehicles are extremely different.

But… You might want to change the first parameter to exercise the throttle more.