does this look ok or do I need to do something more. basically stock setup in ardupilot.
setup: quad rotor setup in I configuration, 24” props 40lbs.
I just want to make sure my filtering is ok before I jump into PID tuning.
thanks
jesse
Looks ok. Maybe even too good.
Do you have vibration absortion hardware?
1 Like
Post a link to the log file. A misconfiguration for data logging can also look like that.
nothing special on the vibration side. carbon fiber tube for frame with pla+ joints. I did have 10lbs of mass simulator attached at the points where the gas engines are in this picture.
running a holybro pixhawk 6c and holybro gps.
picture:
Log:
This is a rig for the darpa lift challenge. I’m trying to get the thing flying with electric only before I go to crazy town with the gas engines. hopefully when all is said and done the bird will lift 180lbs.
I seem to have a problem where the drone wants to tilt as I’m applying throttle stick. if I punch it off the ground it seems to be able to lift off and hover but doesn’t feel very stable and small changes in throttle seem to drive large changes in thrust. going to get the landing gear strengthened and try again next weekend.
jesse
So it’s pretty much no filtering except that the INS_GYRO_FILTER is at 20 Hz (maybe probably appropriate for a machine of this size, but I will not comment much on this).
The actual takeoff attempts look a bit more realistic, with some -40dB of apparently propellers:
and some surprisingly non-quiet VIBE records:
The PIDs are all default as of now, so it’s no wonder it does not feel anything like stable, I can barely see any control at all in these logs. From PID{R,P,Y}.P plots, which for pitch and yaw seem centered and with an accelerating amplitude, I conjecture that pitch and yaw P-terms are too large to be flyable, for roll it’s less clear (if I guess the orientation right, the machine is much wider and has a much bigger inertia in roll).
Not sure how one does it for big and professional vehicles, but if the class of the machine is all new to me, I would be dialling in initial PIDs and filtering at the same time, because until the machine is stable in the air, it does not enter the correct flight regime that would excite all your loose bolts, nuts and wires.
There is not enough hover time in this log to determine configuration for the Notch filter or the vertical acceleration controller gains.
Change these settings and try for a stable hover for a minute or so.
Disable Fast Harmonic logging in the Log Bitmask.
And
INS_LOG_BAT_MASK,0
INS_LOG_BAT_OPT,0
INS_RAW_LOG_OPT,9
Why did you set MOT_PWM_MIN to 990?
Set it back to 1100 or whatever the ESC manufacturer recommends as MIN.
The initial parameters have not correctly been set:
All your compasses are enabled, that is unusual, expecial for big vehicles.
The unusual high MOT_SPIN_ARM and MOT_SPIN_MIN are a consequence of the too low MOT_PWM_MIN .
ARMING_CHECK != 1 is not recomended.
Why did you changed PSC_POSXY_P from 1 ?
Is the INS_HNTCH value taking the Hover throttle and the FM ratio into account?
The altitude controller is wrong (assuming you learned the hover throttle correctly):
I stopped looking further at this point.
amilcarlucas,
LOL, yes I have no idea what I’m doing. just for reference I have never built or setup a drone before. I don’t think I actually said how hard could it be. 
I do however have a degree in mechanical engineering so not completely clueless but this has a steep learning curve.
side question, should I just reset the controller and start from scratch? I’m wanting to try and fly again on Friday. Any suggestions on what to start with?
I have video of this first attempt I’ll post up later.
Yes, using AMC you can select to reset all parameters when creating a new project.
Start from the very beginning and do every single step (optional steps, are clearly marked as optional, and can be skipped)!
From experience with other beginner users, it will take you about 10 hours of concentrated work. You need to read a lot. The good news is that the software guides you in the processes so that you only read information that is relevant for your vehicle and your components and your configuration. It tailors the sequence for YOU.
amilcar,
I got bad news for you my friend, I have spent the better part of 3 days trying to use your configurator and I can’t get the thing to work at all. I started a fresh file and used all the suggested parameters given my setup and nothing seems to work. I might be doing something wrong but I can’t seem to make your tool work. I can’t even get the motors to test spin. it’s even more frustrating when the documentation says one thing and the configurator requires something else.
not sure where I’m going to go from here. maybe some sleep will help.
Sorry to hear that. What exactly does not work?
Motor test only works if the ESCs are correctly configured.
Explain what components you have, how they are connected and what parameters you set, then we can help you.
But for now get some sleep, three days not making progress is very frustrating.
On the bottom of the parameter editor window there is a button: “Zip files for forum support” press it and post the zip file it creates here. It’s a small file that can be attached directly.
amilcar
I wanted to thank you for spending your time to try and make easier for us newbies. it’s undervalued and I try and give you user feedback from a amateur’s perspective.
so I have a few things. first is a video of the first flight from the data log above.
as a point of information, this setup had 10lbs of mass simulator on the ends of the drone to simulate the expected mass two motors that will eventually be attached there. I have since removed this for flight 2 and 3 as I’ll try and make the drone as nimble as possible to start out.
https://rumble.com/v78zt1m-reasonably-adequate-test-flight-1.html
then I did a little rework on my setup working thru the ardupilot setup pages turning down the I term on roll and pitch and updating the throttle settings a little. here are the main parts of the drone to use as reference for the parameter files and videos later in this post.
- holybro pixhawk 6c
- radiomaster rp3 (into telm2)
- sik telemtry (into telm1)
- holybro m9n gps
- hobbywing x6 plus (running with dronecan on canbus1, no pwn out)
- chassis is quadcopter in H frame configuration
- the bird as currently configured weighs in at around 12.7kg
here is the parameters I had setup for the flight today
Here are the parameters I ended up when trying to use the Methodic Configurator with a sample drone in the database.
here is a link to the two test flight from today along with video links of the flights.
Flight 2
https://rumble.com/v7907fk-reasonably-adequate-flight2.html
Flight 3
https://rumble.com/v7909f0-reasonably-adequate-test-flight-3.html
I did get the bird in the air and it seemed reasonably stable for a second try. From what I can tell in the log, the notch filter seems to be doing ok as almost all noise is below -50db. not saying it can’t use some work, but maybe good enough to get started on tuning the rest of the system.
I seem to be having a time keeping the thing at a stable hover altitude and will have to experiment with min throttle settings.
jesse
@onephatz where is the .zip file I asked?
What would prompt you to reduce the I terms to .01 with the P terms at default?
I’ve checked log 3, you barely need any filtering, LPF ever at 50hz looks adequate enough, it can reduce the filter delay from current 20ms to like 5ms.
Fast attitude in LOG_BITMASK will help to check your PIDs.
Uploading log from 1st flight also can be helpful.
1 Like
from my reading, the thing that gets untuned drones into trouble initially is an a PI ratio that has the I too high and results in oscillations and crashes. I have a couple thousand into this bird already so I’m playing it safe at the start and can bump it up as tuning progresses. my understanding is that it will be more “sluggish” during flight but doesn’t risk a crash.
mike, the first flight log was posted further up this thread. I will admit that the poor first flight was likely me overlooking some parameters and being a bit of a newb to all this.
I don’t know where you saw that. For Arducopter
That ratio in most cases should be 1. You will find this in the Manual Tuning section:
Each time the P term is changed set the I term equal to the P term.
1 Like
This is seriously borked. the Number of battery cells in series 12 (12S) and not 2 like you entered in the component editor. That is the number of cells in series, not the number of batteries in series!
And the capacity is 3000mAh ??? for a 24`` vehicle?