Are you really using the “live calibration”? That one is hidden now. The default one is now the “Onboard Calibration”.
If you’re doing the one w/ the green bars that move across the screen, that’s the onboard calibration.
If you’re doing the one w/ the 3 3d crosshairs, that’s the live calibration.
If you are really doing the live calibration, do the crosshairs move as you rotate the copter, and do the dots populate the area around the crosshairs?
If you aren’t really doing the live calibration, you need to hold down shift or control (I forget which) when you click on the “compass” menu item in mission planner. Then it will show the hidden live calibration.
And what worked for me was doing the live calibration first. Then after that saved, I was able to do the onboard calibration. (The onboard should give better results, so you should do it after the live one)
Also, seeing that you were talking about making an 18650 2s, here’s an 18650 4s I made a couple weeks ago.
I’m using battery holders so that I can hot swap and not have to solder onto the batteries. Though, at the expense of some weight. It also makes it a heck of a lot easier to charge your 18650s.
Oh yes, I’m doing Onboard Calibration. Got confused reading through the documentation!
I will try actual live calibration first and get back.
Edit: Well, live calibration worked! I’ve got some new offsets and the headings as seen in Mission Planner are pretty spot on. I also ended up changing ‘Compass Orientation’ back to ‘None’. It’s ‘control’ to show the live calibration option by the way.
The onboard calibration is still not giving me updated offsets however. I’ve tried a few times and restarted the program/flight controller a few times too. I will try again tomorrow. Thanks for all your help @wicked1!
I want to mention something I have noticed w/ this small copter…
Vibrations are causing problems w/ alt_hold.
But, the vibration readout in Mission Planner is always very low.
I unfortunately don’t have logging on this FC. Hopefully some day the onboard data flash will work.
My only reference is the readout in Mission planner. But, adding more foam/protection has not helped (It did initially, but that appears to be OK now). Changing the FC mounting does make a difference.
So, to anyone building a tiny copter, keep that in mind. Vibrations might be causing a problem, even if the vibration read-out in mission planner is almost completely flat, showing no vibrations.
My guess is the vibrations are different (Different frequency or something) on this size copter, so aren’t being detected by whatever MP is using to display vibes.
of other things you can try - set GND_EFFECT_COMP=1.
i also spent some time this weekend experimenting and i can tell ekf3 works better then ekf2, i got my smaller model to fly now almost fine. it still shows from time to time those unexplained yet roll twitches and i start suspecting it is something in the code interfering with the dshot data stream consistency.
here is the param file, you can compare with yours - all modes work fine now, including stability at stab mode full throttle at 100A current drawn.
kakute f7 here has double foam layer - if you have IMU not on foam - use matek`s rubber posts, but, i would simply used any FC with soft mounted ICU. on this smaller build i have kakute hard mounted - wanted to see if it will work, and it does work fine,so, for me now it is a controller of choice.
everything there works, only part i did not do yet is a proper support for an arming switch, but, on a small model imho it is not needed. https://drive.google.com/file/d/1IPEZxpNJMYgCiLSTYazTAXwi_FcO8X-V/view?usp=sharing
PS. it is not only foam that is needed - tape controller around the edges. A simple test is this - get model connected to MP, go to status window, look at the ‘alt’ value and try to blow at the FC. if you see that alt value jumps just from you blowing at the controller - put tape or anything around the edges to stop the propwash from smashing into baro.
Thanks Paul.
I’m getting mine finalized really well too. I flew it last night, and forgot I was flying my Arducopter, and not my Betaflight copter! I ended up raising my PID’s higher than what AutoTune gave me. I started at .03. I’m up at about .05 for P,I for Pitch and Roll. But, my copter is relatively weak. I’ve got all the speed limits increased (atc_accel, atc_rate, and acro_*_p), so it’s very responsive.
At some point I need to try to tackle the hwdef file to enable ESC telemetry and a buzzer… But I’m not a programmer, and am dragging my feet on getting ubuntu set up for the work.
Only other issue is GPS interference. I always get about 10-15 satellites, but high HDOP and sometimes velocity or other GPS related prearm errors. It clears after a while, but sometimes it takes 10 minutes or more. I’ve got a couple other GPS on the way to try.
I’ll post my updated parameters file sometime soon.
If you’re using Windows 10, you can get setup using Windows Subsystem for Linux really, really easily. You can also run simple GUI programs using something like X410 or VCXSRV.
I know you have your camera taped, but in my experience it’s always the video signal that kills GPS. Check for any cabling that’s still exposed. Don’t find ESC’s affect GPS much (but compass a lot).
Those were Leonard’s recommendations. I believe the accel filter down is to make throttle response smoother on these overpowered vehicles. Not positive, though. Gyro up is for faster response.
Thanks for the video advice. I do still have video signal wires exposed. I could switch them to coax cable (receiver antenna wire).
Otherwise, my VTX is definitely causing an issue. Sometimes I don’t get any satellites until I unplug my VTX. Then I’ll get 10 or 12 really quickly. Then I can plug my VTX back in and it keeps the satellites and works!
My matek GPS arrived yesterday, but I haven’t had a chance to try it yet. The antenna is 1/4 the size of the one I’m using now! So, not sure how that’s going to work… But it has filters and is probably better quality overall than the bn-800 I’m using now, so maybe it will work better.
I was thinking last night… Maybe I’d be better off w/ a GPS only gps receiver with a narrower band antenna, rather than the tri-band GPS’s everyone sells these days, that have a pretty wide frequency response to get all the different satellite systems.
With any filter setting you need to set it as low as possible without compromising your performance. That way you get the minimum noise through your system.
I find that I can drop the accel filter down to 10 hz and not compromise my alt hold performance. We actually made it the default but a few aircraft would get an oscillation at 10Hz when setup for 20 Hz so we put it back to 20 Hz.
The reason I use 10Hz is we don’t need the response time for alt hold we need for attitude control. I like to drop the work load of the motors in Alt Hold.
Hmm, now that you mention it, my 250 size quad is one that started oscillating when that default was changed from 20 to 10hz.
But that’s before I had tried tuning any of the other alt_hold related settings. This little 160mm copter definitely had some altitude oscillations when I first used it, but was able to tune them out w/ other settings.
may be you could help me to understand what is the deal here with this thing. I cannot fix this issue - in this video the worst of it is the wild twitch at 1min33sec into it.
it happens sporadically and i cannot find any sign of any hardware issues. yet it is the only model that does it.
is there anything in those parameters that screams fault to you? i got it autotuned several times - sometimes it flies flawlessly, sometimes does same weird twitch mid-flight - once or twice in a 10 min flight.
also, that odd pulsation and a twitch at 5:45 till 6:10 or something when i tried just to guide it smoothly forward - it is not normal. yet, i can`t see in the log what it is. and what bothers me most - is the amount of power it throws into motors - like at the first twitch at 1:33 it pumped in 80A into motors - why?
I can see the twitches are external and not commanded by the autopilot. It may be noise on the signal lines a sync issue or a poor connection.
Your aircraft can draw 80 amps at half throttle by the looks of things so it doesn’t surprise me that it can draw that much current recovering after a twitch.
I see some oscillations so you may be a little close to the edge on the tune.
