I’ve been recently working with a Boxkite VTOL platform very similar to @hwurzburg 's post from a few years ago "Box Kite" Copter Tailsitter Build instructions and thought I’d post some logs here and ask a few questions.
I know I’m using a poor flight controller for this, but it’s just what I had, so I’m not even trying to use bidirectional dshot here because I understood this flight controller has issues doing that conflicting with other tasks. Anyways,
Specs -
Boxkite frame very similar to Henry’s above (foamboard)
4x 1806 motors with 5in triblade props
2 Elevons on the bottom wing (different than the linked elevons Henry used)
Speedybee F405 v3 flight controller
Speedybee Mini 35A 4in1 ESC
ELRS RX with Mavlink setup
Matek GPS/Compass Module
Does anyone have any thoughts about whether this model could use only two motors for fixed wing flight? Would top two or bottom two be better? I’m thinking about biplanes, but I’m not sure if that translates. I tried my model with just two motors the other day (top two), but it consistently had issues in fixed wing mode with only the top two motors enabled in FBWA. I haven’t done any tuning on it yet, but was wondering if I would need to mess with Q_ASSIST amount/length of time it’s engaged to help the plane perform a better transition before it could fully shut off the bottom two motors.
Any help, thoughts, or advice would be greatly appreciated!
I was also wondering how to vary the amount of differential thrust (rudder) applied in fixed wing mode, but of course have full yaw in quad mode.
Log of failed transition to FW mode (I still think it’s incredible that that Ardupilot was able to save it after I switched back into QStabilized mode)
using two motors wont work because the thrust line is so far off from the aerodynamic center in fixed wing flight
and that autopilot is just fine and will work just as well as an $300 autopilot. What
expensive autopilots buy you, in general, is redundancy and reliability over a $50 one
an H7 based one gets you LUA scripting and bigger flash to allow more esoteric features than an F4 base one (but you can use the custom build server to remove unused features and add some of those esoteric ones back in, if need be)
Thanks @hwurzburg I appreciate your response! That makes sense with regards to trying to use only two motors - I’ll stick with four then. I do find amazing that I can get a VTOL to even fly with a 50$ flight controller, but for me that’s one of the amazing things about Ardupilot to begin with! Thanks for your original box kit build as well - it’s been a great reference/inspiration for me to try to play around with VTOLs.
thats what controls it in twin fixed wing configurations…but bear in mind a Copter Tailsitter in FW mode may not use that since its yaw stabilization is active, where all other fixed wing configurations usually do not use yaw stabilization (by default)
Oh, I did see something about that yaw stabilization in the wiki. Thanks Henry. Is there any way to adjust the amount/rate? of yaw stabilization used on a VTOL?
I’m planning to start working on the tuning for this vehicle soon, and after reading through some documentation here - Tailsitter VTOL Tuning — Plane documentation - I am trying to determine if I can use the Quicktune feature on any of the axis of my Boxkite. Because it has elevons, I would only be able to use QuickTune on my roll axis, correct? I can then follow the procedures detailed on that wiki page for yaw and pitch?
if you are able use the QUIK TUNE LUA applet using ALL axis for the tune…it gives a safer and quicker tune…using QAUTOTUNE mode (not preferred any more) usually roll/yaw done independent sessions from roll…anyway QAUTOTUNE takes a looong battery duration
Thanks for your response, Henry. My question was more specifically can I use QUIK TUNE for all three axis of my VTOL in Quadplane modes? I wasn’t clear on if I could actually use it for all axis since the elevons would be affecting things on some axis even in QSTABILIZE. Then of course I would do the fixed wing tune separately from QuickTune.
you should use QUIK TUNE for all axes in VTOL…then once in fixed wing mode use the normal AUTOTUNE to tune for fixed wing…elevons working in VTOL is how VTOL works and the VTOL mode should be tuned with all controls working
Well I think I’m in for some more learning - had a few flights today with multiple distinct aircraft of the same construction where we all had a similar issue, but mirrored, in a way. Myself and one student had our planes give large amounts of uncommanded right yaw in FBWA mode, while two other students had the same things occur, but with left yaw. Some of us were able to fly normally for a minute or two before the aircraft just started to perform its yaw/turning on it’s own. I don’t know what is causing this. I originally thought it was motor misalignment, but I have checked that and all motors are facing straight forward mounted on the foam board wings. The other thing that really stood out to me was the split with two aircraft turning right on their own and two turning left. The only significant physical difference between the aircraft is that two of us have our batteries mounted on the LH side of the center fuselage/strut and two have batteries mounted on the RH side. But I wouldn’t think this would make a significant difference, and if I remember right, that RH/LH battery mount distinction did not line up with the RH/LH uncontrolled yaw we saw in flight.
I’m really at a loss here on what could be causing this - is it something to do with our frame being setup as an “X” quad? Something with a compass anomaly? I don’t really know where to start so any help would be greatly appreciated.
Folder with 4 log files from 4 distinct aircraft, though all are nearly identical in construction/layout.
By default there is no active yaw control in forward flight. You do have differential thrust setup from the VTOL motors, this is only driven by the pilots yaw stick.
You could enable active yaw control/damping as outlined on the wiki:
A bigger hammer would be to enable the VTOL rate controller in forward flight. You could set the two Q_OPTIONS given here: Tailsitter Planes — Plane documentation
Checking my understanding - disregarding the OSD parameters that seem to be causing bad loop rates, would you say that the main issue here is that I was expecting yaw to be stabilized with the differential thrust, but this is not the case in Arduplane by default?
Because the planes keep turning “by themselves” would this point to a physical imbalance in motor speed/output or airframe variations that then leads to them in essence “needing some trim” if this were a manual non-flight controller type vehicle? e.g some type of twin engine fixed wing model.
Thank you for pointing out two methods of remedying the issue. If I understand it, I could setup and tune the yaw damper on the aircraft, or I could set one of these two Q_OPTIONS up. Do you recommend any pros/cons of either? the Q_OPTIONS method seems a bit more straight forward - I’m leaning towards Q_OPTIONS bit 8:Mtrs_Only_Qassist just based off a quick look.
Right. On a “normal” model you would trim it with the rudder servo.
The Q_OPTION method would be quicker to test. It just takes you away from a traditional plane a little, for example the motors will always be active so you cannot glide.
Looking at this a bit more in depth, I’m understanding that I would need to set both bits 7 and 8 to test this, correct? Reading the descriptions, it sounds like if I only have bit 8 set, it is solely forcing Q_Assist to use “Motors Only”. But I have been flying with Q_Assist = -1, so I would never see the VTOL rate controller engage unless I set Q_OPTIONS bit 7 on, and/or enabled Q_ASSIST. Is that correct?
I do find amazing that I can get a VTOL to even fly with a 50$ flight controller, but for me that’s one of the amazing things about Ardupilot to begin with! Thanks for your original box kit build as well - it’s been a great reference/inspiration for me to try to play around with VTOLs.
