I made a first attempt at transition to fixed wing today with my Mini Talon quad-tiltrotor. Flew for a few seconds in Q_STABILIZE - everything going smoothly. I headed in to the wind and hit the FBWA mode.. started transitioning then suddenly chaos ensued, the plane was flipping upside down, spinning around etc so I disarmed and let it fall under its own weight. broke a wing and some 3d printed parts but it’s repairable.
I did however get a log recording (link above)… there are two ‘Internal Errors’: 0x40000000 and 0x 40000200 just after the transition. I’m wondering if this crash was caused by something more serious than poor config?
Grateful if someone could have a look at the log and see what went wrong, thanks.
Just a thought - how should throttle be controlled through the transition? If the motors are throttling up before the tilt servos have reached the right position, the aircraft could flip over. I don’t recall seeing a parameter which controlls this.
I assume that the directions of the servo deflections of the control surfaces for roll and pitch are reversed in the FBWA.
Check that the deflection directions in FBWA mode are as described in the WIKI:
Rolf, I was quite confident that the control surfaces were reacting correctly but I’ve just re-checked to make sure. The control surfaces are definietly deflecting correctly in FBWA.
However, while carrying out this check I noticed that the rear tilt servos now go to full 90 degree vertical, and stay there, whichever mode I am in (they should rest at 80 degrees to allow for yaw - this was definitely working in the first stage of the flight). Note, the servos seem to be orking fine, but Ardupilot is commanding them to full deflection.
However, while carrying out this check I noticed that the rear tilt servos now go to full 90 degree vertical, and stay there, whichever mode I am in (they should rest at 80 degrees to allow for yaw - this was definitely working in the first stage of the flight). Note, the servos seem to be orking fine, but Ardupilot is commanding them to full deflection.
On closer inspection, the rear tilt servos seem to have jumped a few teeth. Pretty sure this was due to them hitting the ground first (no damage to props though) rather than being the cause of the failure - still, I may be on the lookout for alu servo horns for emax mg micro servos for the repair work.
Trying to get my head around this. So you have all 4 motors tilting? Something seems odd, usually after transition the motor power will drop off to a small fraction of the hover power but it looks like all 4 of yours go to high power causing quite a significant voltage drop but probably not enough for a brown-out.
Yes, all 4 motors are tilting; the rear are pusher props, the front are pullers, so when transitioning the thrust axis of the front and rear props should remain nearly parellel and give a nice smooth transition.
Another thing to mention is that the level flight trim in Q_ modes is 10 degrees up from actual level which is used in fixed wing modes - so during the transition the pitch offest should change, from approx -10 (full forward tilt in Q_STABILIZE - plane is actually level) to zero (level flight).
Looking at the logs, the transition was looking OK (aside from the excessive throttle) up to 5:31.7 where something goes wrong. A yaw reset is triggered, EKF lane switching happening etc.
You have 4 tilt servos driven off 2 servo outputs or 2 servos and mechanical linkage to 4 motors? The way it suddenly yaws and spins on axis looks a bit like a problem in the motor angles so perhaps mechanical?
Also you 10 degree pitch thing is odd. Im not sure how ardupilot will cope with a -10 pitch for cruise. But the body is level when the imu says -10?
@Vince_Hogg No it is 4 motors 4 outputs (C3, 4, 5, 6) with 4 tilt servos, each with its own output (C7, 8, 9, 10). No problems at all with this in Q_ modes as it uses tilt-vectored yaw, as well as Q_FWD to great effect. The pitch offset is set using Q_TRIM_PITCH and works correctly when changing between Q_ and fixed wing modes on the ground.
@tridge here’s some photos; it’s not a particularly novel set up:
I was hoping to run a SITL simulation of the transition before trying it in the real world, however couldn’t see a model for quad-tilt rotors (I don’t know how to create one). The standard 4+1 quadplane simulation works fine, however the FireFly6 simlation (AFAIK the only tiltrotor model on the standard list of sims) has a very similar problem when it comes to transition! (ie. flips over).
UPDATE:
I’ve almost completed the repair work; many of the 3d printed parts (motor booms, landing leg mountings) have been repaced, whether they needed to be or not. I made some minor improvements to the printing process for extra strength and rigidity, and focused on getting everything really well alligned this time around. However, I’ve not made any substantial changes to the way the plane will operate.
I flashed the latest 4.6.0 beta (5), and will soon do a test flight in Q_ modes to check all still OK.
Before I try a transition to fixed wing again, I need to test this config in a simulation. Would anyone happen to know of a ready made “quad-tilt simulation” which I could use? Many Thanks
Upon running through mode changes with the plane plugged in to Mission Planner, I noticed some strange behaviour:
Going from Q_LOITER to FBWA or FBWB, the pitch offset changes by -10 degrees, as expected.
However, when going from Q_HOVER or Q_STABILIZE to FBWA or FBWB, the pitch offset does not change! Could this relate to the crash? It sounds like a bug - has this already been raised as an issue in 4.6.0 Beta 5?
