Thx for your reply. Before I ever tried it outside, I tethered the plane top and bottom and flew it in Q-Stabil. It was fine, I could control ele on forward and back, Ail for motor roll diff thrust and rud for yaw. All the controls were correct orientation for this mode so I assumed it would also be ok in Q-hover.
My intention was to only hover a few metres up then land but after throttling up to just over mid, it did a successful takeoff but kept climbing and drifting back past me, so I just put in forward ele to bring it forward, but to me it looked like it jumped into forward flight. I thought this only happened when FBW was selected. With the plane shooting off un controlled I tried to find FBW button but his pure manual. This is more then likely why the plane just stalled and went into a spin.
I am rebuilding now and will recheck all the settings as you suggest, but I am positive they are all correct.
what about CG? was it setup for normal forward flight? if the CG was very nose heavy and the elevons not effective enough, it may be that it just could not overcome the nose heaviness and pull it back upright…like slow flying with full elevator in a nose high position…without vectored thrust (my only extensive flying experience) I would imagine that having the CG just at the slightly positive stability point would be pretty crucial in order to have elevon effectiveness at slow speeds or prop wash impingement…
Well. Balance point was near on the money and the Elevons are a bit barn door ish and the motors are 35x36 with 11x5 props so I don’t think that was or is the problem. What I don’t understand is; did the plane enter normal flight on its own or do the logs indicate the model was still in Q Hover and it just tried to fly in that mode. Rebuild is well on the way, then I can do some more tethered tests .
Still in QHOVER trying to pull the nose up to vertical(cant really tell actual servo direction) but was hard one way when nose was down, and later the other after it flopped over on its back before you hit manual
Any way, thanks very much for you insight and observations. Plane finished, will try tethered next time until I am happy to let it off the lead.
Reading about Stab and hover, I realise that q-hover is not gps controlled, you still need to input flight corrections. This may have been my undoing, should have set up Q-loiter, oops
Hi Tridge,
Is the fix reflected in the latest plane firmware? The firmware that is used when you download it for SITL in other words?
yes, it is in the firmware you get when you start SITL in MissionPlanner
From the log it is evident that the VTOL has exceeded the maximum angle declared (30) in QHover.
This usually denotes a CG problem, coupled with relative aileron control (poor hardware excursion?) and bad copter tuning.
From what I see you have left many parameters at default that should be calibrated instead.
If properly tuned with a full pitch command this cannot happen.
The maneuver that perhaps would have helped would have been to go into QStabilize and put the throttle to maximum until the attitude recovery, switch to Manual Plane imho is not the best choice.
I have taken your observations onboard. Yes, the CG was way out, too nose heavy by a mile and the elevon travel was insufficient. I am tweaking the parameters as well. Now the plane is rebuilt I have moved the battery rearwards behind the main spar and mounted the pixhawk toward the front. According to my CG calc I need the balance point to be 6 inches back from leading edge, it is right on the money now. I increased roll Q-roll, but the throttle hunted, so lowered it back till a quick flick on the roll did not cause the oscillations.
The Q-pitch setting has been increased and the elevons are very reactive so may need a test to make sure I have not made them too sensitive.
I have a couple of questions about a non vectored tailsitter : In Q-Hover and Q-loiter should the max lean angle be no more than 30 degrees off vertical ?: Secondly, if in FBWA and Q-RTL is selected, will the model stay in plane mode till near the rtl point, then transition to vtol and land, or will the model enter vtol mode from FBWA and return slowly in vtol mode, landing at the rtl point?
switching to QRTL immediately changes to VTOL mode and will move in that mode toward home and then land…
switching to RTL mode can be made to travel in fixed mode til within RTL_RADIUS and then switch to QRTL for landing by changing Q_RTL_MODE
OK, I thought that may be the case, thx for the clarification.
Where can i find this model to do some testing with real flight?
In Wingtra One, from internet images I observed the existence of an optical flow sensor or lidar
, which I believe to be of assistance in taking off and landing.
Would anyone know how to implement this functionality?
if you have a downward pointing rangefinder enabled, quadplane will use it in the final landing stage for QLAND to measure altitude instead of the baro and for alt measurements while in QHOVER if within its range…just read the wiki on setting up a rangefinder for downward facing, just like on copters…if enabled, QuadPlane will use it
Is it possible to use precision landing with tailsitters?
great, I’ll read more …
Has anyone tried to use precision landing with tailsitters?
precision landing is not available on Quadplane.
Thanks peter, are there any plans to add it?