hi Otto, sorry, I’d misunderstood your request. I’ve now tried it as you requested, and get this:
I think that means we need a texture?
I think that means we need a texture?
Hello Tridge Sir,
I am going to set maximum spin limit for copter modes by Q_M_SPIN_MAX but dont see any spin change.
Please guide how I can set maximum spin limit for copter mode.
Hello Tridge Sir, how we can set Climb Sensitive of copter modes through param of that tailsitter.
Have been testing Beta 4 with my tailsitter yesterday. Great improvements. Transition is much smoother, backtransition into QHover now working for me.
Flight #4 I had a crash at backtansition @1:30 , dunno why. Think he set desalt to low at transition and though cut throttle for a short time. after this I set min_thr to 0.3
Flight #6 It was going sideways in QHover very fast to test stability.
Also two BT in QHover. https://www.youtube.com/watch?v=-fbucWjaaeA
Flight #7 with eleven backtransitions in QHover. They are marked in the description: https://www.youtube.com/watch?v=nd9cnBXvj_Y
Logs of three mentioned flights: https://www.dropbox.com/sh/lolc2e70zu852zd/AAAzljaqoVcr2YpK8U2axSnSa?dl=0
1 Like
i just uploaded @lorbass originale file to Fusion360 applied polystyrene foam as material and appearance property saved and then exported via A360 to FDX file. it did change the appearance of the wing section in FDX viewer but I have no idea what I am doing with these animated mesh file types.
Wing (2).3ds (23.2 KB)
same thing; it’s not a 3ds so change ext to fbx
@Coby thanks - @lorbass and I have made some more progress on this together, and I’ve also made some progress with @Leonardthall on a skywalker X8 model. I’m thinking we should probably create a separate discussion about model creation for RealFlight, and also create a github repo of models and components (as well as parameter sets)
I’m hoping we’ll end up with lots of models eventually, and for people to be able to modify models to create new ones fairly easily.
I’ve also managed to edit a model in blender, which I’m delighted with as blender is a free tool, so accessible to anyone.
Cheers, Tridge
2 Likes
Does the parameter Q_M_PWM_TYPE (sets normal/oneshot/oneshot125/brushed ESC-type) also set the type of output for the two motors of a twin-tailsitter ?
Regards Rolf
@tridge I would like to give some suggestions regarding the backtransition Code based on my flights few posts ago:
- plane should continue with the rate it rotates in the first half of the transition till the end of transition
- Hover thr changes over flight, should be scaled with Batt voltage
- PIDs should be linear blended from Plane to Copter while transitioning
- PIDs should somehow be scaled like this also for fast hover flight to (maybe) prevent instability while fast hovers (like mine in lasts Post Flight #6).
1 Like
no, it doesn’t. That parameter only affects the MotorN functions, which are not used for tailsitters. You can set the update rate of the motors using Q_RC_SPEED, but you can’t change to oneshot
yes, although it is a bit difficult to achieve with the current code architecture
yes, I agree, and that shouldn’t be too hard to add
ideally we’d have just one set of PIDs for both, and teach the copter PIDs about airspeed scaling. Blending PIDs may work, but I’m not sure it will, especially with fighting integrators
yes, for that we need to teach the copter PIDs about estimated airspeed. Leonard and I have discussed that, but it isn’t implemented yet
These are all good suggestions, but I don’t think I’ll get any of them done for 3.8.3. For future releases I am certainly looking at these changes.
Cheers, Tridge
1 Like
Thanks for the feedback.
May I ask whats the math for scaling the PIDs as you implemented lately?
My PIDs were working fine before the patch with scaling. I now want to recalculate the PIDs that i have to set that it behaves in hover like before the patch.
EG: P = 0.3, I = 0.025 and hover thr is 0.6, whats PIDs do i have to set that it behaves as before the patch?
It scales using throttle level. So if you are at hover throttle then no scaling applies. If your throttle is at 2x the hover throttle then half the surface movement is used. If your throttle is at 0.5 the hover throttle then 2x the surface movement is used.
don’t change the PIDs, but do make sure your hover throttle is set correctly
Cheers, Tridge
Plane 3.8.3 is now released - thanks to everyone who helped test it on tailsitters!
There is still a lot we can do to improve tailsitters, but I think this release is a big step forward
Hello everybody, I have a question.
What is your autopilot orientation and how the transition code works for you?
@tridge Does the transition code takes into account the autopilot board orientation? I’ve tried transition on the ground (holding the plane in hand and switching from Qhover to manual) and the plane tried to pitch up (instead of pitch down). My board orientation is roll 180°, so I think, that it is connected. I will keep on testing.
The orientation of both pixracers in our tailsitters (1 vectored bellyitter, 1 conventional 2-mot-tailsitter) is AHRS_ORIENTATION 16 (ROLL 90) respectively AHRS_ORIENTATION 2 (YAW 90) and all works as expected.
Are you sure that you have set AHRS_ORIENTATION correct ?
Hi Tridge.
Thx for the explanation. I did as you said.
From my log the hover thr is at 0.7. I had set it to 0.6 before.
In both settings the control surfaces are shaking at hover. Before the patch they were not shaking at hover.
strange! Can you post a before and after log to look at?
Thanl you for your reply. I don’t know, what I have seen last time. I tested today, and transition from Qhover to FBWA and back works perfectly. So no problem at all:slight_smile:
before: https://www.dropbox.com/s/4cwr0qptrpdpp0n/00000010.BIN?dl=0
After: Last log of them: https://www.dropbox.com/sh/lolc2e70zu852zd/AAAzljaqoVcr2YpK8U2axSnSa?dl=0
Hi everyone,
I want to first thanks to all developers for this fantastic codes. And this forum give me a lot of information during my built. I have just switched from quadplane configuration to conventional tailsitter one. The project is supposed to become commercial product in the future. However, at this moment, I must first get to know all relevant tailsitter design principles and software. I want to share our first transition test flight with Arduplane 3.8.2 with mostly default parameters. (But I have just seen that 3.8.3 has been released today, will try it later for smoother transition)
From flight log, we’ve found that the control surface size is too small causing the aircraft to tumble in pitch many times especially during lowering throttle while hover and windy condition. However, I think it could be able to stabilize itself back every time due to excessive thrust to weight ratio of this vehicle, more than 2. I still use EKF2 (Not EKF3). Airspeed sensor doesn’t work properly in FBWA too. Will check it later
Here is the link for dataflashlog (incase you guys want to check it):
Regards,
RW