This morning we had our third flight. Both transitions have worked great.
Log: https://www.magentacloud.de/share/e.2saescpx
The PIDs still have to be corrected.
https://vimeo.com/211097038
rolf
2 Likes
Hi,
what type of servo (standard or retract) do you use and how strong is it ? I would like to try a tiltrotor with a Tmotor MN4006 and a 15" prop and I am trying to get an idea what I need as servo. Any suggestions welcome.
we use this D-Power CDS-360BB MG Servos for the fronttiltmechanismus and the yaw-tilt also.
Hi Rolf,
Thank you very much for the info. Your servo is rated 60-70Ncm, I donāt know the size of your prop, do you think a 100Ncm servo will be good enough for the 15" prop ?
12" props. The motor (DYS BE2814 700KV) weighs 105 grams.
Hi Rolf
Can you please share your parameter files
thanks 
looking good! congratulations!
Yesterday we had successfully adjusted the PIDs of the pitch axis.
A little more authority around the YAW axis, we will still adjust.
With the new 40/80 C 4700 4s battery was finally enough power in hover.
So the transitions were both great.
Regards Rolf
2 Likes
The transitions look great, Rolf!
Have you detailed your tilt mechanism anywhere?
that is looking pretty smooth now. Awesome!
Hello,
I am the partner of Rolf at the project VTOL.
The tilt construction is made of PA (3D print)
This design has the disadvantage that at the end points (+/- 45 degrees)
only 70% of the torque of the servo are available. In addition, the precision of the push rod is only moderate. The accuracy of the transmission of our servos is also not very good. Therefore, there are motions of the motor. But in works so well enough.
A better solution would be the position of the servo directly on the axis of the holder for the motor. And a more precise (expensive) servo.
pictures of the design: red = variant
Regards Walter
Hi Walter, and welcome!
I really appreciate the work you and Rolf have done in testing the tilt-quadplane code. When I originally added support I hoped someone like you and Rolf would come along to test it as I didnāt have the right sort of airframe myself. It worked in the simulator, but real-world testing is so important.
Would you and Rolf be interested in documenting what youāve done as a build log on the ArduPilot wiki? Or just writing up a page with advice on building and configuring tilt quadplanes?
We have the basic docs here:
http://ardupilot.org/plane/docs/quadplane-support.html
but a lot has changed, and there is no information on tilt support. I could give you write access to the wiki if you have time to help with documentation.
Iād also like to use one of your great flight videos in the announcement of the 3.8.0 release if that is OK with you.
Cheers, Tridge
Hi Tridge,
Of course you can use the films in the announcement.
We can write a buildlog (Walter) and a first draft (me) of a wikipage proposal and send it to you for proofreading and further improvements. But unfortunately it will takes some time. You will have released 3.8 sooner.
Regards Rolf
Hi Tridge,
we still have some questions on how the forwardtransition works on tiltrotor with normal tiltservos (Q_TILT_TYPE = 0).
The charts comes from the logfile of forwardtranistion from flight above on Apr. 8 with Arduplane V3.8 Obeta3 .Logfile: https://www.magentacloud.de/share/3oqfc0hrfz#$/
Some settings:
RCOU.C5/C6 front right/left motor
RCOU.C8 rear motor
RCOU.C12 tilt servo front motors
RCIN.C1 = ailron stick input
RCIN.C3 = throttle stick input
ARSPD_FBW_MIN 8
Q_TILT_MAX 55
Q_TILT_RATE_DN 12
Q_TILT_RATE_UP 80
Q_TILT_TYPE 0
In the diagrams are marked four junctures A, B, C, D.
After switching from QHOVER to FBWA, the tilt servo runs up to 55 degrees (Q_TILTMAX 55, point A in time). The program then waits until the Airspeed has reached the value of ARSPD_FBW_MIN (8, point B in time). Until then, the forwardtransition runs perfectly without losing even a millimeter of height. But between points B and D in time follow three seconds of shock stasis, in which the rear engine is switched off from C but one can still not adjust full throttle.
In this context, some questions:
What is the meaning of Q_TRANSITION_MS (1000) at tiltrotor with normal servo (time between point B and C ?)
Is full tilting really triggered by Airspeed or by GPS-Speed (Groundspeed)?
Could it be selected, wether triggered by GPS or Airspeed?
Could not a separate new parameter be used for the limit speed instead of ARSPD_FBW_MIN?
Could you set a third speed parameter for the tilt servo that allows between B to C to bring the rotors to the final position more quickly ?
Regards Rolf
Hi @Rolf,
I had a look at the log, and I think I see what the issue is. Once we get past the transition time the throttle is constrained to rise at the same rate that the tilt increases.
The simplest fix is to increase the Q_TRANSITION_MS so that the rear motor stays active for longer, and the tilt has a chance to get further forward before it shuts down the rear motor.
We can also look at changing the strategy it uses.
It the the time after we reach the transition airspeed before it shuts down the VTOL control. During this period the VTOL throttle is slowly lowered. That lowering makes sense for non-tilt quadplanes, but I think its counter-productive for tilt-quadplanes. Iāll think about whether it should be disabled.
airspeed, but if an airspeed sensor isnāt available then it estimates airspeed using ground speed and the wind estimate.
I think it doesnāt really make sense to use groundspeed if airspeed is available
we could, but I donāt think that will really help solve the issue
yes, we could if that turns out to be the best approach
Cheers, Tridge
Thanks for looking through. Walter and my opinion is also to use a third parameter for the tiltservospeed after reaching ARSPD_FBW_MIN would be the best approach.
I looked at this again today, and decided that instead of a new parameter that it would be best to just make it always use a minimum of 90 degrees/second for that stage of the tilt. Iāve created a PR here:
The reasoning is that once weāve shutdown pitch stability control by shutting off the rear motor that we want the motor tilt to be all the way forward as quickly as we can. The 90 dps limit is the same as we use for the transition to manual mode, and is only there to be nice to the tilt servo so it doesnāt try to move too quickly.
Do you want me to create a prebuilt fw for you with this change?
Cheers, Tridge
@ tridge:
Excellent solution. We canāt flight until thursday due to bad weather(forecast). If so far the changes will be in master, we can also download it.
Rolf
@ Tridge
As the first phase of development for a tiltrotor tricopter plane weāre making, Iām currently building a tilt-rotor tricopter on a T frame, with the two tilting forward motors mounted on the top of the T, and a larger fixed rear motor and prop mounted on the bottom of the T frame. In order to test this version of the code without wings, Iām wondering what would be the best setup to ensure that the rear hover motor remains on to produce lift at all times, and if that is at all possible fly that at the moment?
The difference being that when this version of aircraft flies forwards it will not get any lift from wings (as there are none yet and I would only like to use the smallest highest aspect ration wings possible to sustain cruise) and the rear motor must maintain some rotation to produce enough lift from the rotor both when it is stationary in hover and whilst it is moving forwards (including for pitch control).
Currently with tilt rotor support as Rolf is using in the above video, Iām assuming the only motor that would be activated by going under Q_Assist_Speed would be the fixed mounted tail rotor? The forward tilt motors would need to be rotated upwards to maintain lift whilst slowing down. However, as would be the case for the tiltrotor tricopter explained above with the forward motors tilted for forward thrust, I would need the rear motor to stay on even in forward flight, and also to maintain pitch control even without wings. Is this currently possible with arduplane and if so what would parameter setup could possibly work?
Regards
Sam