ARWing Pro, the road to VTOL

Поділитися
Вставка
  • Опубліковано 6 лис 2024

КОМЕНТАРІ • 21

  • @RavenFPV_NL
    @RavenFPV_NL Рік тому

    Nice!

  • @AerialWaviator
    @AerialWaviator Рік тому +1

    You didn't mention what flight controller software, or how configured for first wing maiden. As you did with the VTOL setup having an altitude hold, or loiter and RTH really help if encounter an unexpected situation.
    The WTF moment near the end was likely caused by trying to transition to VTOL while flying downwind. It's always best if do transitions into the wind. Better at altitude first.
    With tailgater VTOL like you built it's a good idea to decend slowly; too fast and the upward wind stream will cancel out wind from the propeller airstream over the control surfaces.
    Overall, are making good progress for first time. A wing is more challenging as centre of gravity range is more more narrow than a conventional (tail attached) wing.

    • @dabit4840
      @dabit4840  Рік тому +1

      I used Arduplane 4.3.1. This maiden was close to default settings; I basically only enabled the VTOL functionality with the correct frame class and motor/servo assignments and setup the attached devices such as the MavLink serial port to the Pi, DShot for the motors, ELRS RX, etcetera.
      I was not yet transitioning back to VTOL when WTF happened.
      I still think the WTF moment at the end was just too little airspeed. The ARWing tends to stall a wing first, and it happened in a (too) slow turn.
      It looks terrible on image, but not much was broken. Some previously glued pieces of foam came loose and had to be glued back together, and the electronics ripped loose of the double-sided tape.
      After this video I did a few more flights, and the VTOL lacks serious authority. Even a few meters/sec of wind prevents the plane from staying in place while hovering. I loaded it up to maximum takeoff weight to create a smuch airflow over the elevons as possible and enlarged the elevons, with little success. I also failed to do the transition from forward flight back to
      Currently I mounted two smaller (2207) motors on the stiffened elevons and increased servo size so I have vectored thrust. Testhover in the garage was not a complete failure, but to call it succesfull is a stretch. Lots of oscillations until the plane is in the air, not very stable positioning. Definitely needs some PID adjustments. And probably some balsa elevons and upgraded hinges too.
      Waiting for better weather so I can give it more space than 1x1x1 meter to do it's thing.

    • @AerialWaviator
      @AerialWaviator Рік тому

      @@dabit4840 One technique is have the two wing motors on servos, then will have vector thrust for yaw, pitch and roll .. independent of control surfaces. Vibrations have potential for causing problems, as can effect IMU/gyro. Motors on elevons is not a setup I'm familiar.
      Re: WTF. It might have been case of just flying slow downwind and turning at slower airspeed as you suggest. Flying downwind can give impression of plane flying faster than it's true airspeed.
      Good luck with weather windows. You seem to be making progress. :) Arduplane/copter forum is a great resource.

    • @dabit4840
      @dabit4840  Рік тому +1

      @@AerialWaviator : I am aware of the tilt-rotor tri/quad configurations. Would be the easy way out, but the current purpose of all this fooling around is to explore a bit more.
      For anything serious a quadplane is probably the best option since both vertical and horizontal flight can be optimized separately. With tail/bellysitters or tiltrotor there is a compromise to be made; a motor+propeller combination that works well in hover is not very efficient during forward flight and vice versa.
      However, I think there is beauty in simplicity. The beauty of a VTOL tailsitter is that it only requires an extra motor compared to a regular plane. Actually not even 'requires', more 'desires', and you can 'pay' for that extra motor by getting rid of the rudder if you had one.
      This is why I want to try mounting the motors on the elevons. It keeps the simplicity.
      35 grams of motor+prop gives a lot of thrust nowadays, no issues there. Motor+prop acting as a gyroscope is an issue, stiffness of the wing trailing edge is an issue too, especially on my butchered ARWing test platform.
      If it works, great! If it does not, I can learn why nobody does this, which is also useful.

    • @AerialWaviator
      @AerialWaviator Рік тому

      @@dabit4840 Have fun experimenting. Makes for great learning opportunities.
      My next experiment is just 2 motors (and a servo to pitch each motor), no control surfaces.
      Like you I enjoy the process of learning. :)

  • @TheHillsHaveFPV
    @TheHillsHaveFPV 4 місяці тому

    needs a camera tilt for landing/takeoff

  • @BrainDeadEngineering
    @BrainDeadEngineering Рік тому

    So when you click transition, imagine its on a switch ? what happens ? the pid driver is shut down? I look up to you ! I feel same way, hard to find runways for planes ! VTOLS is the way !

    • @dabit4840
      @dabit4840  Рік тому

      If you switch from a VTOL mode to a fixed-wing mode the plane transitions from vertical to horizontal and continues as a regular plane.
      I am still not able to reliably switch from forward flight back to vertical. Sometimes it works, often it does not. Wind had a lot to do with that

  • @vtoljet
    @vtoljet Рік тому

    Mine failed in the flight transition and was wrecked
    nice stable fast plane

    • @dabit4840
      @dabit4840  Рік тому

      Sorry to hear about the wrecked plane.
      I was a bit surprised that it went so smoothly. This was on Arduplane defaults. I can tell you I shitted my pants the moment I pushed the 'RTL' button.

  • @gpopsk
    @gpopsk Рік тому

    what speed did it had when you transitioned back?

    • @dabit4840
      @dabit4840  Рік тому

      I did not transition back. I managed to tipstall the wing in a tight slow turn before I could start the backtransition.

    • @gpopsk
      @gpopsk Рік тому

      @@dabit4840 ok, but what speed did you had before doing this manuavers? do you have a log file? I am interesting in making this work

  • @farkhodkhikmatov4738
    @farkhodkhikmatov4738 Рік тому

    Do you pay for 4g telemetry? Can you guide us how to set this up?

    • @dabit4840
      @dabit4840  Рік тому

      Yes, you do pay for the mobile data use.
      However, MAVlink telemetry does not use much data. Total data usage during the various field tests is a measly 15MB for now. I paid 18 euros for 5000MB prepaid data. Thus, I used up for 5 cents on data. Or to put it differently: I can fly all year for 18 euros worth of data. I could even have bought a smaller data bundle, but once I start to push camera images through the data usage will increase.
      Add in a pre-used DW5811e cellular modem which I bought 3pcs for 20 euros and a Pi Zero which I got for free but which is

    • @farkhodkhikmatov4738
      @farkhodkhikmatov4738 Рік тому

      @@dabit4840 thank you for the detailed message 😊! I will try my best to set up. I will probably use raspberry pi since I have them laying around and weight does not matter. Also why do you say you wanna change to a jettison nano? Does it have a lot more processing power? It's very big ang bulky compared to this, or am I wrong?
      I am actually building a very big tail sitter, hopefully it will be very stable and easy to fly. I set up my model in realflight combined with sitl seems like it works 🤷🏻.

    • @dabit4840
      @dabit4840  Рік тому

      @@farkhodkhikmatov4738 : In the future I might want to use an extra camera connected to the single board computer to provide an FPV image when out of range of the regular (currently analog) FPV system. Or just for spectators; if they connect VLC on their phones to an RTSP server relaying the FPV feed they have image. Not really because I need it (in The Netherlands I am not allowed to fly BVLOS anyway and spotters at some distance are not allowed either), but just because I think it is fun to tinker with.
      In that case, the Jetson Nano is capable of h.265 (HEVC) video compression, the Pi is not. Saves roughly half the mobile data. The Nano itself is not big BTW; half the size of a regular-sized Pi. It does need a carrier though, but I am capable of building one that would fit both the Jetson and an M.2 LTE modem in about the space a regular Pi would use.
      At this moment I am having other issues. I cannot replicate the vertical takeoff in this video anymore; plane rises vertically, flips over backwards, flies upside down for a while, and crashes. After 10 crashes or so one of the motor mounts gave up. Bummer. Time to figure out how to analyse the logs. As I said in the video; Arduplane is quite new to me but I am experienced witth the Betaflight/iNav blackbox logs.
      I am interested in using Realflight or maybe X-plane as SITL target too. Quite a chunk of money, but if it prevents crashes it might be worth it.
      Do you know how hard it is to replicate a real-world plane with reasonable accuracy in either of those? I can model them in CAD, but I guess that's just the easy part?

  • @dragobert
    @dragobert Рік тому

    Hi! Can you please share your parameter tree? =)

    • @dabit4840
      @dabit4840  Рік тому +1

      Sorry, missed this comment.
      Here you are: www.icecoldcomputing.com/misc/ARWing_VTOL_after_vtoltune_autotune_26nov2022.param

    • @dragobert
      @dragobert Рік тому

      @@dabit4840 Thank you😊