"Box Kite" Copter Tailsitter Build instructions

Since I got one request for this and created instructions, I though I would share with all.

Foamboard “Box Kite” Copter Tailsitter build guide
utlra cheap, very relaxing “wright flyer” style copter tailsitter

Parts List:
3- 20inch x 30inch Foamboard sheets
4- 1804 2300-2500kv motors and 8- 2mmx6mm bolts for mounting
4- 4x4.5x3 props or 5x4.5x2
1- 4 in 1 20A ESC
1- 3s, 1000-1300mah battery
2- 9g servos, preferably metal gear train
1- OmnibusF4Pro autopilot running ArduPlane firmware
1- BN880 GPS/compass (I used a BN220 + discrete compass)
1- RC receiver of your choice, with single wire protocol output (SBUS,CPPM)
{I used FrSky r-xsr with Fport to get telemetry back to TX}
8- bamboo cooking skewers for elevon push rods and cross bracing
6- Dubro 1/4A control horns
6- Dubro micro EZ connectors

  • .042 wire for pushrods and elevon link rod ends
  • fiber re-enforced UV resistant packing tape (for hinges)
  • Hot Glue and spray contact cement
    4- motor mounts, either 3D printed or likewise fabricated from lite plywood (STL
    files attached)
  • optional, 3mm corex used for center panel

Construction should be obvious from photos:
1.cut two sheets in half in the 20 inch dimension…you now have 4 10inch x
30inch panels… now cut one of the 10" wide sheets in half and trim 3mm off each
end of the 5inch KMF2 doublers you just cut for the side pieces to fit…these are
the top airfoil sheets (KMF2) and strength doublers
2.glue the shorter sheets to the longer sheets, leading edges flush, these are the
wings…cut out 1.5 inch elevons, 45 degree angle, flip over elevons and trim to fit
flush again,add tape hinges in normal foamy plane style
3.from the other 10" sheet, cut three 10 x10 inch vertical members for ends and
center and cut remaining full sheet for an extended center section 10 inch tall and
13 long with tapered nose(3" nose for extending battery location)…and use
remaining to double the nose portion…laminate the center sections together…cut
recess to fit KMF2 top airfoil 5inch sections…alternatively, you can cut a single
3mm corex panel for the center section (shown in photos)…its more durable and easier to attach
velcro and autpilot,etc with double sided foam mounting tape…
4.Glue all the vertical sections together, add bamboo cross braces, tape leading
edge and ground contact points (not elevons they should be trimmed a little to
avoid contact) with tape for durability
5.Make pushrods for top to bottom elevons
6.Add equipment and setup ArduPilot, be sure to check motor orders and
rotations, just like a copter…my param file is attached
7.go fly! pick a zero wind day…you can also gently hover in your garage,if
careful…definitely in your yard if you have a 10mx10m space…
note, you may have to lay it horizontal for 20 seconds if you get “DCM pitch/roll
inconsistent” prearm failures, then back vertical and it should arm now.
-elevons upper and lower are linked with bamboo link rods…music wire CA glued
to ends after wrapping with thread or heat shrink…an old foamy plane method of making light
strong pushrods
-try to mount everything as far forward as possible as shown in the photos. If you
do, then the battery will only have to be a little beyond the leading edge to
-joints are butt hot-glued but KMF2 doubler is shortened on each side to allow end
joints to be fitted, as is center section fitted to KMF2 step for flush joint it entire
-kmf2 doubler attached with contact cement
-CG is at ~15% of chord, I printed a cg hanger to help determine that…finger tips
under top wings works also
STL Files:
Motor Mount 3D design dragonfly mount | Tinkercad

Parm diff from default Plane:
(you may want to setup OSD to your liking if you FPV this…but choose light camera and vtx)



Nice job, Henry! A great example of thinking “outside the box”. :wink:



We might want to replace one of the videos on that page with this one
which is a bit more informative about the frame. I think it’s an
analogous frame to the X-Frame biwing?

1 Like

what telemetry module/display is that used in the vid?

1 Like

Hi Henry
how can i adjust the pixhawk in case there are no flow on aerodynamics surfaces in the aspect of yow
I’m building model airplane like H-WING VTOL

if you have no flying surfaces on a Copter tailsitter, you want it to be in QASSIST mode all the time…Q_OPTIONS bit 7 I belive…@iampete can confirm

1 Like

Does the latest 4.1 firmware have no servo control?Will it be more difficult to fly without a steering gear?

I am not sure what you are asking…4.1 definitely controls servos and motors according to what SERVOx_FUNCTION is selected for an output…
what do you mean by steering gear? that is not a term used in ArduPilot

I think KMF2 is referring to the “stepped” airfoil Kline-Fogleman KFm2 described here: https://en.wikipedia.org/wiki/Kline–Fogleman_airfoil

If that was a question, then yes it is

Thanks for the confirmation and I should have phrased that as a question.
Are you using just 2 elevon servos or is there a total of 4 as in the parts list?

Look at the photos at the top… Two servos, one wings set of elevons are linked with bamboo control rods to those driven by the servos on the other wing, to save weight

1 Like

You list 4 servos, not 2, in the parts list. So the parts list should be changed to show 2 servos, is that correct?

Thanks for the nice simple design to experiment with tailsitter airframes. I’m building one now.

1 Like

done…thanks for catching

Hi Henry,

I have a build of your BoxKite VTOL design that I’m doing some initial testing on (too windy to do hover tests outside right now). It looks to me, in the video you posted, that the control surfaces are moving in VTOL modes but it’s hard to tell for sure. Is there some setting that determines whether the control surfaces are used during VTOL modes for a tailsitter? Here’s a video of my “hands-on” testing:

BoxKite VTOL indoor testing

Pitch and Roll authority seem fine but I think I may need to bump up the Q_A_RAT_YAW_P somewhat.

yeah the control surfaces move in VTOL mode…and with your small surfaces YAW in VTOL will be sluggish…will investigate if there is a setting for surface gain in VTOL only…dont think there is one…but could be wrong…you can disable the surfaces in FW mode in a Copter tailsitter with Q_OPTION bit 8, but dont think there is anything for in VTOL…you really need the surfaces assisting in VTOL since the CG is counter-acting the motors pitch and roll torque a good bit due to pendulum effect…

edit: confirmed no way to stop surfaces in VTOL, but Pete Hall has a PR pending Plane: tailsitter: add gains to scale control surface vs motors by IamPete1 · Pull Request #19463 · ArduPilot/ardupilot · GitHub

I was able to some short hover hops tonight once the wind calmed down. It went pretty well. The yaw authority is actually quite good although the the control surfaces don’t seem to be contributing to that. Here’s a link for the very short hover hops:

Earlier in the same video I bumped up the throttle, picked it up, and yaw’ed it back and forth but there doesn’t seem to be any control surface movement then either. This was all in QSTABILIZE mode.

The control surfaces do move just fine in manual mode.

I haven’t tried changing any of the PID parameters yet. They’re just set the way you recommended them above.

Anyway, it’s a nice VTOL test vehicle. Thanks you for sharing it. I’ll keep you posted as I do further testing.


Hi Henry
I apologize in advance for my English
I am currently building a VTOL Tailsitter quadplane with no control surfaces. It looks like X-wing from “Star Wars”. And I have nice flight in QSTABILIZE mode, I have nice transitions to FBVA. But his FBVA is a problem, he starts to swing sinusoidally, do somersaults and fall, I have been selecting the parameters for several months. At first I thought it was the PID settings, but by making them very high or very low, I got the same result, so I started going through other parameters, but also with no result. If you can help, I would greatly appreciate it. If necessary, I can provide flight logs. I fly on Arduplane firmware 4.2.0

I can try to look at a log when I get some time next week…no promises…post it