Dual Tilt-Wing issue

ArduPlane VTOL Plane
1

Hello:

Need a little hand holding here.

I’m building a quad tilt-wing. I’ve flown the frame as a quad, using Copter. I have the latest Plane installed now and the Q functions enabled. What I’m running into is kinda strange.

First thing is I don’t have the wings installed due a 3D printer malfunction, so waiting to get that back to fabricate the wing sections. I haven’t flown it in any of the Q modes yet.

My present flight modes, on 2 separate Taranis switches using “replace”, are:
QLoiter
QHover
FBWA
QStabilize
Manual
QRTL

The tilt servo is on Pixhawk 2.1 Servo9.

On the bench, no props and disarmed, I can get the motors to tilt from vertical to horizontal in both Manual and FBWA. If I arm the frame, Manual gets me full down tilt but FBWA does not. FBWA will twitch the tilt servo but nothing more. The odd thing is if I add THR, the tilt servo starts following it, so that at full THR, the motors are horizontal. Let up on the THR and the motors move back vertical.

The other oddity here is if I flip in and out of Manual while the other Taranis switch is in FBWA, the motors will go full down and stay there.

I’ve read about and changed the Tilt_Assist and Q_Tilt-Max functions, etc. along with the “Arspd” parameters and nothing seems to get the motors to tilt when armed and in FBWA. I’ve even used compressed air to simulate air speed thinking that param needed to see something. Nothing changes Re FBWA.

Also, being new to Arduplane in general, I am unable to get a Taranis radio off failsafe to work. I can invoke a manual F/S and the flight changes to QRTL. I’ve set the Taranis custom F/S to the same value but Mission Planner shows me the last flight mode before turning off the radio. Kinda nervous about flying in that situation.

Build:
-Frame and arms are Home Depot square 15x15mm aluminum towel bar and Actobotics U channel and bearings

  • 2.1 Pixhawk
    -Sunnysky X4108S 480KV motors
    -Hobby King CF 15x5.5 props
    -Lumenier 4 in 1 35A ESC/PDB
    -Gopro on a gimbal
    -Old DJI F550 gear struts

When I get the printer back, print up wing ribs and skin with 1/32" balsa, using the square arm tubes as the spar with additional 5x5mm CF tube as secondary. There will be a total of 8 wing sections 2 each per arm in and outside the motors. Aerodynamic controls will be ailerons on the outer front panels and elevators on the inner rear panels. Rudder mounted aft. Wing span looks like 42-48" x 2.

Any thoughts are appreciated.

2

Hello again:

OK, I worked out something that I could understand with tilting the motors/wings and got the rest of the frame put together. What I did was have manual control of the wing tilt for normal operations and give tilt control to Pixhawk in a QRTL situation.

My flight modes are Qloiter, Qhover and FBWA.

Took off in Qloiter, moved over away from the pad so we could have a better viewing angle due to sun position and switched to QHover. Started forward with wing tilt near 45deg and as the frame accelerated, I pushed in more wing tilt and the frame switched up to FBWA. The flight mode change happens when the wing tilt is past 45deg per a Taranis mix.

It flew straight for about 2 seconds, then pitched up vertically and tumbled. I moved the wing tilt back to vertical and added full power.

Kodos to the Arduplane developers - Pixhawk recovered it and I hovered home and landed from 200yds out.

The data log shows it pitching up right after the flight mode change to FBWA and I can’t find anywhere there was a control input from RCOUT, so I’m suspecting it’s a physical problem with the frame. For instance, the elevators might be on the smallish side.

Anyone out there doing a quad tilt-wing frame that has some tips?

20181116_092844-ssm
20181116_092844-ssm.jpg1328×747 299 KB

3

A flight log would certainly help anyone to see what might be happening.

I have a question:
Have you set the CofG as per a Canard?
Or as a Quadcopter?

Here is some info on Canard CofG calculations.
https://rcplanes.online/cg_canard.htm

Its just that it sounded like a bad CofG, and for an aircraft I would check that first.

4

Hey Mike:

Thanks for the reply.

Tomorrow I’ll see if I can make sense of the canard measurements and plug them into the formula and see what pops.

Having the front wing at a slightly higher AoA might be problematic.

Yes, the C.G. was calculated as a quad.

5

Hi,

The closest thing I have to your design is the Ares V-Hawk X4 but it flies stock, not APM. However, I have a few words of advice for your test flights.

First, replace the camera and gimbal with equivalent dummy weight for balance. Perhaps you just had them mounted for the photo. Second, disable your Air Speed sensor. You should not be using this sensor on initial flights. Parameters like Q_ASSIST_SPEED and Q_ASSIST_ANGLE should initially be set to 0.

It will be interesting to see the log and if you were really still in FBWA mode during the pitch up. It would also help if you could post your param file and tilt rotor/motor connections to the Pixhawk.

Cheers!

azsz2700

6

Good morning:

Mike B. - See the canard formula image. If all the calculations work out true, then I need to move the C.G. about 125mm forward. That’d be some serious re-work.

Greg:
The param file is from before I made your suggested changes181117-01.param (16.9 KB)
CANARD-FORM

20181117_091550-ssm
20181117_091550-ssm.jpg1594×896 218 KB
ATT
ATT.jpg1920×1080 168 KB
.

Images of Pixhawk 2.1 wiring and a screen grab of the graphed ATT showing where the flight mode changed and the pitch up.

The GoPro is on it’s second life after spending 5 days in 40ft of water in a fresh water lake in MT, so don’t much care if it takes a dive. Same with the 2 axis gimbal.

Working on the datalog upload.

Ron

7

Check your “Canard Half Span” and your “Wing Half Span” i think they should be the same judging by the picture. Also the if the wings are the same size the canard CG should only be slightly (20-40mm) forward of the copter CG, and when the motors are tilted forward they might move the CG forward naturally, once the CG is adjusted for the canard, the copter CG should be pretty close. To get both CG’s closer you could try reducing the front wing span 10%

8

Clinton:

Yes, I see the mistake I made now with the Canard half span.

9

Ron,

Just pull the flash card and copy it direct onto your PC/Laptop. You’ll probably need to host the file somewhere since the .bin file is usually quite large.

Cheers!

10

Greg and Mike:

Thanks for your time. I’ve decided to go back to something I know and have disassembled the build for quadcopter reassembly.

Well, not completely generic - I’m gonna get it flying as a quad, then add a wing in the middle. Hopefully, the wing takes some of the load, while the quad motors provide control. Also gonna have the wing set at a couple degrees positive AOA and have a servo keep it there via Pixhawk gimbal pitch tilt output. Will limit the quad max angle to 25deg to start, so the wing will only have to pitch around 27-28deg.

If it flies OK, but sluggish in roll, I’ll add on the ailerons from the tilt-wing and have them help stabilize using the Pixhawk gimbal roll output.

Thanks again,

Ron