Dual Rotor Non Thrust Vectoring Tail-sitter Failing to transition back

[Stefard_D_Souza1 (Stefard D'Souza)]

Hi guys,

I have been working on a Dual rotor tailsitter, i have an issue where the tailsitter drops throttle for an instance when transitioning from FW to Qhover & Qland, this causes the tailsitter to become unstable and i end up transitioning it back to FW and then belly landing it.

This i have not observed when transitioning from FW to Qstabilise as i keep the throttle fairly high up.

The message that i get is Vtol Transition Done,timeout

I have tried this at Vertical transition angles of 70 degrees,80 degrees and 20 degrees. all have the same effect, ive tried at transition rates of 30 degrees/s and 50 degrees/s.

PS the Tailsitter weighs 2.65kgs.

Ill attach the link to the Telemetry logs for the one of the flights.
Tailsitter Telemetry Log - Google Drive

Requesting you guys to please help me narrow down and solve the issue at hand.

[Farkhod_Khikmatov]

This is a very common problem we need to address. I have the same issue actually. Do you use an airspeed sensor?

[Stefard_D_Souza1 (Stefard D'Souza)]

Yes, i do use an Airspeed Sensor.

indeed, really hoping someone knows much more about the issue and the solution that needs to be implemented.

[Stefard_D_Souza1 (Stefard D'Souza)]

@tridge Requesting you to please shed some light on the issue at hand.

[Satoru_Sasaki]

Tailsitter fails to transition back to vertical, due to lack of stability and controllability in vertical mode. I cannot say much with TLOG, you need to share .bin log. Does your design meet basic requirements? 3-50% of large elevon to wing chord, prop wash covered etc?

This thread has lots of info.

Dual-motor tailsitters

This is a place to discuss support for dual-motor tailsitters with ArduPilot. This is for any fixed-wing airframe which sits on its tail to startup, and has two forward motors for thrust. The motors use differential thrust for roll when hovering, and elevons for pitch and yaw. Support for this type of airframe is in master now, but has only been tested in a simulator. We’re now waiting on the first real aircraft flight tests. Mark Whitehorn may well get there first, as he already has one flying…

I learned a lot of tuning for a year, with advices from veterans and numerous crashes.

Eclipson E-VTOL, 3D printable dual motor tailsitter

10 inch prop config is good, tuned further and elevon vibration is less now. Short video of flying around on an autumn sunny no wind day, Log is here. [Eclipson E VTOL flight highlight]

[Stefard_D_Souza1 (Stefard D'Souza)]

Thank you for the valuable input. i believe the Elevon Area is within the 33-50% as suggested for tailsitters.

could you please elaborate a little bit more on the Prop wash covered part?

the elevons are in fact directly under the prop wash.

PFA the link to .Bin Logs to one of the recent flights

2024-01-02 16-34-58.bin

We flew the system in FBWA, FBWB, Qstabilise and Qloiter mode,
Attempted multiple back transitions into different Q_modes
Initially the Q_tailsit_ang_vt was 70 degrees, which we then changed to 20 degrees noticing that when trying to back transition, the aircraft wouldnt cross more than 20 degrees pitch angle.

[Satoru_Sasaki]

Thank you for posting .bin log. I am no expert but can see your plane is not responding enough for desired pitch control. (Pitch is not reaching DesPitch). Plane will fail to transition and stay in vertical, or timeout to the level desired. Looks like there are some wind (3-4m/s) condition.

Wash from larger dia prop covers elevon more and can give better control. I experienced noiceable improvement when changed 8 inch prop to 10 inch.

I would tune better. Increase P, D at the level vibration will start soon. PID gain scaling will help, applied disk theory.
Information on this tailsitter tuning guide is all valuable. Start from finding appropriate FF value, then tune rest.

Since you use H7 based PixHawk 6C, you can run new quicktune lua which is very helpful.

Good you have large elevons. For plane design, I am advised to move CG as aft as possible, to the level you need FBWA in plane mode. It will improve vertical pitch manuevability.

Transition parameter will not improve transition success, but for smoother transition. QStabilize and FBWA is basic mode to use.