Instable Yaw control on differential trust; adding additional servos on top of 4 motors and 4 servos

[CristianoUspace]

Hello everyone. I'm using INAV 9.0.2. on an Axispro7 FC (9.0.1 firmware), trying to set up a VTOL flying machine. You'll find a diff_all file attached.

Fine with setting up the mixed profiles mode. One profile is just a quad, with tilting motors ahead. After transition, it turns to be an aircraft with two pulling propellers ahead, ailerons and pitch. For Yaw control, I use differential trust of the two engines. But this doesn't work properly: as I increase trottle, one propeller (the left one, motor #2) is spinning more.

INAV_9.0.1_cli_20260527_174636.txt

Initially, I can correct with the stick, finally it diverges: If I try to take off on a runway, the aircraft ends up spinning on itself before gaining enough speed for takeoff. That's one issue. I tried to work out adjusting PID and lowering the index of YAW in the mixing tab, but so far I couldn't succed.

So far, I've 4 engines and 4 servos. The mixer table is full, in both profiles.

Now, I'd like to have one or two servo ports, using unused UARTs on the FC. I'm using a Specktrum RC with SRLX, with several channels. I would use them:

1. to control yaw aerodinamically by a rudder, by RCH Flight - New default PIDs for navigation #4;
2. to control flaps, e.g. by RCH Flight - NAV_POSHOLD messes up with yaw control #6.

I followed a tutorial to enable SBUS on a free port (e.g. UART 3), then on CLI "set servo_protocol = SBUS". As soon as this is done, the 4 servos commanded through traditional mixer table stop working.; that's the main issue.

Secondary issue: I understand that the whole radio signal (i.e. 16 channels) would be forwarded through that port, which means I need another piece of hardware (e.g. an ESP32 board duly programmed) to finally command the additional servos. That's not a big issue. I can manage it, I already tested the signal and, yes, I can read some bits coming out of the UART 3 to the ESP32 board. Likely I just need to find the correct library to process such signal and dispatch commands to servos. But first I need to solve the first issue :-(

Any suggestion for a simplier configuration?

[sensei-hacker]

"set servo_protocol = SBUS". As soon as this is done, the 4 servos commanded through traditional mixer table stop working

That sounds like you want to set servo_protocol SBUS_PWM

But that board has seven channels plus the LED pad makes 8 separate channels. isn't that enough for?:

Four motors
Ailerons
Elevator
Tilt servos
Plus a rudder if you wanted one!

Are you duplicating some output channels on multiple pads, or can I not count because I just woke up and haven't had my coffee?

Note differential yaw can very easily result in a spin when you make a turn while flying slow, like you are coming in to land (and are already at a low altitude with no time to recover). Be careful with that.

[CristianoUspace]

Hi there, thanks for replying. I'm currently having 4motors, then (for aircraft configuration):

1 servo port is driving tilt servos (2 servos with Y cable)
2. servo port is driving aileron left servo
3. servo port is driving aileron right servo
4. servo port is driving pitch servo

I might put the two aileron servos on a Y cable, but I'm using mix to trim both servos like flapperons..
screenshots profile 1 (QUAD)

screenshots profile 2 (AIRCRAFT)

Logic Condition 0 is reserved for Quad mode. Logic Condition 1 starts moving forward, still all 4 props are spinning, Logic Condition 2, tilt in full forward position, only two props spin. I'd love to add one more servo to replace differential trust (which you confirm has some wird behaviours) with traditional rudder.... (or at least decrease a lot the differential part).

Thanks for supporting. Cristiano

[sensei-hacker]

You could put a servo on the LED pad if you wanted to.

I might put the two aileron servos on a Y cable, but I'm using mix to trim both servos like flapperons

Yeah I don't think I would add a whole other microcontroller doing sbus to PWM just to add flaps to a VTOL. Unless I just wanted to. :)

[CristianoUspace]

Super. One more servo would be pretty enough. Servo #2 will serve tilt, I’ll change alilerons to y cable to Servo #3 add flaps to two more servos on another Y cable connected to Servo #5 (giving up flapperon mixing) and set servo # 6 to serve a rudder.
To do so, I set S9 output from auto to Servo and add a mixer rule in the servo section to the newest servo number (I reckon it should be #6),
mapping it to stabilised yaw or (conditionally) to RC channel 4. Is that right? Thanks again

[sensei-hacker]

Yep, that sounds right.