@ApoorvThapliyal : the problem (Error message: "set_position_target_local_ned: coordinate frame 0 unsupported") appeared in QGC after updating VOXL SDK from 1.5.0 to 1.6.3.
The error blocked me from being able to use QGC to calibrate or arm the drone. Stopping and restarting voxl-vision-hub solved the problem.
For now I'm using voxl-vision-hub as it comes on SDK 1.6.3 (no changes yet) and have not made changes to the .conf file (yet).

@Eric-Katzfey : I'm not running any custom software other than voxl-vision-hub.
If I stop and disable voxl-vision-hub, the "set_position_target_local_ned: coordinate frame 0 unsupported" error goes away !
With this I am able to complete calibration (compass and IMU) successfully in GQC, BUT drone fails to arm in Position mode even though QGC shows "Ready to Fly" and all Sensor status show as "normal".
Checking VIO in VOXL Portal shows correct operation of VIO: VIO plot follows drone motion, which seems to indicate that VIO is working fine, but
QGC shows "Yaw estimate error" , and "--" as altitude estimate, which prevent the drone from arming in Position mode.
Then I re-enable and restart voxl-vision-hub.
Now GQC receives altitude estimates and I'm able to arm the drone in position mode !
So the trick seems to be to disable/stop voxl-vision-hub, calibrate IMU if needed, and enable/restart voxl-vision-hub, to be able to arm in Position mode.
Seems odd but I'll take it ! Let me know if you have any comments or suggestions. Thanks.

Hello -
After upgrading SDK from 1.5.0 to 1.6.3, I am unable to calibrate the Starling 2 in QGroundControl (Version 4.4.4)
Error message: "set_position_target_local_ned: coordinate frame 0 unsupported"
It seems the Starling 2 is sending a Mavlink message that is trying to set the coordinate frame to 0 (MAV_FRAME_GLOBAL), which cannot be processed by the local NED setpoint command.
Any suggestions on how to correct this error would be appreciated.
Thanks

@Ben-Linne : Thanks for the excellent instructions. Worked perfectly first shot. betaFPV is back to normal and binds to the radio controller ! Best regards, Hector

@Ben-Linne : The blue LED is stuck on solid blue as soon as you provide power. The receiver seems stuck on bootloader mode. I've followed several methods available to re-flash the receiver, but communication is not possible. It seems the receiver is dead.
Question: is this behavior expected (betaFPV receiver unable to communicate) when upgrading the SDK to 1.6.0 or 1.6.3 ? Is this receiver still supported by the new versions of the SDK ? Let me know - Thanks.

Hello,
I recently updated my Starling 2 from SDK 1.5.0 to 1.6.3. At the end I got the following error message:
FAILED TO EXECUTE the following steps: voxl-elrs configure.
My receiver is the BetaFPV ELRS 900 v1.1 II, 915 MHz. I connected the ELRS receiver to a FTDI USB interface and tried to configure it by the following ways:

• using ExpressLRS configurator tool
• using ModalAI ExpressLRS configuration tool
• Power off the drone completely
  . Hold down the receiver's boot button, Power on the drone while holding the button
  . Keep holding for 1 second after power on, then release
  . Run the following command:
  python3 -m esptool --port /dev/slpi-uart-7 --chip ESP8266
  --baud 115200 --before no_reset --after soft_reset
  write_flash 0x00 /usr/share/modalai/voxl-elrs/firmware/3.5.3.21/BETAFPV_900_RX_via_UART/BETAFPV_900_RX-3.5.3.21-20251205.bin
  None of these options worked. It is not possible to connect to the receiver - it seems dead.
  QUESTIONS:
• is the BetaFPV ELRS 900 v1.1 II, 915 MHz, no longer supported in the latest versions of the VOXL SDK ??
• If the answer is yes, which is the latest version of the SDK that I can use without bricking my receiver ?
• I tried to purchase a new receiver chip from the ModalAI store, but is not available. Where could I get one ?
  Please let me know at your earliest convenience. Thanks.

@Alex-Kushleyev - thanks, I submitted the RMA request, will let you know after I have the chance to replace and test the replacement parts.
Thanks, Hector

@berayksl : If your problem is with the height estimate and your VIO is working well and gives you good estimates of altitude, try disabling the barometer as well: SYS_HAS_BARO = disable, SYS_HAS_GPS = disable.

@berayksl - we have an indoor lab . There is a large steel table on one side of the room that affects the magnetometer. The drone has to stay away from that object to be able to arm. Large ferromagnetic objects definitely affect the magnetometer. Can you arm and fly position mode in a regular room with magnetometer enabled and GPS disabled ?

@berayksl : IN my case the problem has not appeared again (fingers crossed !) since I deactivated GPS from the EKF2.
From what you described above, it seems you also deactivated the magnetometer (SYS_HAS_MAG = 0), have you tried to enable that ?
Lack of magnetometer may lead to unstable EKF2 sensor fusion ?