New Pixhawk Cube Orange+ hardware incompatible with MATLAB PX4 Toolbox (PX4 v1.14.x)?

Hi all,

It appears that the Pixhawk Cube Orange+ has undergone a hardware change since the MATLAB R2024a PX4 Support Package (PX4 v1.14.0) was released. In my lab we use Cube Orange+ for custom flight control law testing, and a recent purchase of four Cube Orange+ units (ordered August 2025) has revealed a serious compatibility issue:

• Older Cube Orange+ boards (purchased before Summer 2025) work perfectly with MATLAB R2024a PX4 Toolbox (PX4 v1.14.0).
• Newer Cube Orange+ boards (purchased Aug 2025) are unusable with PX4 v1.14.x–based MATLAB builds, but work fine with newer stock PX4 firmware.

On the newer Cube Orange+ boards, I observe the following behavior:

• Stock PX4 v1.15, v1.16, and v1.17 build and flash without issue. RC input works correctly, and peripherals such as GPS, TFmini LiDAR, and airspeed sensors behave as expected.
• MATLAB R2024a PX4 toolbox builds (PX4 v1.14.0), which still work on older Cube Orange+ boards, cause the Cube to play the “error tune” immediately after boot when flashed to the newer boards.
• RC input is completely missing. The Radio setup screen in QGroundControl shows no channels, and MAVLink Inspector shows no RC_CHANNELS or RC_CHANNELS_RAW messages being published.
• RCIN does not respond to standard SBUS input.
• When ELRS/CRSF drivers are enabled, some switch activity appears in the Radio screen but is glitchy and not reliably received by the Pixhawk.
• TFmini LiDAR modules no longer publish to DISTANCE_SENSOR in MAVLink Inspector.
• PWM_MAIN channels 1–8 are disabled after uploading the v1.14.0 firmware.
• PWM_MAIN_TIM* parameters are missing, similar to the behavior reported previously on the Pixhawk 6C.
• From what I have seen so far, this issue appears to persist with MATLAB R2025a+ (PX4 v1.14.3 firmware), so the proposed 1.14.3 fix that helped the 6C does not appear to resolve this for newer Cube Orange+ boards.

Tests I have already tried on the newer Cube Orange+ units include:

• Flashing stock PX4 v1.17, fully setting up the board in QGroundControl, verifying that RC and TFmini work, then flashing MATLAB-generated PX4 v1.14.0 firmware (that still works on older cubes) → error tune and no RC/TFmini.
• Flashing stock PX4 v1.16, setting up the board, verifying RC and TFmini, then flashing MATLAB-generated PX4 v1.14.0 firmware → error tune and no RC/TFmini.
• Flashing stock PX4 v1.15, setting up the board, verifying RC and TFmini, then flashing MATLAB-generated PX4 v1.14.0 firmware → error tune and no RC/TFmini.
• Flashing stock PX4 v1.14.0 directly from the PX4 GitHub repository and setting up the board → error tune and the same loss of RC/TFmini.
• Performing a full wipe (SD card erase and parameter reset), then flashing MATLAB-generated PX4 v1.14.0 firmware that still works on older Cube Orange+ units → error tune and no RC/TFmini.

From these tests, I have deduced the following:

• The same MATLAB PX4 v1.14.0 firmware remains fully functional on older Cube Orange+ boards.
• The same newer Cube Orange+ hardware works correctly with stock PX4 v1.15 and newer.
• The newer Cube Orange+ hardware consistently shows broken RC/IO/TFmini behavior with PX4 v1.14.x (both stock builds and MATLAB-generated builds).

This makes me suspect that there has been a hardware revision in the Cube Orange+ and that the IO/RC/PWM layer in PX4 v1.14.x (including the IO firmware) is no longer fully compatible with these newer boards.

A few questions:

• Is MathWorks aware of this issue with newer Pixhawk Cube Orange+ hardware and the PX4 v1.14.x–based support packages (R2024a–R2025b)?
• Is MATLAB R2026a (or a future PX4 Support Package update) planning to move to PX4 v1.15 or newer, which appears to work reliably on these newer Cube Orange+ boards?
• Has anyone else seen this behavior on newer Cube Orange+ units, and if so, has anyone found a workaround or partial solution (e.g., custom IO firmware, patched board configuration, or updated support package)?
• Has anyone confirmed whether this issue is also present in the Cube Orange (non +) variants?

Thank you in advance for any guidance, support, or suggestions you may have.

-A

[Michael]

We have seen this behavior, but no work around yet

Answer 1

[Anthony]

I reached out to MathWorks and received some very useful information:

1. Launch & run the hardware setup screens from MATLAB and in the “Test Connection” screen, upload the firmware on the Cube Orange+ board. This would upload v1.14.3 now, but that’s fine.
2. Close hardware setup screen in MATLAB.
3. Launch QGroundControl (QGC). Wait for it connect to the newer Cube Orange+.
4. In QGC, navigate to MAVLink Console.
5. In the MAVLink console, run the command “bl_update /etc/extras/cubepilot_cubeorangeplus_bootloader.bin”
6. You would see the below expected output when the bootloader is updated correctly.

nsh> bl_update /etc/extras/cubepilot_cubeorangeplus_bootloader.bin

INFO [bl_update] image validated, erasing bootloader…

INFO [bl_update] flashing…

INFO [bl_update] verifying…

INFO [bl_update] bootloader update complete

Reboot the Cube Orange+ hardware.

Based on our preliminary testing, this has fixed the issue. I confirmed this also works for stock v1.14.0.

[dpb]

Appreciate posting the resolution and it's bound to help anybody else who may find it with same upgrade.

Answer 2

[dpb]

"Is MathWorks aware of this issue ...?

A search for "Pixhawk" on <the public bugs database> didn't return anything, but that doesn't prove they may not have internal data, but it wouldn't be at all surprising if they haven't yet discovered such a vendor change.

To ensure they are aware and to have any chance of getting a patch any time soon, submit the issue as an official support request at <Product Support Page>, the Answers forum is not an official support channel so it cannot be relied upon for direct Mathworks response to issues.

[Anthony]

I have also submitted a bug report/request on this topic.

Thank you for your comment.