Yes, Hadron is working when plugged into J6 of VOXL2 Mini with default kernel.

You just need to have the following sensormodule files in /usr/lib/camera/:

com.qti.sensormodule.boson_0.bin com.qti.sensormodule.ov64b40_1.bin

And instead of gpio6, you need to use gpio30 to enable the CCI mux, so that communication with the EO camera is enabled:

voxl-gpio -m 30 out voxl-gpio -w 30 1

I will update the Hadron documentation..

Alex

@Moderator got it. Thanks!

@Jetson-Nano
YES!!! So happy to hear that.

@wifa799 ,

Please see the following post : https://forum.modalai.com/topic/4274/hadron-boards-to-order/ (at the end)

In short, Hadron works in J8 only using the kernel variant that works with M0173. I need to update the documentation.

So you can have 3x tracking, 1x hires, 1x TOF, and Hadron (IR + OV64B hires). That is 7 cameras which is the maximum number of MIPI cameras that VOXL2 supports. I know you did not ask for the other hires camera (recommended IMX412 M0161 module), but just wanted to mention that.

This set up is not standard, but you would just need to set up for a standard Starling camera setup (C27) then go from there.. copy the sensormodules (camera drivers) for Boson in slot 4 and ov64b in slot 5 and update your voxl-camera-server.conf. I can help you with that if needed.

Alex

@Andrew-Jue , you can find the latest firmware for the IO board in /usr/share/modalai/voxl2-io-tools/firmware/modalai_m0065_firmware_v0_2_RC1_f94baad1.bin on your VOXL2 or on gitlab: https://gitlab.com/voxl-public/voxl-sdk/utilities/voxl2-io/-/tree/master/voxl2-io-tools/firmware

You can update the firmware using our voxl-esc tools, as documented here : https://docs.modalai.com/voxl2-io-firmware/#how-to-update-firmware

This version of the firmware is the latest and has been stable for over one year.

Alex

@jonathankampia There is some relevant info here: https://docs.modalai.com/voxl2-external-flight-controller/
M0062 is more rare of a product for us so isn't listed there but you can find more info here: https://docs.modalai.com/m0062-datasheet/#j9---2w-uart--i2c-port

@Alex-Kushleyev

Thanks for the explanation — I now understand how to use the TOF sensor.
I’ll use the ROS point cloud directly for my research.

Hi @smiley
We don't offer longer ones and we have not tested them.
Sorry I can't be of much more help here.

Usually in this situation, I suggest our customers to go and challenge their Mechanical and Industrial Design Engineers to better constraint image sensor placement in the drone given these fundamental constraints.

@Daniel-Rincon I'm hoping you are still around as we might be having a similar issue. When you say documentation, do you mean these: https://gitlab.com/voxl-public/support/voxl-train-yolov8

What specific helpers did you add with what model conversions? (Reference: I also have a custom yolov8 model trained for an output of two classes.)

@gestrada1150 We have an experimental fork of betaflight running on VOXL 2 now. However, there are no official plans to support this.

@Vinny Thank you for your valuable feedback.

I’ll keep the points you mentioned in mind while doing the bench testing.

@holmeslice Regarding having the external flight controller, it's really only needed if the I/O limitations of VOXL 2 are a problem. For example, if you need CAN bus then you should probably use the external flight controller since VOXL 2 has no CAN bus support.

@dan64 Unsure if you have resolved this issue, but I am going to add some more information to anyone that stumbles across this. There was a bug in the voxl-mavlink-server setting and sending all System ID numbers to 1. The deb file below should solve this issue where the MAV_SYS_ID can be changed an reflected on the VOXL2.

http://voxl-packages.modalai.com/dists/qrb5165/dev/binary-arm64/voxl-mavlink-server_1.4.12-202509181025_arm64.deb