@Eric-Katzfey I tried examining the logs for default modal AI px4 firmware IRIS configuration and my customized firmware with vtol modules enabled. Strangely, the CPU usage seems identical and above 75% even in the default case.

Log plots here: https://drive.google.com/file/d/1UZfAYMRQ-VfYnyf1u__l9NAUHnM8Q0PJ/view?usp=sharing.

I assume one of the cpus is the apps one and the other is the dsp one. This is still strange, since according to my comment above the total cpu usage of the apps was slightly over 70%.

@garret said in Firmware Structure (Adding VTOL support) - Clarifications:

total CPU usage is only slightly over 70%

As for the gazebo sending messages, still looking into it. However, I did some quick diagnostic and found that gazebo's publication rate seems sufficient:

IRIS case root@nikita-ThinkPad-P16s-Gen-2:/usr/workspace/voxl2_hitl_gazebo# gz topic -d 1 -z /gazebo/modalai_world/modalai/mag Hz: 89.91 Hz: 88.87 Hz: 76.84 Hz: 105.65 Hz: 87.52 Hz: 90.77 Hz: 83.98 Hz: 85.79 Hz: 57.25 Hz: 134.63 Hz: 82.31 Hz: 88.04 Hz: 48.27 Hz: 227.83 Hz: 87.30 Hz: 75.15 Hz: 60.36 Hz: 105.97 Hz: 101.91 Hz: 92.13 Hz: 89.28 Hz: 83.36 Hz: 88.95 Hz: 88.49 Hz: 55.99 Hz: 156.46 Hz: 85.59 Hz: 90.70 Hz: 50.35 Hz: 129.66 Hz: 92.53 Hz: 87.62 Hz: 84.50 Hz: 66.56 Hz: 90.34 Hz: 89.11 Hz: 73.71 Hz: 58.14 Hz: 117.21 Hz: 88.86 Hz: 49.12 Hz: 179.37 Hz: 91.40 Hz: 89.39 Hz: 62.59 Hz: 108.00 Hz: 87.86 Hz: 75.88 Hz: 84.17 Hz: 77.79 Hz: 72.78 Hz: 87.25 Hz: 93.21 Hz: 64.89 Hz: 73.10 Hz: 92.97 Hz: 53.24 Hz: 196.72 Hz: 90.09 Hz: 89.55 Hz: 48.60 Hz: 121.67 Hz: 68.07 Hz: 96.66 Hz: 91.59 Hz: 89.99 Hz: 91.62 Hz: 72.11 Hz: 91.43 Hz: 78.47 Hz: 91.06 Hz: 81.71 Hz: 72.23 Hz: 93.85 Hz: 82.83 Hz: 84.47 Hz: 73.80 Hz: 114.91 Hz: 85.34 Hz: 95.42 Hz: 79.51 root@nikita-ThinkPad-P16s-Gen-2:/usr/workspace/voxl2_hitl_gazebo# gz topic -d 1 -z /gazebo/modalai_world/modalai/baro Hz: 45.30 Hz: 33.52 Hz: 40.19 Hz: 38.64 Hz: 39.93 Hz: 45.78 Hz: 41.65 Hz: 42.63 Hz: 44.85 Hz: 39.09 Hz: 45.65 Hz: 36.76 Hz: 41.51 Hz: 43.36 Hz: 34.07 Hz: 47.10 Hz: 45.87 Hz: 39.17 Hz: 53.24 Hz: 40.17 Hz: 37.47 Hz: 37.41 Hz: 50.90 Hz: 37.27 Hz: 53.60 Hz: 38.04 Hz: 45.86 Hz: 43.55 Hz: 43.98 Hz: 44.47 Hz: 40.48 Hz: 39.65 Hz: 44.65 Hz: 43.66 Hz: 41.32 Hz: 42.98 Hz: 40.23 Hz: 44.23 Hz: 42.25 Hz: 30.15 VTOL case root@nikita-ThinkPad-P16s-Gen-2:/usr/workspace/voxl2_hitl_gazebo# gz topic -d 1 -z /gazebo/standard_vtol_hitl_world/standard_vtol_hitl/mag Hz: 261.34 Hz: 268.29 Hz: 165.38 Hz: 301.10 Hz: 219.42 Hz: 318.94 Hz: 258.53 Hz: 260.81 Hz: 259.69 Hz: 261.20 Hz: 230.35 Hz: 246.97 Hz: 105.02 Hz: 27011.70 Hz: 584.93 Hz: 245.19 Hz: 210.23 Hz: 137.50 Hz: 317.01 Hz: 217.37 Hz: 213.23 Hz: 245.10 Hz: 102.70 Hz: 7085.62 Hz: 255.86 Hz: 214.26 Hz: 358.80 Hz: 135.56 Hz: 601.92 Hz: 197.99 Hz: 289.66 Hz: 146.01 Hz: 710.52 Hz: 285.12 Hz: 183.18 Hz: 283.64 Hz: 267.36 Hz: 196.66 Hz: 293.54 Hz: 261.75 Hz: 189.79 Hz: 285.15 Hz: 269.74 Hz: 210.62 Hz: 247.68 Hz: 278.95 Hz: 92.07 Hz: 9.93 Hz: 39858.11 Hz: 249.81 Hz: 2192.14 Hz: 263.26 Hz: 255.81 Hz: 268.84 Hz: 204.21 Hz: 119.26 Hz: 9.63 Hz: 9.72 Hz: 9.71 Hz: 9.96 Hz: 2776.08 Hz: 133.24 Hz: 274.92 Hz: 823.84 Hz: 202.26 Hz: 276.13 Hz: 276.10 Hz: 216.36 Hz: 339.81 Hz: 270.43 Hz: 266.76 Hz: 261.38 Hz: 265.82 Hz: 268.36 Hz: 256.52 Hz: 276.77 Hz: 93.41 Hz: 1257.61 Hz: 263.08 Hz: 263.41 Hz: 146.94 Hz: 270.98 Hz: 243.40 Hz: 283.23 Hz: 150.90 Hz: 1511.17 Hz: 274.21 Hz: 258.02 Hz: 144.53 Hz: 2885.13 Hz: 247.98 Hz: 137.59 Hz: 829.43 Hz: 247.74 Hz: 254.85 Hz: 264.31 Hz: 209.19 Hz: 275.45 Hz: 267.43 Hz: 270.33 Hz: 245.57 Hz: 228.72 Hz: 232.26 Hz: 264.85 Hz: 93.60 Hz: 49265.94 Hz: 499.46 Hz: 233.43 Hz: 273.17 Hz: 234.92 Hz: 233.14 Hz: 258.48 Hz: 278.42 Hz: 225.91 Hz: 280.44 Hz: 135.68 Hz: 276.36 Hz: 134.54 Hz: 282.50 root@nikita-ThinkPad-P16s-Gen-2:/usr/workspace/voxl2_hitl_gazebo# gz topic -d 1 -z /gazebo/standard_vtol_hitl_world/standard_vtol_hitl/baro Hz: 132.49 Hz: 129.18 Hz: 129.00 Hz: 130.49 Hz: 111.43 Hz: 151.83 Hz: 97.87 Hz: 133.37 Hz: 108.81 Hz: 145.29 Hz: 88.34 Hz: 257.54 Hz: 127.70 Hz: 102.37 Hz: 131.12 Hz: 124.66 Hz: 104.16 Hz: 135.17 Hz: 117.07 Hz: 123.88 Hz: 134.75 Hz: 130.16 Hz: 134.76 Hz: 70.86 Hz: 577.87 Hz: 123.25 Hz: 120.28 Hz: 104.71 Hz: 179.62 Hz: 125.18 Hz: 119.61 Hz: 121.50 Hz: 148.17 Hz: 126.87 Hz: 103.36 Hz: 132.89 Hz: 145.11 Hz: 114.83 Hz: 113.25 Hz: 148.53 Hz: 132.36 Hz: 128.08 Hz: 124.23 Hz: 72.73 Hz: 282.04 Hz: 127.52 Hz: 128.52 Hz: 124.72 Hz: 121.43 Hz: 112.05 Hz: 126.82 Hz: 108.68 Hz: 165.72 Hz: 128.03 Hz: 124.98 Hz: 100.47 Hz: 9.51 Hz: 137.65 Hz: 93.98 Hz: 88.69 Hz: 135.25 Hz: 129.01 Hz: 130.18 Hz: 89.62 Hz: 143.84 Hz: 126.38 Hz: 138.17 Hz: 133.42 Hz: 122.09 Hz: 136.52 Hz: 136.22 Hz: 117.27 Hz: 54.04 Hz: 39575.75 Hz: 160.93 Hz: 127.02 Hz: 87.96 Hz: 260.10 Hz: 93.48 Hz: 110.70 Hz: 144.94 Hz: 129.48 Hz: 132.85 Hz: 119.29 Hz: 123.63 Hz: 143.99 Hz: 103.66 Hz: 136.10 Hz: 9.17

Weirdly there is some frequency jumps with VTOL world. Could this be the cause of the error? If you have an idea where these come from , I would appreciate the hint.

Although, when looking at the logs again, this is the magnetometer plots:https://drive.google.com/file/d/1yxWfiHGunIUJAk8WNnSOku5CxXks31so/view?usp=sharing

Apologies for small dot thickness, but it can be noted that magnetometer coming from the vehicle is at twice the frequency that it is coming from the sensor. Ironically, these rates are also off from the frequencies pasted above. For IRIS default the vehicle's magnetometer is at about 2Hz, and same for VTOL. Sensor frequency for IRIS is about 1Hz and same for VTOL. I wonder how this is possible, since for EKF2 to run smoothly it requires at least 5Hz update rate on sensors. I assume I am misinterpreting the data, so if you spot a mistake in my analysis please let me know.

@TomP , oh yeah, my bad.. you found the right one 🙂

@arcj and @Nse-Obot ,

Yes we just saw that there is an issue in the build.

However, it looks like the following workaround is working:

use voxl-cross:V2.7 use ./install_build_deps.sh qrb5165 sdk-1.3 to install the dependencies ./clean.sh ./build.sh qrb5165

Please try!

We will work on fixing the build for voxl-cross:V4.0. The native build on VOXL2 is also failing at the moment.

Alex

There might be a problem with the voxl-camera-server according to the output of systemctl status voxl-camera-server so I upgraded to the latest version in gitlab (2.2.4) and still get the same status output below.

voxl2-mini:/$ systemctl status voxl-camera-server ● voxl-camera-server.service - voxl-camera-server Loaded: loaded (/usr/bin/voxl-camera-server; enabled; vendor preset: enabled) Active: failed (Result: exit-code) since Thu 2025-07-03 17:42:10 UTC; 2min 6s ago Process: 2832 ExecStart=/usr/bin/voxl-camera-server (code=exited, status=255) Main PID: 2832 (code=exited, status=255) Jul 03 17:42:08 m0104 systemd[1]: voxl-camera-server.service: Main process exited, code=exited, status=255/n/a Jul 03 17:42:08 m0104 systemd[1]: voxl-camera-server.service: Failed with result 'exit-code'. Jul 03 17:42:10 m0104 systemd[1]: voxl-camera-server.service: Service hold-off time over, scheduling restart. Jul 03 17:42:10 m0104 systemd[1]: voxl-camera-server.service: Scheduled restart job, restart counter is at 3. Jul 03 17:42:10 m0104 systemd[1]: Stopped voxl-camera-server. Jul 03 17:42:10 m0104 systemd[1]: voxl-camera-server.service: Start request repeated too quickly. Jul 03 17:42:10 m0104 systemd[1]: voxl-camera-server.service: Failed with result 'exit-code'. Jul 03 17:42:10 m0104 systemd[1]: Failed to start voxl-camera-server.

@george-kollamkulam : the problem with the drone shooting straight up in position mode seems related to bad estimation of altitude. In particular, if you are flying indoors (no GPS signal accessible) GPS needs to be eliminated from the altitude estimation since the lack of GPS signal in the sensor fusion seems to cause the problem.
IN summary, if you are flying indoors and experience this problem, try disabling GPS in PX4 by setting the parameter SYS_HAS_GPS to "disable" in QGroundControl. That seems to have resolved the issue for us - at least for now.

@taiwohazeez The path is https://gitlab.com/voxl-public/voxl-sdk/utilities/voxl-mpa-to-ros2

Git will always require a username / password if you're cloning via http

You can set up an SSH key tied to your gitlab account and clone via SSH to get around this

I grabbed the /usr/bin/quectel-CM from another VOXL2 on an older SDK (1.4.1) and swapped out the one that came with 1.5.0 and everything is working again...

Hi again,

From investigating the voxl-vision-hub code with added prints for debug, I see that the setpoint SETPOINT_POSITION_MAGIC_NUMBER is sent to mavlink-server from two functions -

(1) _send_current_position()
sending the current position, and type-mask ignoring vx,vy,vz,ax,ay,az,yaw_rate

// fetch latest position and attitude from px4 itself so the setpoint // we are about to send is a close as possible to where we currently are

(2) execute_setpoint_position_command()
sending the setpoint as recieved from voxl-planner

The implemented behaviour of (1) suits the situation of position+yaw command. However for vx+vy+z+yaw_rate command I think something is missing.

Any help will be appreciated.
Thank you, Meytal

The overall goal is just to get time synced between the Lidar and the VOXL2's IMU. Oddly enough, I was looking through the voxl-kernel for qrb5165 and the driver for PTP is there. Is the physical layer chip not able to support PTP?

Where I see the driver in the Kernel:
https://gitlab.com/voxl-public/system-image-build/qrb5165-kernel/-/tree/qrb5165-ubun1.0-14.1a/drivers/ptp?ref_type=heads