ModalAI Forum

@alex-kushleyev @eric-katzfey Apologies for the direct mention — I hope you don't mind. I posted this question set about two weeks ago (and a related product question three weeks ago), but haven't heard back yet. We are preparing to purchase a Starling 2 Max (C29, with the PMD ToF) as an evaluation unit for an indoor GPS-denied inspection project, and just three answers would unblock our decision: Can VOXL SLAM / VOA (with the dense PMD ToF on the C29 configuration) reliably handle very narrow corridors — around 300 mm width? Any references to pipe/industrial inspection use cases in similar tight spaces would be very helpful. For repeated missions in the same indoor space, is voxl-mapper + AprilTag re-localization the recommended pattern to bound Open-VINS drift (no loop closure)? What is the practical minimum illumination for reliable VIO with the tracking cameras — and do you recommend onboard IR/LED illumination for dim industrial interiors? Even brief pointers to relevant docs would be greatly appreciated. Thank you for the great platform and documentation!

Thank you for your reply. After upgrading the Voxl2 SDK, two cameras were recognized on the J7. For now, I have successfully recognized four cameras using the ov7251 via the J6, J7, and J8. If the J6 recognizes only one camera, does that mean the J8 will also only recognize one camera? I want to have Voxl2 recognize all four ov7251 cameras and one Boson camera. So, that is a total of five cameras connected. Is it currently impossible to connect five cameras with Voxl2? I would like to know if there is a way to make this possible. Thank you in advance for your answer. [image: 1784163730641-%EC%8B%9C%EC%8A%A4%ED%85%9C%EC%9D%B4%EB%AF%B8%EC%A7%80.jpg] [image: 1784163780359-kakaotalk_20260706_162025619.png] Current test configuration / 4 cameras recognized [image: 1784163653454-kakaotalk_20260715_143030812.png] Configuration I want to recognize in the future / 4 ov7251s, 1 Boson+

Digging up an old thread, any updates on the move to Ubuntu2.0 Kernel?

Hello @Alex-Kushleyev , This is exactly what we needed ... thank you for turning the combined flip + GNSS-fix driver around so fast! Confirming our plan against your docs: We're on the encoded (ISP) path ... large_encoded for the pilot's low-latency FPV feed, not MISP ... so the ISP should pull the reversed bayer from the flip sensormodule and debayer correctly (no R/B swap). We'll run the version-agnostic combo: the flip sensormodule + en_rotate = true, plus maxRAWSizes=20 in camxoverridesettings.txt. Bit rate: going with 2088 Mbps. Our GNSS is a Septentrio mosaic dual-antenna running moving-baseline heading, so GPS L1 integrity is critical for us ... your analysis showing 2088 sitting ~10 MHz off L1 center with only a 1–2 dB drop is exactly the margin we need. We'll keep our FPV mode clear of the three forced-2200 modes (staying 1080p ≤120 fps). We'll validate by logging the Septentrio C/N0 and satellite count with the camera streaming vs. off, in representative daylight (noting your point that under-exposed frames worsen the interference). One quick check: for our single hires IMX412 on the M0173, is the default i2c 0x34 the right choice, or is there a scenario where we'd want the 0x20 variant? I'll report back the before/after GNSS numbers once we've tested ... happy to share the data if it's useful. Thanks again, John

This is the commit that I'm planning to propagate: https://github.com/modalai/px4-firmware/commit/c4fd40afc9f0c79a83a633023013f11b1185cf23

Wanted to check in again. Any advice on swapping spektrum RC for ghost atto on voxl2?

Thank you very much, Vinny, for your time and response.

@moderator Thank you very much for your detailed explanation of the VOXL 2 Mini ToF sensor. This information will greatly assist us in finalizing our approach for integrating the sensor into our UAV platform. Thank you once again for your continued support and guidance.

the hexagon sdk has been unlocked. It is under the miscellaneous folder: https://modalai.filecloudonline.com

Yes, your understanding is correct. To see the PX4 IMU values use standard PX4 commands. For example, on voxl2 you can use the PX4 listener commands: voxl2:/$ px4-listener sensor_accel TOPIC: sensor_accel sensor_accel timestamp: 30123125 (0.006725 seconds ago) timestamp_sample: 30122829 (296 us before timestamp) device_id: 2490378 (Type: 0x26, SPI:1 (0x00)) x: -0.05360 y: 0.01754 z: -9.68999 temperature: 34.62560 error_count: 0 clip_counter: [0, 0, 0] samples: 20 voxl2:/$ px4-listener sensor_gyro TOPIC: sensor_gyro sensor_gyro timestamp: 37716660 (0.003955 seconds ago) timestamp_sample: 37716405 (255 us before timestamp) device_id: 2490378 (Type: 0x26, SPI:1 (0x00)) x: -0.00409 y: 0.00796 z: -0.01429 temperature: 34.32669 error_count: 0 clip_counter: [0, 0, 0] samples: 20

Yes, thank you Vinny.

Yes, unfortunately some of the dependencies have changed. Can you checkout tag sdk-1.6.4-1 of voxl-mpa-tools and then in the docker ./install_build_deps.sh qrb5165 sdk-1.6. That should work.

Hi @adam These claims remain true. We are exploring some items in the future for individual certification to ease downstream efforts by integrators, but as of today, the efforts still largely fall on the end-product integrator (which is unavoidable no matter what certifications we employ) and we can help where needed if any requests are specific and actionable.

Hi @muqing-cao Correct, M0213 does not have a USB hub on it like M0090.

@RoyAzriel 1.6.4 release is tested on Starling 2 now, go ahead and use that from the link I emailed you.

Hi, I would like to know what wifi module is on the drone. The one mounted on the drone is different than AlfaNetworks? Adam

Hi @awagner Apologies for the delayed response. It looks like this post slipped through during our recent forum and documentation site refactor. Are you still experiencing issues with your M0041-4? A few clarifying questions will help: When you mention "that is dead too", can you please clarify what exactly is no longer working? The M0041-4 power module, the VOXL 2 Mini, your power supply, or some combination of the three? Can you share more details about your ESC, make, model, and how the battery pads are connected to the power module? A photo of your wiring connections would be very helpful here. On the spark at plug-in: direct battery connection is the standard way to power up a drone, so that in itself is not unexpected. The spark on connection is not uncommon with a 6S pack due to charge balancing and inrush transient at the moment of contact (especially with our FPV ESC with near 1,000uF of capacitance). That said, we do include protection on the M0041-4 for this reason, so something unusual may have happened in your specific setup, which is why we'd like to see the connections. As a general note, we typically see XT-30 used up to 4S and XT-60 up to 6S. Running 6S on an XT-30 is outside the more common usage pattern and worth flagging as a potential factor. We had some folks mistake a 4S ESC for 6S due to exactly that on early Stinger models. We've since remedied that. One practical note: depending on what has failed, a replacement power module may be a faster and cheaper path than a full debug effort. That said, we do want to help you identify the root cause so this does not repeat. Replacing hardware without understanding the failure just puts the next module at the same risk. Let us know and share photos if you can. Again, sorry for the long delayed response. Thanks!