ModalAI Forum

We do not support DSHOT when using Voxl2 + IO Expander board. However, OneShot125 is supported (though you would need to re-build voxl-px4 pakage and change the following scale factor : https://github.com/modalai/px4-firmware/blob/voxl-dev/src/drivers/voxl2_io/voxl2_io.hpp#L133) -- set MIXER_OUTPUT_TO_CMD_SCALE = 125 instead of 1000. This would reduce the actuator latency by 1-2ms (as long as the ESC supports OneShot125).

Hi @electroAJ, thanks for the detailed writeup and follow-up SKU info. From a hardware standpoint, the 200mV droop you are observing on VDC_5V_LOCAL and VDC_5V_LOCAL_USB1 during ToF activity is not a concern at the rail level. That kind of transient is most likely localized contact/pin loss at the connector interface rather than a bulk supply issue. On a 5V rail, 200mV is well within normal operating margin. The eFuse on that rail does not assert until approximately 4.4V, so a 200mV droop from 5V puts you nowhere near that threshold. Downstream logic is regulated from a 3.8V VBAT rail that has demonstrated margin well above the minimum needed, and that rail is permitted to droop past 3.3V before any real issue occurs. The ToF-induced droop is not your RC dropout culprit. Additionally, the M0129 VOXL Mini ESC power stage has been proven to the same load capability and test rigor as our standalone M0041 power module, using the same parts, so the supply itself is not a suspect to me. Would hate to see you chase a red herring. The I2C errors and UART dropout sequencing are worth looking at together as they may share a common upstream cause. Beyond that, this one is likely better handled by our software team who can dig into the dmesg and service state in more detail. I will flag this thread for them.

@Alex-Kushleyev Hey Alex. Thanks for this input. I will incorporate your feedback as we continue initial motor testing and let you know if we need further guidance. Much appreciated!

@Jetson-Nano Given the crash and hardware issues I'd recommend sending this board in for an RMA

Hello @jameskuesel , Sorry for the confusion about the lens change. Please see the details below. The original lens in the IMX412 modules (M0107 and M0161) had manufacturing issues where the internal components of the lenses were loose and vibrating in flight, causing shaky video. Due to quality / vibration issues of the original lens (part number M10000580) shipping with IMX412 (M0107 and M0161) modules, the default lens was changed to M10001028 (late 2025). However the new lens does have much narrower FOV. When purchasing the standalone camera, you have a choice of lens: M0161 with old lens (wide) : MDK-M0161-1-07 (3.1mm EFFL, M10000580 lens) M0161 with new lens (narrow) : MDK-M0161-1-08 (3.24 EFFL, M10001028 lens) It seems that there maybe some discrepancy in the manufacturer stated EFFL for the new lens (3.24 vs original 3.1) because the actual FOV change is more significant than the change in the effective focal length would result in. We are looking into this. Starling 2 and Starling 2 Max drones are currently shipping with the new lens (narrow FOV). We do currently have stock of the old (wide) lenses that should have no issues (according to the manufacturer), however we cannot provide any guarantees for this lens due to past history of issues. If you would like to update your drone / camera with the original wide lens, please contact us and we will make it available for you to purchase (lens part number M10000580). Alex

@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!

You are welcome! By the way the Starling 2 Max plastics kit (linked above) contains both types of front camera mounts (with and without the TOF sensor).

@Robbie-McCue Thank you for your feedback I will look into this.

@NikosMav Yes, things will get hotter when running QVIO. But once you arm and start the props everything will cool down. When reflashing the SDK I would grab one of the official releases from our downloads page and not a nightly.

What i meant was that sometimes, depending on the frame design, the orientation of the cameras may change slightly over time. This can happen as the plastics can change shape due to exposure to heat, sunlight, humidity (depends on the plastics), bumps and crashes, etc. Additionally, there has been some research showing that the camera + lens intrinsics change over time / temperature, especially if the lenses are made of plastic (which is the case in our old OV7251-based fisheye modules, which use plastic lenses - although we have not characterized them vs temperature). It is possible to use stiffer and more stable materials (metals, carbon fiber) to ensure that the sensor mounting remains consistent over time. I hope that answers your questions. Alex

Starling 2 Max is designed for 4S batteries only (two 2S Li-Ion battery packs, which are connected in series inside the vehicle). The Mini ESC, which also acts as a power source for VOXL2 inside the Starling 2 Max, is rated at least for 4S input (depending on the version of the ESC). There are several versions of the Mini ESC to provide the users some flexibility, in particular the voltage of the VOXL2 regulator (5V for VOXL2 and 3.8V for VOXL2 Mini). Since Starling2 Max uses a VOXL2, the ESC part number will be either M0129-5 or M0129-65. An updated (6S) version of the Mini ESC was released after the release of Starling 2 Max (new ESC part numbers: M0129-63, M0129-65). This ESC update extended the supported voltage for the ESC (not for Starling 2 Max) from 4S to 6S. The 4S version of the ESC has been phased out because the 6S version outperforms the original version even at 4S voltages (due to better Mosfets), so the 6S version is an improvement over the original 4S version of the ESC, while being fully backwards compatible. Using the 6S-rated ESC in the 4S-rated Starling 2 Max is completely safe. If you use the Mini ESC outside of the Starling 2 Max, you need to verify the part number / sticker on the ESC before plugging in 6S battery (M0129-63, M0129-65 only) The ESC variants and part numbers are documented here : https://docs.modalai.com/voxl-mini-esc-datasheet/ Alex

Hi @kristi, I'm experiencing a similar, though maybe not identical issue. Would you mind sharing the repeatability of your problem? is it consistent? thanks.

@alex-kushleyev @ModalAI-Team . Could you guys help me with this? thank you!

@Omri-Eival The Zorro Blue User Manual should answer most of your questions, with a shell script to copy all of the latest data to a handset in the provided zip file. While you are there, if your firmware is out of date it may be worth going through the update process.

Hello, I would like to follow up on this topic. We are currently evaluating the feasibility of using Starling 2 for indoor autonomous flight applications in Japan. Understanding the recommended Wi-Fi configuration for operation in Japan will help us continue our evaluation. If anyone from the ModalAI team could provide guidance, we would greatly appreciate it. Thank you very much.

@mrawding , I am sorry, but it looks like we are not going to be officially supporting Ubuntu2.0 Kernel on VOXL2, as we are working on enabling next gen hardware.. We did start the effort but it is not going to be finished for the general availability for VOXL2. Alex

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.