Hi @a-w , If you're running on target, you'll need an arm64 container flavor (not x86/64), see this docker file as an example: https://gitlab.com/voxl-public/voxl-docker-images/voxl-docker-mavsdk-python/-/blob/master/Dockerfile#L1

@Kasum-Hussain actually sorry my mistake i thought i saw video2 but no if i go into dev folder there is no video2 only video0 video1 video32 video33

Thanks this fixed it. It was connected backwards. Maybe this can be marked a bit clearer in the docs, I think it’s quite unintuitive having the cable going inwards the board when not using the extension.

Hi, I am facing the same issue [image: 1624369128135-1a3fbb9a-05ba-4660-8b39-9e658f6a2d35-image.png] [image: 1624369154908-ca2512c7-9c2e-4b01-864d-c55f02b944f3-image.png] Also, I don't have a USB Expander and Debug Add-On Board (https://docs.modalai.com/usb-expander-and-debug-manual/) Is there any other way so we can Force VOXL into Fastboot Mode. (like any onboard button/ any other technic) please let me know.

It all depends on your setup. With our flight deck setup we can run all 4 cores at 100% with just the fan. A fan is much more effective than a heat sink. Still might require a little experimentation though

I was able to resolve this issue by trial and error of the camera orientation, it seems like the documentation is outdated and does not provide an accurate reference for a starting point to obtain the proper values.

Also, you could do a quick test on DSP side, just write more data to UART port than the baud rate can handle and see what the write() function returns and how much time it takes. Usually the write() function is non-blocking, as it will hand off data to underlying layers to take care of UART writing. But I suspect that if the SDSP UART system runs out of buffers to put your data in, it will probably block until buffer space for writing frees up instead of dropping data. I have not confirmed this, but that's my expectation.

@Daniel-Ryu thanks for posting the resolution!

@Chad-Sweet Hey, I updated my voxl to the newest system image and after doing so my docker is using docker-start.service while before it was using docker-daemon.service. Not sure why this has changed, but I have updated the above instructions and wanted to let you know so you could change this in the documentation. Is there a reason why this would have changed after flashing a new system image onto the device? Both times I configured my docker by using voxl-configure-docker-support.sh

@gauravshukla914 no problem, glad it helped

@Alex-Kushleyev great! That’s a point what I need. I’ll try it Thanks,

@Chad-Sweet voxl-dfs-server is for voxl-vision-px4 ver 0.8.1. version 0.6.8 is not used voxl-dfs-server. could you please let me know based of voxl-vision-px4 ver 0.6.8? Thanks,

@Alex-Gardner thank you!

@FourPointTwoMm Interesting, I was under the impression that the Microhard modems would auto-reconnect to each other when they came back in range. This is something we will have to test.

The Yocto OS uses opkg which is similar to, but not compatible with, aptitude Aptitude is used in Debian/Ubuntu distros. You can run one on VOXL using Docker as documented here: https://docs.modalai.com/docker-on-voxl/

@Chad-Sweet Update: I have performed the calibration process as you mentioned for both Stereo camera and tracking camera by following the calibration documentation https://docs.modalai.com/calibrate-cameras/ though they are not moved from their original position of VOXL flight deck. [image: 1622039548849-2744ae31-4605-48ab-be86-bbdcc97f51cf-image.png] While performing the calibration for stereo cameras I have also fixed the focus for both left and right camera. But after the calibration process the collision prevention does not seem to be working any more. By that I mean the the obstacle topic does not seems to detect anything even it is very close or far from the wall. I also tried with different object to see if the camera can pick up the difference but it does not seems to be consistent any more. At 0 degree angle it mostly shows 0 m. It only shows some value while it is in the ground but once flying it is completely 0 m in withing +- 30 deg angle range. After several try I decided to perform the calibration process again, and after the process I found same result. I also tried to change the CP_dist from QGC parameter ranging from 0.5- 10 meter but it does not seems to work anymore. Then i realized I have made a big mistake by not saving the default calibration.stereo.xml file for calibration and then I found the sample calibration file in the voxl-camera-calibration git folder which then I replaced it with the calibrated file. Once this process is complete I then fly the Modal AI m500 and it worked almost as before but the CP_dist value of 1.5 m seems now too much and we reduced it to 0.8 meter. This the sample calibration.xml snippet from the git hub link: [image: 1622038820107-1362c676-a1ef-4bae-85b4-ee0e6799befa-image.png] I also noticed the focal_length_x and y for calibrated camera is ranging from 488-490 for both camera value which is much higher value after the calibration process compared with the sample calibration file shown above. Please let us know what can be done to correct this calibration problem or at least how I can obtain the default factory calibration data from modal ai.