Hi guys, I’m having big problems using Hobbywing Uavcan/Dronecan combo on PX4. I already use these products successfully on ardupiot but on px4 they seem not to be “compatible”. Do you have experience on this? This seems strange to me since the “CAN” products for pixhawk normally work well on both forks.
Good morning guys, little update on the situation.
It appears that Hobbywing Dronecan/Uavcan ESCs are not yet supported by px4. I honestly thought that since they were “ESC Dronecan” they would be supported by any device piloting Dronecan… but this is not the case.
To confirm this, Hobbywing itself responded that it is in contact with the PX4 community (I hope it’s true) to resolve the problem.
Since these motor/esc/led combos are very widespread and functional, it is a real shame that they cannot yet be used on px4 in CAN, but only via PWM.
If any community developers read this post can they provide their version of events?
When I visited the Hobbywing booth at the AUVSI XPONENTIAL exhibition in San Diego last month, I got the same answer. They were not ready for DroneCAN yet.
Instead, T-motor has unveiled a new FOC ESC lineup that supports DroneCAN. They say they’ll start selling in a few months.
And I’m using the ESC of KDE. Some of the KDE products have updated firmware to support DroneCAN since last year. There were minor issues in updating the firmware, but the new firmware worked fine anyway and succeeded in flying the 50kg hexacopter VTOL UAV.
In the case of servo, my best experience was using MKS’ DroneCAN series.
Hitec also sells the DroneCAN servo. But I remember the last time I used it two years ago, Hitec’s DroneCAN servo was only supporting part of the DroneCAN actuator message.
The worst experience was Currawong. They’re just slightly remodeling MKS’ servo and inscribing their logo and selling it for several times more expensive.
Currawong has been continuously promoting they can support DroneCAN for several years. However, after we purchased Currawong servos, they kept telling us to wait, saying that the firmware supporting DroneCAN was still under development. And when Currawong released new firmware earlier this year, they said that new servos would need to be purchased to use DroneCAN. They told me the reason of that is because the processors in the previous servos did not have enough flash memory to upload the new firmware.
Thanks for the detailed report.
Hobbywing, upon request, releases a good part of the documentation relating to their protocol. I said “their protocol” because it is clear that it is a “hobbywing can”.
It would be fantastic to know if at this moment someone is working on the integration of Hobbywing CAN on px4, our team would be able to develop this thing… but the doubt is that of starting a long and demanding job only to then discover that someone else already working on it… and therefore it would be a double job.
Reading this and seeing someone integrating Hobbywing XRotor Pro H200A 24S BLDC over DroneCAN with PX4 I have to say they generally work but Hobbywing seems to have misinterpreted the DroneCAN standard and implemented it their own way ignoring clear definitions. The result of this I’ve seen is that commands work but telemetry is 1 indexed instead of 0 indexed and hence off by one. I would love the manufacturer to read the standard they are implementing and not ignore it and if they did then fix it. But since people are shifting it to PX4 doesn’t support Hobbbywing I guess the expectation it that we know about every misinterpretation of the standard and implement quirks to work around it. I might just do that but I have to say this is not a reasonable workflow of “supporting” a standard.
It says `zero is first ESC. This should match the index into the cmd[ ] array in RawCommand`:
Agreed. And unless there’s a way to determine which variant of hobby wing firmware is in use, you can’t know if you should use 0 indexing or 1 indexing. I’ve reviewed the hobby wing PRs in the past and left them open, it’s my opinion that hobby wing needs to fix their firmware and provide an update to their users before we accept support and commit to carrying the hack.