Txs for the report of DShot not working on the Cube/Pixhawk2. I guess you’re loading the ChibiOS and have FRAME_TYPE = 2 for hexacopter? I’ve tried setting these two parameters on a Cube and it seems to be working OK although I don’t have DSHot ESCs to test it with - still, it’s at least not freezing up.
Could you provide a log file or parameter file from the board so that I can try and reproduce it?
Hi Randy.
14 days ago I wrote here the following. Is there a chance of getting my Rangefinders working again. I like the connection to the AUX because it never failed.
14 days ago:
I have loaded on both my Helicopters Copter 3.6.0 rc10 the ChibioOS version. On both Helis with identical FC Pixhawk 2.1 cube and LIDAR-Lite 3 Laser Rangefinder. In both Helicopters the Rangefinder is not working with ChibioOS. But loading back NuttX Copter 3.6.0 rc8,rc9,or rc10 the Laser Rangefinders are working. This is repeatable.
I must let you know I do not connect my Lasers with the I2C cables.
It works perfectly on both Helis but with NuttX only.
Hi
Many thx for your help. I will be able to fix it
Regards
Hi Randy,
Thanks for your response.
Yes its Chibios hexa and Frame_type 1
I attached my settings file.
The error occuers on rc11 too, and it does not matter if the escs are connected to the Pixhawk or not.
I made the changes described here : (i currently do not use telemetry)
http://ardupilot.org/copter/docs/common-dshot.html
Hexa_settings.param (14.2 KB)
Thank-you! Will investigate the SD card.
i think i have replicated your pulsation exactly when i dropped PSC_ACCZ_FILT value to 10. cannot recall now if it was done in combination with INS_ACCEL_Filter set to 10 as well or not. but i think it was not driven by INS_ACCEL_Filter value, only the first one did it.
when restored to 40 it smoothed it, and now i will test it at 80 - tomorrow.
when it was set to 10 it started bumping up model up and down, very weird. try it.
INS_ACCEL_Filter i also have at 40. andi have a feel the smaller model is the higher that value needs to be.
now, i see a bunch more of filters in the PSC section and not sure what those need be set to, but one above was indeed the one that made it to pulsate up and down.
Daniel,
I’ve tried reproducing the issue on my Cube using mostly the same parameters as in the Hexa_settings.param file but sadly I can’t. Obviously this doesn’t mean there isn’t a problem, just that i can’t easily reproduce it.
To help remove some other causes, could you try disabling precision landing and the range finders? PLND_ENABLED = 0, RNGFND_TYPE = 0, RNGFND2_TYPE = 0.
@Daniel_Met I’m the author of the DShot code. I’d like to look into this, but I need your full parameter list, or even better a tlog or DF log.
Cheers, Tridge
@Daniel_Met one thing to check carefully is what your BRD_PWM_COUNT is set to. It will need to be 6 for the setup you describe.
Hi @Daniel_Met, I’ve now reproduced the DShot hang bug, and I’ve put a fix here:
I have put a pre-built firmware for you to test here:
http://uav.tridgell.net/DanielMet/copter-3.6-beta-dshot-fix-cubeblack.apj
this fixes the issue for me, but I’d appreciate confirmation from you.
Thanks for reporting this bug!
Cheers, Tridge
Hi Tridge,
yes this seems to have fixed the issue.
I’ll leave the pixhawk running in the background a bit to make sure it doesn’t occur again.
Thanks,
Daniel
edit
yes it seems this solved the Problem
Thank you for the quick fix
Hi Tridge, the issue you found - could it be same thing causing unexpected twitches? I have a feel it is code, Leonard looked at log at says it is external, hardware or sync loss and i think hardware is most likely not the reason. Could that bug you found be causing any pauses in the dshot stream to escs?
and, that fix is not in the master yet, is it correct?
Paul,
The fix has just gone into master so it will probably be a few hours before it is available using the magical Ctrl-Q on MP’s Install Firmware screen. Until then, I’m guessing, but I think it should be safe to try the binary that Tridge has posted above to see if it resolves the issues you’ve seen.
We’re planning -rc12 to go out with this DShot fix quite soon. Probably Friday.
Thx, i got fix into the code, issue i had was indeed hardware - an interference of an unknown source into a single less than a 2" signal wire from controller to ESC. After replacement of that wire with a shielded wire all issues were gone. I guess it is now a given that DSHOT requires coax cables or shielded multi conductor cables for FC to ESC connection. It is new.
What is also odd - i have had this same model on this same drivetrain on same FC on single wire on betaflight, and it flew with no visible glitches. go figure.