LibrePilot Forum

General Category => General Discussion => Topic started by: pj72 on April 20, 2018, 06:28:26 pm

Title: PPM+outputs Makes RX go into Failsafe or Failed to update one or more accessory
Post by: pj72 on April 20, 2018, 06:28:26 pm
Hi,

I am running r589. All looks good on the transmitter set up using PPM input. When I change to PPM+outputs for an Octo build the RX seems to go into failsafe or similar. The outputs seem to work. I can get all 8 motors to calibrate and spin, but no activity from the reciever.

Any ideas?

Thanks

Paul
Title: Re: PPM+outputs Makes RX go into Failsafe or Failed to update one or more accessory
Post by: f5soh on April 20, 2018, 08:34:43 pm
Hi Paul, welcome.
Which board do you use ?
Title: Re: PPM+outputs Makes RX go into Failsafe or Failed to update one or more accessory
Post by: pj72 on April 21, 2018, 12:07:30 am
Hi,
Its a Revolution Board.

Thanks

pj
Title: Re: PPM+outputs Makes RX go into Failsafe or Failed to update one or more accessory
Post by: f5soh on April 21, 2018, 12:54:17 am
Seems the issue is already solved in pios_servo.c (https://bitbucket.org/librepilot/librepilot/pull-requests/500/lp-584-fix-adc-on-servo-outputs/diff)

Please try flashing the firmware attached using the rescue method (https://librepilot.atlassian.net/wiki/spaces/LPDOC/pages/5669250/Firmware+Rescue).
Title: Re: PPM+outputs Makes RX go into Failsafe or Failed to update one or more accessory
Post by: pj72 on April 23, 2018, 07:46:17 pm
Nearly there. I used the rescue method and seemed to get somewhere.
I then experienced the following issues:

When the output is configured as PMW no issues.
When the output is configured as Dshot this also seems to work. (I have some issues with the 0-1000 channel input so not using this)
When using oneshot or multishot I experience the same Failsafe issue as earlier.

I noticed an issue with Bank5. If I moved output 7 in bank5 to output 9 in band 6 and leave bank 5 configured as PMW I think the issue goes away.

I think this is viable workaround for me, but it feels like there is still an underlying issue here.

Thanks

pj