Update drv_pwm.c #153
Update drv_pwm.c #153
Conversation
Detection & fix for the "8 channels in 18ms Frsky problem" see http://diydrones.com/profiles/blogs/why-frsky-cppm-signal-is-so-disappointing . Note: failsafeCheck not altered!
| bool sync = diff > 2700; // rcgroups.com/forums/showpost.php?p=21996147&postcount=3960 "So, if you use 2.5ms or higher as being the reset for the PPM stream start, you will be fine. I use 2.7ms just to be safe." | ||
| last = newval; | ||
|
|
||
| if(frsky_18patch) { |
|
If you can correct the above stuff, I'll merge this. |
|
I will do another pull request with a better formated version now. |
Reformated the code, tested it again - works. The failsafeCheck is not changed and probably needs further attention from the BF folks.
|
oK! On Fri, Aug 1, 2014 at 5:37 AM, Crashpilot1000 notifications@github.com
|
|
Just doing a shorter version it is as good but cleaner. Have to do some testing. Will commit it when done. |
|
OK On Fri, Aug 1, 2014 at 11:27 AM, Crashpilot1000 notifications@github.com
|
Shorter and working.
|
That is it from my side. Plz test it out as well :) |
Besides the Frsky 18ms fix it also verifies the channelnumbers, that will most probably also fix some hiccups some ppl see. I can not verify this because I don't see those hiccups with my gear. BF failsafeCheck logic unchanged may need attention.
|
The last commit also does a better check on the channel order, there is a good chance it fixes those hiccups some ppl see (can not check it, because no hiccups with my gear). |
fiendie
force-pushed
the
master
branch
2 times, most recently
from
0e8347c
to
fed0125
Compare
|
Can one of the admins verify this patch? |
|
Just a whitespace issue. We could merge this. |
|
I do not like this patch. It adds a workaround to a specific problem that has a proper fix (27ms firmware). Also as it generates 'sync' from thin air so it is possible to drop out of sync when the input is without sync for a longer period of time due to glitch or what ever. As per the channels order, a more simple way is to just set chan=255 when an invalid pulse is detected causing the rest of the frame being ignored. |
|
To be even more safe against glitches the whole frame should be ignored if any of the channels fall out of the 750-2250 range, setting chan to 255 would just ignore input after the glitch. For better resiliency the captures should be saved to 'temporary' table and only copied to captures when next sync is detected with proper number of channels sampled. This approach will also ignore the 'corrupt frames' from a 18ms frsky receiver. |
Detection & fix for the "8 channels in 18ms Frsky problem" see http://diydrones.com/profiles/blogs/why-frsky-cppm-signal-is-so-disappointing . Note: failsafeCheck not altered!