Where to connect ppm reciever to the board?

[Max]

I've come to discover that we may be triggering our PPM pulse samples based on the wrong edge (rising vs. falling).

Can anyone post PPM sum traces from their Rx(s)? I'm looking for if the signal is positive (normally low) or inverted (normally high). These FrSky captures (below) are "inverted," at least for our purposes (purple, green, and orange lines). The individual channels shown in the first capture would be considered positive, I believe.

Most of the captures I've found online from Rxs with PPM sum output are also inverted (including the Ardu/Paparazzi PPM encoder). The current AQ code is triggering based on a positive signal. When I invert it in the code, the PPM decoding of my FrSky Rxs improves significantly (way fewer invalid pulses and no dropped frames). It does work fine the current way, but the results are much cleaner when triggering on the correct edge. And it does resolve the issue I was having with the 27ms FrSky firmware.

Thanks,
-Max

FrSky D8R-XP PPM sum 27ms frame on top, with direct output from some channels following.

Comparing FrSky D8R-XP 27ms and 18ms PPM sum frames; 2 Rxs, same Tx (channels 6-8 not sampled).

[SuperSharpShooter]

Hi Max, where do you invert the triggering in the code? Now that I've gotten the first compiled firmware flashed with signaling, I figure I may as well get this changed too if it will improve the performance with the FRSKY receiver.

[Max]

In the latest code there is a macro at the top of ppm.h to control the trigger direction. But I would hold off on messing with that... I now think my problems were unrelated. I spent a lot of time debugging PPM issues... seems like it's pretty easy to introduce noise into the stream (as others have found as well). The latest repo code should further help reveal any signal problems and help with diagnostics. If your radio_quality is staying steady and the new RADIO_ERROR_COUNT is low, you're fine (the error count can actually be viewed with the latest version of QGC, but it is mislabeled in the graph viewer as "LOG_STACK_FREE_INIT").

-Max

[SuperSharpShooter]

Ok , yeah my radio quality stays at 100, so I won't touch it for now.

Thanks

[Max]

Glenn, if you have the time, give the latest rev from the repo a try and check how your quality and error count looks, just out of curiosity. It should be much more sensitive to detecting any signal issues. Not saying you have any... and a few glitches or errors isn't a problem.

-Max

[SuperSharpShooter]

So that's compiling from the latest source? I can give that a go tomorrow and check it out. I haven't flown more than about 30 metres away yet, so now that my copter is handling nicely (it really is very nice seeing it sit there in DVH compared to how much drift there is in loiter with an APM!) I feel better about giving it a push.