Problems with PPM going out of control

[apejovic]

After having my rev1 board sitting in a box for many months, finally I have found some time and decided to put it on our “labquad” prior of using it for one of my bigger octos.

I have loaded fw 6.7 b1400 and done the static calibrations which had beautiful curves, much better than my precious ones done last year with my first rev 0.

Everything was ready for the calibso, I mounted the board on the quad, setup the limits on the radio and checked on the GUI that radio connections was fine, all OK. Quad had props on (normally I remove them)) but I simply wanted to test if the board was arming and give some throttle just to check if props directions where correct, arm give some throttle and everything seemed fine…..until all controls started to move randomly making the quad to take-off and hit the wall. Even after catching it and throttle completely cut the motors were still accelerating. There was some kind of issue with ppm.

I decided to take out the board and do some tests with the AQ and my radio on the bench with everything unplugged and minimizing any possible interference with ESC´S and other devices, but unfortunately sooner or latter I was always able to reproduce the problem, at some point all channels get out of control.

I did the following changes but unfortunately the problem persists.

1-Tried two different radios, x9 with original fw and another x9 with er9x (both have and are used for over a year on a daily basis with other systems and they work great)

2-Tried two different TX modules (FRSKY) and two different RX (D8R-XP) as I did not flash them to bypass the 18ms ppm issue I always used a maximum of only 6 channels.

3. -Checked and rechecked pins soldering on port 14

4. - Tried different ppm/servo leads

5. - Used 5v output from AQ to power RX and also tried an external bec

6. -For connecting to PC I have used a 3dr 433 radio modem but also done tests with my FTDI.

With any of the previous combinations the problem always appears, and it is curious how sometimes ppm input simply does not work.


Here are two logs showing the problem, I attach both a jpg and the entire log.

log 24

Complete log 84mb

Note that after 20000 radio was actually off and rx detected it (failsafe deactivated) so no signal was sent to AQ

Here there´s the complet log 16mb

I hope that somebody could help me, and thank you all in advance.

[Max]

Hi apejovic,

Do you have any other way to verify the PPM output from the Rx, or have a logic analyzer or scope to examine the signal? The problem sounds very strange. Quite a few people use PPM and a lot of them with the exact Rx you describe (including myself). It should work even with the older 18ms Rx firmware.

I did have one board which had noticeable PPM signal quality problems (though not nearly as bad as your logs show and no loss of control). The problem turned out to be a bad solder on the serial ("FTDI") connector ground pin. I suggest very carefully re-checking all the solder joints and wiring for loose connections. Not sure what else to suggest. I suppose it could be a hardware issue, if everything else is ruled out.

You could also try an X8R Rx with SBus output instead... but I think your PPM problem is actually an indicator that something else isn't right.

Cheers,
-Max

[apejovic]

Thanks again Max, you´ve been alwayso trying to help me in my AQ adventure.

You know my experience with multirotors and the way I work fly.

I did check all solderings and will recheck again but the curious thing is that actually these logs were taken with the board on the bench, so there´s really very little I can do apart of checking again port 14 and FTDI grounds.

I do not have a logic analyer but can check ppm output on one of my YS-X6 or any of the AlexMos boards I´ve got around, if the problem persists it will discard the AQ but the curious thing is that I did the tests with different receivers, rx etc etc.

As for the X8R that is maybe a solution for me but I have no experience with sbus, does the rx come with the cable that goes to AQ or should I buy it somewhere else, if so where?

Thanks

[Max]

Yea, I've seen you do some amazing video stuff with inferior hardware, so I'd love to see you flying an AQ. I'm sorry you seem to hit these stumbling blocks though... maybe it's the disgustingly nice tropical climate you live in.

Anyway, I agree that with the FC out on the bench with minimal wiring should be the best test. That is how I eventually tracked down my bad solder issue as well. In fact I had even re-soldered that connection once already, but it still wasn't good. For some reason I've had trouble soldering to the AQ ground plane on both my boards, especially at the serial and main power connectors. Probably because it's big.

With an LA you could see what was going on while actually connected to AQ, which would probably be the quickest way to find the cause. I have a feeling it's not the Rx/Tx.

I could send you a firmware build using a different M port for PPM capture... could be something wrong with that input I suppose. Do you have any others with 5v already jumpered? Though you could power the Rx from any 5v.

Speaking of which, have you tried powering Rx via external source (V. reg/bec)? Also, as a curiosity, does anything change if you only hook up the signal line between Rx and AQ (no gnd or pwr)? That's not a good way to run it for flying, but it might tell us something.

For SBus you need a simple serial signal inverter (transistor and 2 resistors), though the trickiest part can be finding the little JST connector for the AQ side. Easiest is to find a pre-crimped pigtail, though I've managed to make a couple of my own as well (using pins/housing from digikey). You can also find the inverters pretty cheap online, depending on shipping. The inverter is described here: http://autoquad.org/wiki/wiki/autoquad- ... taba-S-Bus You have to power the Rx with 5v from somewhere else, as described in the blue note.

If it was practical to mail things back/forth I'd check it out for you, but... Of course in your case I'd also be willing to do some on-site consulting!

-Max