Polarity Issue on Trainer Port via H-Beat Pin?

[MarenB]

Hi,

there seems to be an issue with the PPM-in using th heratbeat pin in the module bay (X9E at least).

I was using the famous DIY-headtracker for a while on my old Taranis and it worked fine. On that radio I connected it to the trainer jack, which was internally routed to heartbeat pin (at that time unused by firmware). The tracker was then plugged into the module bay.

I built another one to plug into my X9E, but this radio does not detect the correct PPM signals. it seems to trigger on the death time between two pulses. Between two PPM frames, the signal is high. The PPM pulses are also high, just the breaks are low:
http://www.rcgroups.com/forums/showatt.php?attachmentid=4959223&stc=1

Is there a chance, that this signal shape causes a false auto-detect of PPM-in polarity?

[projectkk2glider]

We had a change in the algorithm that decides if incoming PPM signal is valid or not. Perhaps this is causing your problem. Is logical 1 state between frames a valid signal or not? Looks funny to me.

[projectkk2glider]

Did you by any change tried the same with the plain Taranis and OpenTX 2.1.x series where the same algorithm is used. If the problem is duplicated there, then at least we can rule out any hardware problem on X9E.

[kilrah]

PPM can have either polarity and it shouldn't have any influence, the Taranis will happily take both.

[projectkk2glider]

True, but he polarity of pulses is in his case positive and the idle state is also positive!

[MarenB]

Yes kilrah - but in my point of view, this signal shape is not consistant. It's high between the frames and high for a pulse - that let's the Taranis struggle with it, perhaps.

My old Taranis is still on 2.0.17. I need to make an adaptor harness to try the new headtracker on that one, but I am pretty sure it will work. Both headtrackers use equal hardware and firmware, only the interface to the radio is different. I also checked with an oscilloscope, the signal is as intended (as on the picture I attached above) and equal on both units.
The X9E detects a signal, "TR1" i.e. in the mixer line becomes bold, I get confirmation by haptic and the mixer output jumps to something close above 0.0%. However, it doesn't change that when I move the HT, that is what makes me believe the radio might trigger on the gap between the pulses.

Btw. when was that change made? I can't recall it 100%, but I might remember, that it worked once evene wit the X9E (a couple of weeks ago).

Oh wait - I had a false wiring on my HT supply voltage, the arduino was provided with full battery power (replaced the Arduino afterwards, of course). Is there a chance to check wether the X9E got damaged by that?

[kilrah]

Yes kilrah - but in my point of view, this signal shape is not consistant. It's high between the frames and high for a pulse

No, it's high idle and low for the pulses, perfectly normal.

[MarenB]

No, it's high for the pulses as well! Low are the breaks between the pulses, that I confirmed by oscilloscope.

[kilrah]

Then you're not referring to the correct thing when you mean "pulses". The "pulses" are the channel separators, which all have the same length, and are low in your capture. The variable channel durations obviously have the same level as the "idle" frame separation level.

http://api.ning.com/files/KWePTtollo4*Z6ijOwVqLAtitTvNj5O*1hu7jQtEG5-w0qpo3yoV5eE1CZQX4MqV1*oNoPVPNe66wWgeSaeJDQ__/PPMframeafterPPMencoder.jpg

[projectkk2glider]

@kilrah again you are right, his signal is just inverted nothing else. So do we handle inverted PPM input in 2.1?

[projectkk2glider]

@MarenB do you get voice announcement when the PPM signal is detected? Do you also get trainer lost announcement? (you nedd files "trainko.wav" and "trainok.wav" in SOUNDS/en/SYSTEM folder)

[MarenB]

I get a haptic alarm in both cases. Will try using voice file.

Thanks guys, have a couple of items to test now. Hopefully I did not fry the input by overvoltage. Should be possible to confirm or deny this by testing S.Bus-Traininer-in.

Will return after I got more data.

[projectkk2glider]

Hopefully I did not fry the input by overvoltage

Hardly, the input is protected by the reversed diode so it only triggers on low voltage.

[projectkk2glider]

If you get many trainer OK and trainer KO messages, then the new PPM validity algorithm is causing this.

[LapinFou]

Maybe it is not the same issue, but I played with a X8R plugged in the JR bay ("Master/SBUS Module" mode) in order to achieve a wireless trainer link. When I switched from one model to another on the master radio, I got many trainer OK and trainer KO messages. After a power off/on cycle on the master radio, it came back to normal.
The same behavior happened when I switched from one model to another on the student radio (both models use the same RX ID, since the X8R is inside the master JR radio bay).

[MarenB]

Hi there,

finally managed to test with "TrainerOK" and "TrainerKO" sounds and indeed, that's what I hear. But just once any time I plug/unplug the headtracker, not multiple times.
Next test will be with this HT via an adaptor cable to my old Taranis D.

[LapinFou]

When I described my problem, I thought it was correlated to me problem. In fact a new ticket #2999 has been opened. So, please don't take in account my comment below.

[MarenB]

Gents,

thanks a lot for your comments on my issue. Finally, OpenTX 2.1.5 solved the problem :)

[theailer]

I'm on OpenTX 2.1.8 and I have this issue with the same headtracker.

[projectkk2glider]

I am not sure why @MarenB closed this issue, we haven't done anything in code. Perhaps he can explain how he made it work.

@theailer what are your symptoms, do you get "TrainerOK" and "TrainerKO" sounds? Did it work in previous OpenTX versions? Does it work with any other hardware? Please provide as much information as possible.

[theailer]

@projectkk2glider
Sorry for my brief comment.
It worked fine on my fly sky th9x with smartieparts board and er9x software. I'll check for the trainerKO and OK sounds this evening. I had it on mute when i tested yesterday and there were haptic feedback and I could also see on the channel display that the input halted.

[theailer]

I could try and flash 2.0.x if that would help debugging.

[kilrah]

The feature doesn't exist in 2.0, so no it wouldn't help...

[theailer]

Ok, I get the trainer recovered and trainer lost messages intermittently when moving my headtracker. Not all the time though but pretty much. It still works perfectly on my er9x th9x radio but not on the taranis plus.

[kilrah]

Can you confirm which port you're feeding the headtracker data to?

[theailer]

When looking at the back of the radio its the left one. The other with the smaller bezel is for headphones?

[kilrah]

Yes. But this issue is about using pin 2 in the module bay, not the trainer port...

[theailer]

Aha, sorry.

[theailer]

Sorry for messing it up.
Anyways, tried the headtracker again on my old flysky th9x and it does the same thing there. So it was the headtracker all along.