Joystick controller using pc sound card [WORKING]

I've written a Linux program that takes the input of a joystick (Logitech G940 in my case) and outputs a PPM signal from the sound card that is used to drive a transmitter (devo8 in my case).

The program uses the Joystick++ library from Drew Noakes (shipped with the program) to read data from joystick, and the portaudio library (available in most Linux distributions) to drive the sound card.

You may need to amplify the signal coming out from your sound card. I used an old HI-FI amplifier. Start with a very low volume and go up slowly until the TX sees the signal.

Here are some pictures of the setup: imgur.com/a/zlXnR
And here a video of me flying the Walkera miniCP:

The readme file contains some instruction to setup and run the program. It is not stupid-proof but shouldn't be too hard to use.

ORIGINAL MESSAGE

I'm looking forward for the reversed usb-hid, in the meanwhile I've written a small C++ program in linux that takes input from a joystick and modulates a PPM signal through the sound card.

I captured the signal with another pc, measured it and looks fine, but when I try to send it to my devo8 nothing happens. I'm not sure if there's an issue with the shaping of the signal, the setting of the devo or maybe it's just my sound card that cannot output enough voltage...

Any experience?

Here is the signal captured with the second pc (voltage scale is random: it depends on the mic gain)

I attached to this post the full program. A precompiled executable for x86_64 architecture is included together with the sources. To run it you'll have to have a Linux OS with Gnuplot and ALSA, everything should be clarified in the README file.

The PPM structure is easily modifiable in the joy2ppm.cpp source file. Let me know if you can take something out of this! I suspect that the signal have to be amplified (or the TX modded) in order to be able to drive the TX.

I guess that the part that you need is this:


I took care to match the centerPW and deltaPW between my code and the Devo config, however I wasn't sure about the notch (400us?) and the total pulse length (22500us?). I'm also not sure if the filling part of the pulse should be up or down (I tried both but no luck).

When I switch to MIX2 the TX doesn't take input from its sticks (I think it's correct) but it also doesn't react to the ppm modulation. Unfortunately I don't have a real oscilloscope available so I cannot do precise measures, I'm quite confident with the time structure, but I have no clue about the voltage.

I still didn't get the communication with the TX but attached there is a much improved version of the program. It contains a couple of scripts that should make the compilation and the run trivial (executables for LINUX 64 bits are still shipped so there is no need to compile on this system). I also moved all the configuration to a .conf file so that every parameter of the signal generation and joystick mapping can be easily accessed. The README file contains updated instructions.

I definitely got it right. So it must be a voltage problem. I think that my sound card output is too small to trigger the acquisition in the TX. Do you know what is the required voltage gap? I may try to put in between an amplifier...

It would be great if you could try the code, it should be pretty straightforward to run it. Please report me any issue and I'll provide assistance ASAP.

Yep, I use the mono cable which was supplied with the TX. In the next days I'll write a similar program to produce a sine wave so that I can measure the output voltage with the simple tester that I own and eventually try to amplify it. Thanks for the assistance!


PS. It is quite like that my signal is below the threshold: en.wikipedia.org/wiki/Line_level

And I got it! The level was too low. I put in between an old HI-FI stereo amplifier and it worked. I had to fine tune the volume so I don't think that the TX accepts a wide range of voltages. However starting quite low and going up slowly it is quite straightforward. I also had to fix some math in the code.

Now the major problem is the delay between a joystick movement and the sound modulation. I guess I should be able to remove it redesigning the code with better libraries. I'll have a look if I can employ something which is also multiplatform.

Updated the program, now uses portaudio. It works very well! You find everything in the first post.