|
Status
Feb 6, 2014 6:26:45 GMT -5
Post by Frisbone on Feb 6, 2014 6:26:45 GMT -5
I'd be surprised if you lost them - at worst they are still stashed. Git is built around branching and merging branches. If you keep everything in the master branch the process I tend to use is:
Commit (local) - commits your updates to your local repo Pull (remote) - merges in any updates from the remote Push (remote) - sends back your updates to the remote
You can't reverse the push/pull order because remote will refuse since you are out of date.
But so long as the merges aren't conflicting - it should work fine.
|
|
|
Status
Sept 1, 2015 17:51:37 GMT -5
Post by Frisbone on Sept 1, 2015 17:51:37 GMT -5
So good news today. I finally got some stability out of the RTIMULib library working with the 9150. I got all the original code back in and had tracking angles sent to the java app that shows pretty good relative position - good reason to believe that my delta angles will be accurate going into the joystick conversion routine.
Next step is to resurrect the xbox and COD and see if I can fine tune the conversion (but first of course I'll need to make sure everything that worked before still works).
|
|
|
Status
Sept 3, 2015 20:28:36 GMT -5
Post by Frisbone on Sept 3, 2015 20:28:36 GMT -5
I ran into a problem today once I tried to enable PWM output for the joystick. It seemed to deeply interfere with the separate thread polling the IMU (supposedly at a higher priority). I surmised that all that needed to happen was for the PWM commands to go out when a sampling needed to occur and perhaps that was enough time for a buffer overflow (FIFO overflow) to muck everything up.
I decided I needed to implement interrupt handling for the 9150. Unfortunately I had troubles there. I added the correct register changes in the libraries initialization and then added an isDataReady() imu method so I can clear the int status register - but the ISR never got triggered.
I was able to verify that the 50us pulse was being triggered every 20000us (50hz) on the o-scope. Must be a problem with the interrupt configuration on the PI side.
|
|
|
Status
Sept 8, 2015 19:11:52 GMT -5
Post by Frisbone on Sept 8, 2015 19:11:52 GMT -5
Yeah! I had some really good progress today. ISR working now (old wiringPi library malfunctioning) and I have it working in lock-step with PWM output using a Mutex. PWM output is queued also and output all at once for both axes so there is no delay. The Angular rate conversion to PWM output needs to be fine tuned quite a bit though. I had to change the coefficients to get left/right looking proportional but the amount of rate change didn't make sense for the amount of movement. Up/down (declination) was in favor of up substantially - will probably need an independent equation for that axis.
All in all a good night!
|
|
|
Status
Sept 14, 2015 5:05:43 GMT -5
Post by Frisbone on Sept 14, 2015 5:05:43 GMT -5
As of last night we have pretty good movement on both axis and enough progress to move on to other areas.
|
|
|
Status
Sept 15, 2015 19:46:27 GMT -5
Post by Frisbone on Sept 15, 2015 19:46:27 GMT -5
Got jump/aim sort of working tonight. Need to work on thresholds so it takes less effort and is more reliable.
|
|