There is a preview on GIT right now. I flew a version of this
The PWM output has been set to 400hz (to counter the low pass filter in most Turnigy PWMs)
The DCM's Roll and Pitch gains were lowered to .03 (recommendation of Hein Hollander)
Mavlink has gotten a re-work for performance and memory savings.
Added a scaling factor to camera Roll And Pitch CAM_P_G, CAM_R_G in the params list.
Fixed some bad PID values for Alt hold that were giving folks trouble.
Exiting a WP for another is faster and more controlled now.
Added user hooks for those who want to execute their own code inline.
Increased loiter speed to center when the copter is flown > 400cm (now 800cm).
Loiter in action.
I'm waiting for test.....
What does instant_PWM do ?
Will the 400hz PWM ouput also be compatible with supersimple card programmable esc's ?
400 hz works slightly better with cheapo Turnigy style ESCs that most people use, than instant PWM.
There is a low pass filter on these and the higher rate out PWM pushes the throttle changes past the filter faster. Instant PWM is more nimble, but it's tied to the main loop speed which doesn't work quite as well for these ESCs.
Jason is it possible the 400hz output could cause more of those intermittent clicks/ pops which are present in the motors at lower RPM. The reason I ask is I am noticing this effect more now with the jdrones 30A ESC's than when I was on 2.0.45 The secondary effect seems to be more interference noise in the sonar signal. Its probably not connected but I needed to ask.
Maybe, you can always turn it back to instant PWM to see.
The clicks are probably sonar noise getting into the control loop. Try to add a capacitor to the Sonar and shield the cable,
It was a great pleasure flying today using 2.0.48 with MP 1.0.82 on a custom Quad frame.
The proportional gain on roll was set to 2.0. Pitch the same. The Loiter was lowered to 1.5. The Yaw integral was lowered to .001. The rest factory (Jason) default :-). Holding a loiter waypoint still has the ability to cycle into a circling oscillation. It happens because of the frame attempting to seek center. This kind of reminds me of a neutralization circuit on a amplifier that needs to be be built. What happens is the copter at times can jump into a circling loop depending on external forces and how it approaches the center. It may be possible to take one circle or even a half circle of flight to determine the center to offset the circling. What was also interesting is the GPS coordination. If the copter was moved at least 50 meters to a new waypoint there was a 50 50 chance it would not circle. Going back to a GPS orgin that circled would allow circling again. I think this is a condition that is a inherent characteristic using a copter with waypoints. This also was the most smoothest flight I have ever done. The altitude held better with the sonar. Letting go of the mode 2 right thumb brought the copter to a near standstill. I will make a video of this. Really neat. Attached is the flight log.
I agree - .48 really the best ever (stock quad frame). Whataever you did to loiter from .47 was a big improvement - now it goes into a smaller less agressive circle and almost seemed to be trying to hone in on waypoint. Under .47 it would go into a big fast circle away from waypoint.
My only minor whine is the large yaw deadband, it seems to take some retraining of my finger but a little smaller would be cool.
It flies like a dream even in crappy wind. Jason's magic gets better and better.
Just tested this earlier on 2 non-ardu frame / not default PIDs
- stable was good but with sudden wind/breeze the quad tends to oscillate a bit and the quad becomes sluggish. If I turn up either rate or stab the quad will wobble a bit and doesn't stay level. PID on windy and calm seems to have build a gap
- Loiter goes into a circle pattern like how the arduplane suppose to do when entering loiter but with 3m diameter only and it moves fast
-circle is still not working as expected, it will correct yaw but it seems to cause more error and unable to follow the circle. I hope it will do the same as what loiter is showing (now) but correct the yaw to point center
- AH hold quite hard to tune (Baro only), if I understand correct its a PIPI alt_hold/throttle_hold but still playing around
-tried ADC 1 but the quad gets worst on windy condition
all this is being compared from v47. Will do more testing on the weekend
I got loiter working today for the first time using 2.0.48. I used default PIDs for loiter/nav.
The conditions were quite gusty and it was awesome to watch it tilt into the wind to maintain position.
After a while I walked up and under the quad and sure enough it popped up to maintain altitude.
Then I turned around and walked back. A couple of guys were watching and their jaws dropped when they saw that I had my back to the quad. After I landed I explained to them that it was in GPS position hold.
Very cool stuff! Good work Jason and the diydrones community!
.48 is very smooth. Minor tweaks to gains for my non-standard frame. Sonar is awesome, but baro is a bit jumpy for me. Loiter is still okay, but I need to tune it- very aggressive circles yet.
I still need to turn up telemetry rates in planner to see if the Mavlink rework will let me have decent data throughput and still be able to send tuning parameters reliably. In the past, if I had telemetry rates turned to a usable level, parameters would fail to send properly (APC220, 56k serial, 19.2k radio data rate).