Just posted this version which has an increased performance DCM implementation. This allows the main loop to run 50Hz faster without increasing the CPU load. I've flown this for 30 minutes and it's an improvement in every respect.
Circle mode now seems to be working just fine. This should be the last beta! enjoy testing!
Jason
Here is the Link to a download:
Replies
I got two flights in with my quad yesterday afternoon. The first flight I used the same PIDs' I was using on 2.0.40. The flight went extremely well. ALT Hold, Loiter, and RTL worked flawlessly. It was a bit breezy so the second flight I increased the rates a little to compensate, which resulted in a crash:( The crashed happened when I flipped the quad into ALT hold. It began to oscillate, flipped, and hit the ground. Easy fix, arms broke where they are supposed to, no other damage to frame or body. Replace short inner ferrel tube peices and fly tomorrow. Camera mount survived, GoPro clip that holds it didn't, buy new one today. I have attached the log file to veiw.
I dont know if it was due to the rate increase, or a bug. I will say that when I flew the Y6 afterwards there were times when it would want to fly off away from me as if headed for a waypoint (Auto not configured in modes). It got me wandering if there is a bug in ALT Hold. flew my Y6 next with no problems except for ALT hold, not every time, but did notice when I initially engaged ALT Hold it would want to veer off. I would flip to Stabilize then back into ALT Hold and no Problem, perfect hold. RTL on the Y6 works better than it did in 2.0.40. It came back like it was on rails, performance that Jason claimed. I will post some video of the Quad crash, and Y6 flight shortly.
One more question if I may...
I'm planning on implementing a led lightshow on my ardu hexa. For that I'd like to move the update_motor_lights() function to the 50hz loop and greatly expand it to make colored ledstrips on my arms flash in different sequences and speeds.
Would that hurt performance of the main loop or is it no problem doing this ?
I'm not really sure I understand how the loops are timed. As I see it, the main loops run at their frequencys (250hz, 50hz, 10hz, 3.33 hz, 1hz) and there are still some cpu cycles to spare.
Or am I wrong and is the mainloop just running as fast as possible and anything I add will slow down the main loop ?
Could it be that the stick center drift is a little worse again then before ? I have filterpads soldered, balanced props and brain on a foam pad. But when I liftoff I can leave the sticks centered, but after fighting the wind a little, center is off. I'v always had this more or less, but seemed a tad worse now (or could be because it was a little windy when testing ?)
Wasn't there a way to speed up the correction ? How was it done ?
Otherwise I'm quite happy with 42 so far. Eager for the wind to go down a bit so I can test some more.
Great work, really nice to see arducopter approaching maturity :-)
I have one update left which deals with Guided mode. It's done and posted to GIT. So I'm really to release it now.
Jason
Jason, i still need to reverse Ch2 inside parameters to get the right motor setup on a hexa X frame. Do you think it does come from an error in the motor_hexa.pde file or an inversion elsewhere ?
Do you need to reverse Ch2 for quads ?
Very nice Jason, just out of beta when i've finished my frame.
I've just added the missing motor tests for octa quad and octa V. Upped slightly motor_out values as well in the test because it was a bit low for Pentium ESCs.
Just loaded 2.0.42 and repeated my 2.0.41 tests on my standard frame. Flys great - and I got a nice loiter at 30 feet over the house and trees in a 5 mph wind. It loitered for about 2 minutes and I do need some minor tuning, but it was really as good as I've ever seen it. Altitude hold seems rock solid with minimal pulsing at sonar and baro altitudes. Checking my logs, sonar alt and baro alt are dead on - no weirdness, no noise. Did a 360 on simple mode and that seemed good also.
The feeling in stabilize is awesome and yaw seems even snapier. I like it!
I've only flown 1 flight, but this version is sweet.
Thanks Jason!
Does this means the entire fast loop is now running at 250Hz, including all the stabilization code? Are you anticipating any effects on PID values due to this change?
Sounds good!
Jason, what about AUTO mode? Any improvements?
-
1
-
2
of 2 Next