This beta brings support for the new MediaTek GPS unit. The MTK works great and it's very light.
I've also added some bug fixes including:
- Auto Trim - (on by default) When you enter any flight mode from Manual, you'll save your current trim values. Great for trimming your plane in the air.
- FBW issues (turning too much in one direction) caused by incorrect or unset radio limit - fixed with new algorithm
- Radio Mixing - lowered the mix value for more responsive sticks
Let me know if you have issues, and I'll update the release version.
Jason
Comments
Can you zip up the code you flew and send it to me or post it here? I think you were missing one small fix.
here my report.
Testflight 8.7.2010
AP 2.6.2
Navgation.pde patch 4
Ardupilot_2_6.pde patch 1
IMU V2
Speedsensor
Total 4 Flights
General
Navigation.pde version 4 works well and smooth. See below.
1. Navigation good, does not go around but thru WP, direction clock or counterclock.
2. The loiter path is oval. See KML's below. 90° twist compared to 7.7.2010
3. RTL centers nice over home. Same shape and proportion as Loiter.
80m Radius good
60m Radius the kidney is hard to see in flight
40m Radius kidneyshape better visible
25m Radius kidneyshape well visible
Conclusion
Logical error in calculating the diam of loiter track.
BR Christof
APflight_2010_07_08_09_14_54.kml
ArduPilot_DataLog_2010_07_08_09_08_01.txt
APflight_2010_07_08_09_04_48.kml
ArduPilot_DataLog_2010_07_08_08_57_20.txt
GPS_MTK_IMU.pde
But tomorrow I will test 2.6.2 the latest first.
Christof
Mega will have upload capability when it's released.
The PID configurator is even a better idea. And please implement somehow the transmission by xbee. So it can be used to first generate the PIDconfig file and after to adjust the values without reconecting, rebooting ect.
After the testflights from last week, my pins on the AP board have no more goldplating left and soon they will be weared completly ;-)
But dont we need a database with some reference PID's according to type of airframe? I do not think that you can calculate these values, not even estimate without reference.
Christof
We'll be moving the PID values in Mega to a separate file from the other settings. This should make sharing much easier and I think 2.6 should get the same treatment.
I might make a configurator in Flash that helps people pick their PID values and spit out a header file. Maybe it's something we can host on the site.
Jason
that sounds good. To bad I left this morning without the last patch but I will try it tomorrow.
Please dont forget to check on the position for LOITER when you look for the endless circling. The center should be where LOITER is switched and not the next WP.
The N times Loiter is a nice feature. I will do the testing. btw where is the plane going after N times?
I would like to set up a PID settings database. I am sure that we could categorize the airframes and calculate an average setup to start with when all Ardu Pilots help and post their (good) values.
airframe name, plane type, wingspan, roll type (rudder or aileron), max mecanical aileron/rudder in ° up&dn/left/right with 100% servo, PID servo, PID nav.
What do you think, would that help to get good start values?
I thank you for all the effort you do to get the AP working.
I've replicated your flight pattern in the simulator and the last fix solves the Loiter shape problem. There was a slight mistake on the distance calc which causes the kidney bean shaped loiter.
I will look into code to avoid the endless circling. I have an idea of a new mode based on Doug's Loiter N times mode. We can count the revolutions around the WP and limit them to 2 or 3.
Thanks for all your help. This was a tricky bug. And now the LOITER code is better than ever.
-
3
-
4
-
5
-
6
-
7
of 12 Next