I finally got up the courage to try a mission with 2.05. I set up a very simple mission - takoff to 10 meters, loiter time 60, land. All the coords were very close to each other so it should land about where it took off. The first two times I started the mission, it went up a few feet and then bounced around the ground like it was at altitude and loitering. The third time was very different - Up it went very high - seemed like more like 100' rather than 10 meters - and I was sweating bullets. It then loitered for what seemed like much more than 60 seconds, but I bit my lip and waited. It then began a landing pretty much on target, but at about 20 or 30 feet it decided to head towards a corner of the park and I aborted, though it might have landed OK if I had more room.My theory is that there may be probs in the altitude sonar/ baro switch. When I lock height, it bounces a lot if within sonar range, and sometimes just shoots up without stopping. I verified the sensors by loading Ardupirate code - unit is very accurate on sonar height with this code, so I know the hardware is OK.The landing way off target was puzzling. I let the GPS run for a while, but maybe coords changed. I wish I would have had more room to see where it ended at.I also assumed that mission end at 'land' stops the mission, but soon realized I had to kill mode switch as soon as it landed or it will repeat the mission. Might be nice to wait for a throttle toggle to start again.Anyone else seeing this?
I tell more than 10 times that arducopter have a serious problem with altitude hold.
The first problem are in the sensor.pde that used a code diferente that ardupirate / bosh paper.
I changed the code on ACM2 sensor.pde by the code from ardupirate or bmp085 demo and the altitude hold work much better but continue with serious problem in the algorithm.
I am working on ACM2 ALT_HOLD code when i can (personal interest) because i have my own stabilization code and now are the time to put a alt_hold in my own code.
Comments
I tell more than 10 times that arducopter have a serious problem with altitude hold.
The first problem are in the sensor.pde that used a code diferente that ardupirate / bosh paper.
I changed the code on ACM2 sensor.pde by the code from ardupirate or bmp085 demo and the altitude hold work much better but continue with serious problem in the algorithm.
I am working on ACM2 ALT_HOLD code when i can (personal interest) because i have my own stabilization code and now are the time to put a alt_hold in my own code.
Jason, lock at sensor.pde and tell anything.
regards.
You may have had you throttle Cruise value set incorrectly. You need to hover in Alt hold first to set it.
60 seconds is a long time. GPS is only so accurate. I will be doing more testing this weeked.
Jason