Just loaded 2.0.39 from MP .54 (stock AC frame, MB1200,compass). I tested sonar alt hold and loiter - seem very good now. Sonar hold stayed within a few inches (under .38 I got +-1 foot slow oscillation, no osc now). Loiter stayed good within a few feet. It flew very well, no yaw probs, etc.
I then screwed up my courage to try a mission again. I wanted something very simple that would be repeatable, so I just did takeoff to 2m, loiter at 2m for 20 s, then land.
I the flew a little alt hold and got ready for mission. The first time I ran it, it went a little hi but not bad, then did a very wide loiter (60-100 feet) for a couple circles, then went into land and came down fine, close to where it should. The next run of same mission did not go very high and did big aggessive loiter a couple of feet high- very scary, but I waited for land and it went down in about the right place. Third time not the charm, tho. It went up fairly high and then started a huge very aggressive circle, when it became apparent it was headed off way too fast I aborted, and will have fun fixing arm tonight.
Is my program bad? What changes each run? Log attached
I also want to thank Chris for the 'Meet the Developers' - I have long wondered who and where these brilliant people are, and was surprised to see Doug lives about 30 min north of me. Thanks again to all!
Replies
Could the attribute of having different result be an issue on how the GPS fix or issue with GPS that it fluctuates that causes different result?
@ Marcelo I also have Gaui 330X and FY tech both are good but no where near as good as Ardu, which as you mentioned why we are here :) Now all my frames have arducopter on them and now they fly as good even better than stock 344 and Fy-91 Now, I have at least 15-cycles with 3 different version of ACM and no crash. (so far) Prior to this I have 2 major crashes and some minor take off issue but now I found a way to make them always work.
@ Vincent I think you posted that question a couple of times in different threads. I am not sure if you posted the picture on your main forum but if not can you post the connection between your sonar and ardu on your thread and we will try to help you if we can. could just be some connection issue. Hope its not the board or the sonar.
Edit ok ignore it found your other post, its seems to be working on v37 as you noted it could be other issue then but mine is working with same sonar on v38
Do you have barometer covered with felt ?
This is just a theory of mine for the shooting up too high, but here goes. In the code, sonar is only considered fully when below 4m. Between 4 and 6 meter baro is mixed in and above 6 meter only baro values are used. Above 10m barometer reading, sonar reading is not considered at all (the sonar calculation is not executed)
So if your barometer values are not too good, and you shoot up above 4m, then baro might keep sonar from taking over again ?
The loiter aggresiveness could maybe be contributed to not so precise gps readings ? Maybe first time gps readings were pretty close to where you already were. Second and third time they might have been further off, causing the circling ?
why doesn't sonar work on my quad??
C'mon it is not fun to have to fix an arm where supposed to be a controlled flight... if 1st and 2nd worked, why not the 3rd!!! we all encourage and support programmers& this proyect to become almost 90% reliable and get out of beta soon... congratulations for the test we need more and more reports and updates on this kind of tests!! right?