Worth reading for the gripping detective work on that corner-case DCM bug alone!
Bravo to Andrew and the team on the DCM fix.
I'm positive that you've found the issue that eventually caused my 1st high alttiude return attempt (NTNS-2) to lose attitude reference (Link). This flight experienced 2-3 Hz oscillations due in part to autopilot gains robustness issues. I had determined that the accels and rates were not saturated but I never really understood why the DCM fell apart. The next two flights had much more robust gain sets and had no issues with the DCM. It's still good practice to make sure autopilot gains are robust even though we now have better DCM robustness with 2.76. I'm looking forward to seeing results with the new build.
I don't know if what I was experiencing with 2.75 is what was fixed in 2.76.
With 2.75, when switching to RTL, plane was flying nicely back home, loiter once and then went into spiral dive with full throttle as if it decided to commit suicide. switching quickly to MANUAL saved the day. can't remember this ever happening with 2.74 or previous even though Andrew claims this bug was there in 2.74 as well. Anyway, I have flashed mine with 2.76 and will check again tomorrow
Just to clear things up - this was not a stall.
the plane started pointing its nose down and in a steep dive, wings leveled and motor at almost full throttle was flying towards the ground.
Do I have to do all calibrations again when updating from 2.74b to 2.76?
No, you don't need to recalibrate.
Thanks for your fast responding.
greetings from the Netherlands,
Do you have telemetry or dataflash logs of this flight?
I believe my MP was OFF during these test flights so I will need to download logs from APM.
I will search for the relevant ones and will PM you.
Thanks for looking after us!
After first flight tests with 2.76 I have at least two new problems:
1) Bad log recording. I have problems to display, graph and analyse logs since 2.76 upgrade. It seems like now there is a new column of data after the first one that makes titles to be displaced backward one position and mismatch with data shown in the table. To explain it better, please take a look to the following screen shots:
2.74b log: I can graph all data (IMU, for example) whiout problems:
2) worse take off with 2.76 than 2.74b. I don't know why but throttle starts later in hand launch after upgrade firmware to 2.76 until the point that now it don't take off. I will attach tlogs in a next reply.
Please, find here attached the related two tlogs to complete comparison
I am still doing more research about my problems with 2.76 and I have found a most tricky problem that could explain what I posted before. I lost a lot of parameters when I upgrade from 2.74 to 2.76! Concretely the following ones dissapear after upgrade:
By the way, I fly with an APM 2.6.
If I downgrade the firmware to 2.74 I recover them and I lose them again if I repeat the firmware upgrade to 2.76.
What I wonder is how I flew this morning with this setup without crash my plane. It hardly took-off in my first flight but it flew rather well without rare behaviours (although I didn't saw too much and I had to land shortly because of the mist). My plane did not take-off in my second attempt and then I decided to check all well before to try it again.
Some one else have had problems like me?. In the mean time I will get back to 2.74 until these issues were resolved.