First enemy: WINDY DAYS.
It was windy but ive managed to find some less windy areas to fly today,
i decided to give it a chance, after checking all's OK in the living room, hovering about 10cm...
GPS hold seemed to work, as at one point it was drifting a lot with the wind and suddenly changing direction back to the launch point (not what you see on the video)
ON THE VIDEO
I have no idea what happened, i engaged GPS hold a few seconds after it flipped over,
and anyways, it flipped in the direction of the wind, which means it didn't flipped by trying to compensate and fly against the wind back to the launch point...
my guess is that was slightly unstable because of the wind and finally the wind caught it, gave it a push and it flipped....
have any other idea of what happened ?
I wanted to test some equipment,
here you can see the transmitter of the wireless tracking system mounted on the copter:
Red LEDs strip
The lipo alarm with the camera power supply behind:
the 5v BEC for the APM and receiver:
The wireless pinhole camera:
Didn't get to testing the wireless camera,
but the custom wireless tracking system was working great from about 100m away !
Hi, did you solder the wires together? Did it help?
I myself experienced 5 crashes like this, firstly thinking it's me doing something wrong or the propellers getting loose. But now I know it's the hardware or software (ie. some-ware :) ) failing. My ArduCopter is always swaying in the same direction and eventually sometimes flipping and crashing. I've already changed the ESC and a motor (but the connectors are still the same). I've also flown both ArduCopterMega 2.0.026 and ArduPiratesNG v627, both with the same effect. So hopefully soldering will help...
yes it's the Pirates firmware...
dont know about logging, but im going to solder these now...
Actually I've seen those connectors fail and still be plugged in. They loose connection slightly and cut off the ESC. I've also seen the DCM struggle to recover from Gyro saturation in tough conditions. This sounds like what some of your struggles were. If you were logging you could check. Are you running Pirates? Do they have logging?
yes i have them, but they haven't disconnected !
after the crash all motors worked fine.
so it can't be that...
question, do you have the bullet connectors soldered to the wire?
if yes, then cut em off and solder the cables without using the bullet connectors.
i run into a similar issu - the copter just flipped and crashed.
removing the bullet connectors and the copter stays in the air.
if i get it right it is the overcompensation or some kind of out-of sync- of the pid calculations that make it finally positive feedback and thus going with the wind direction instead of against it... ?
but this case can occur in strong winds when it has to compensate more and more quickly with much more changes in the wind forces, right ?
This shouldn't happen when 'estimation' is easier thanks to having no winds and no other external wind forces...
Not to jump to conclusions, but maybe this is an example of what Gary has pointed out to William in the DCM algorithm? Check these comments for more information:
after replacing the broken propS, i could get it hovering again in the living room.
this time it had some drift to the side and front, i think the arms might have been turned a bit so the motors are not totally vertical. possible ?
but it's hard to know because it's a large quad (28cm arms, 12" props) in a small place with little throttle...