I just finished building my first quad and encountered strange issue during second test flight (ardu-copter V3.1.5). Few seconds after lift-off APM stopped responding to throttle commands and started gaining altitude. The quad was still not stable as none of the gains were tuned. I have helicopter flying experience, thus I was testing quad in "Stabilize mode". When it went to RTL it started gaining altitude and I kinda freaked out not understanding what is happening, so moved throttle to 0 and switched between modes (I was not flying in an open area and quad was flying towards a tree). That did shut off motors and quad crashed through tree into ground. Not much damage, only one arm bent.
At home I reviewed flight logs and only then realized that it entered RTL mode caused by battery fail-safe. However, fail-safe is set at 16V and according to the logs voltage never fell bellow ~18.6V. Capacity was also set correctly. Any ideas what could have caused it to enter RTL mode? Maybe some kind of a bug or wrong data from voltage sensor that was not captured by logging?
Any help is appreciated!
I am attaching whole data log to the post and failure data here (I deleted IMU data as it is not as relevant, but available in full log):
CURR, 317, 93192, 1897, 1863, 4876, 66.89587
GPS, 3, 482316000, 1807, 8, 2.23, 54.9890627, 23.9474907, 0.37, 81.09, 0.09, 64.79, -1.010000, 426767
GPS, 3, 482316200, 1807, 8, 2.23, 54.9890632, 23.9474913, 0.59, 81.21, 0.28, 64.79, -0.980000, 426968
GPS, 3, 482316400, 1807, 8, 2.23, 54.9890617, 23.9474924, 0.75, 81.73, 1.01, 64.79, -1.300000, 427168
GPS, 3, 482316600, 1807, 8, 2.23, 54.9890610, 23.9474936, 0.83, 81.85, 0.64, 64.79, -1.230000, 427369
MODE, RTL, 366
ERR, 6, 1
GPS, 3, 482316800, 1807, 8, 2.23, 54.9890597, 23.9474934, 0.91, 81.92, 0.63, 64.79, -1.140000, 427570
DU32, 7, 272761
CURR, 258, 95655, 1933, 1389, 4940, 70.64372
Your BATT_CAPACITY is 3670mAh.
your FS_BATT_MAH is 3600
So you are asking for FS when there is less than 3600mAh left.
your CURR.CurrTot indicates that you've used 70mAh right before RTL.
So.. everything is fine. (which is pretty usual when people blame it on "bugs")
you should read documentation on FS_BATT_MAH http://copter.ardupilot.com/wiki/arducopter-parameters/#Failsafe_ba...
Indeed mistake on my side. I thought that FS_BATT_MAH is how much battery is discharged, not how much is left.
BTW: be sure to remember that you cannot discharge a Li-xx battery to 0% while flying. Whatever your battery capacity is , you should not enter more than ~80% of that.
Yes, I perfectly know that. When I entered 3600mAh I thought that would disable capacity fail-safe (as it would never reach 3600mAh discharge). Boy I was wrong. I trust voltage readings more in determining battery state. I will disable capacity cut-off as in my opinion voltage cut off will work fine. But at first I need to fix the flying machine.
yes, voltage is a better indicator, less precise prediction when you are 50% into the flight time, but more accurate prediction of the end, like "5minutes left."
it also handles half-charged batteries, and different capacities better when some of your packs get old.
How much voltage (per cell, LiPo) would be good for a Battery Fail Safe?
I did set mine (3S) to 10.5V (so, 3.5v per cell) but the drone come back too soon.... (because the voltage drops while on load....)
What voltage per cell would you recommend?
That depends on how far you fly from landing spot. But I would suggest ~3.7-3.8V per cell as it will be much better for pack (although you will lose a minute of flight)
well, I do not use battery failsafe, but I get a warning over HoTT and telemety, and OSD at 14.2v (<=3.55vcell)
- the setting depends on the percentage of battery used at this point (if larger capcity, this means more flighttime left, and voltage drop is lower.)
-you are using 3S - that probaply means higher current - and the cell may drop more while climbing etc.
in short : you need to base it on your own experience, and the distance you may fly.