i have a S1000 with around 10kg and the cube pixhawk onboard.
I set filters to 10hz and autotune aggr to 0.07, did a autotune at no wind.
flew ok, a bit too twitchy. Pids were not usable in wind, a lot too high. So i did another autotune in 6m/s wind. Still a bit too much. I lowerd the values manualy a bit. than it flew fine.
now i have the problem that when i do automissions after 10 minutes the copter comes back and is shaking… pids too high right before landing. what could that be?
You do not have the battery fail safe configured and enabled. FS_BATT_VOLTAGE is 10.6. For 6s this should be 21(3.5 volts per cell). FS_BATT_MAH is 0. It should be set to the battery capacity.
And this is HUGE… You are sucking too much current out of the battery. If the power monitor is properly calibrated, the logged end voltage is 20, or 3.3 volts per cell. If the PM is properly calibrated for current you’re pulling a total of 9869.522 mAh.
The wobble you are seeing is because by the end of the flight battery voltage has dropped so far the PID controllers are loosing authority. I see this all the time at the end of a flight when I’ve hammered my racer, and I’ve seen it more than once when I’ve had an APM or a Pixhawk battery failsafe misconfigured and I’ve flown too long.
I ran some guestimates on your aircraft. With a 10000mAh battery Your projected flight time to battery exhaustion is 10.5 minutes. Your flight time with a 20% reserve in battery capacity is 8 minutes. With a 12000mAh battery your projected flight time to battery exhaustion is 12.62 minutes. Your flight time with a20% reserve in battery capacity is 10.018 minutes. Your g/w is 7.99.
One other thing I noticed is current draw “clips” at 60A. I’ve never seen a draw this high, so I don’t know if this is a limitation of the PM you are using, a logging limitation, or a limitation in Mission Planner…
In the Auto mission log is shows your Rate Pitch/Roll filter set at the 20Hz default. Didn’t you say you set these to 10Hz? Advisable with such a large multirotor.
Ah, I see you were referring to the accel and gyro filter settings. Try the Rate filter setting also.
Why on Earth do you choose not to use the battery fail safe? To my way of thinking (Politically Correct response follows) that is just begging for trouble, so to keep the censors and PC Police off of my back I’m bowing out.
I have to comment here, and hopefully my pointing out your irresponsibility, as well as @oldgazer1, will cause you to take heed.
I have just completed a repair of an Octo where the user saw a low voltage warning and thought “I usually get 5 more minutes from this battery so I will complete this leg of the mission”.
Bad idea one.
At the end of that leg the battery voltage warning was going critical so he thought “I will just take manual control and fly it quickly home” (it was over 200m away)
Bad idea two.
So by the time he gained correct orientation the battery voltage was hitting 3.2V per cell (16000 7S) and the craft was getting hard to control.
It lasted for another few seconds before losing stability (no current left) and going in.
If all you are looking at is current consumption and flight time you are making a very grave error.
By the time your current starts to drop it will be all over.
And the reason for the above described crash was an ageing battery that lost voltage sooner than expected.
Now the user knows to keep a more careful eye on the voltage, which is why I take such pains to carefully calibrate the battery sensor.
first I changed the gyro filters back to 20hz and the acc filters to 10hz, this was done by mistake.
did another autotune, on roll/pitch and set the PID voltage scaling from 25v to 20v. now it flyes perfectly.
for the voltage failsave. i am very ware of that. and for the whole mission i watch the voltage closely.
below 20.5v i do a manual emergercy landing on the spot. but that never happened, because i know my batterys and i plan my missions with power consumption in mind.
voltage failsave is dangerous for me because i want to have manual control all the time.
Thanks for your reply.