Is it possible to overpower a design?

  • APM 3.1.2 (Pixhawk)

  • 2.070 kg total take-off weight

  • Tiger U3 motors with 13" props

  • It tunes in a fixture just fine. The P, D and I values all respond as you’d expect

  • MissionPlanner accelerometers and gyros seem to be working fine

  • Compass is calibrated

But when I pop it off the ground, it comes up evenly but starts veering uncontrollably in one direction (and not always the same direction). It’s like it just takes off a little bit vertically, but skids across the ground at an increasing rate of speed.

The catch is, I expected the total weight to be a full 1.0+ kg higher and I did make a switch from 3S to 4S at the last minute. Between those two changes, I have a theoretical lift to mass ratio of 3.5 with 100% throttle. Theoretically I should be hovering at no more than 40% (and probably more like 35%) throttle.

Could that be the reason it behaves that way? Is there a throttle-to-mass ratio that is just too much for a Pixhawk/APM 3.1.2?

Thank you in advance.

Did you tune with the 3s and then swap to 4s? the tune will be different between 3 and 4s.

Stu

In your tuning video in your fixture you are in ground effect all the time

If you have too much power available you will find the vehicle twitchy but 30-40% empty is reasonable.

I still recommend you go out side and do an auto tune. Also look at your log files and see what your average throttle is and set THR_MID to that value. That will help make the transitions into and out of auto smoother

  • I did the tuning with the 4S

  • Outstanding call on the ground effect. I hadn’t considered that. I need a taller fixture for sure

  • I’d love to put-in nominal values and autotune, but I’d probably get it about three feet off the ground before it starts “skidding” to one side (imagine it staying in the right orientation, but moving laterally) at around 20 mph. This isn’t a drift – picture a car at speed hitting ice.

I’ll pull some videos together if weather permits.

That’s what’s vexing me about this build. It isn’t flipping on take-off (i.e. motor in wrong direction). And in Mission Planner, the orientation sensors appear to be correct.

One dumb question… In a +4 configuration, the front and back motors are clockwise, and the lateral motors are counter-clockwise, correct? I would think that could cause a skid (imagine trying to correct and just getting more of the same orientation). Also, is there any chance my accelerometers were calibrated with a left-right swap or a front-back swap? Would a craft skid like that under those circumstances?

This isn’t like anything I’ve seen before. I’ve done a kit S800 evo and a full-blown custom octo-8/15" prop model both with A2 controllers. Those things popped off the ground first try. When I find out what I’m doing wrong, I’m afraid it will be embarrassingly dumb.

Are you sure you have GPS lock when you take off?
Can you put the vehicle into Loiter and have it hold position?
Are you sure you don’t have some trim on your transmitter?

I’m going to try that.

I need the weather to cooperate, but I have a fresh batch of 13" props that arrived today. I’ll try to post videos.

I really appreciate all the help you guys are offering. I’ve had great success with DJI A2s but I clearly see the benefit of the Pixhawk/APM/Open Source movement. I’m one of those customers that are dying to be converted.

Let’s get this thing off the ground and I’ll post videos!

It flies! It flies!

I picked a nominal P value of around 0.093 and it worked fairly smoothly. I didn’t have enough room to run autotune (especially because a failure meant a water landing).

  1. It wants to drift to the right – I have to fight the controls a little to keep it from wandering off

  2. When I flip it into loiter mode (with a good GPS lock), it goes from drifting to the right to a full-blown takeoff to the right. I have to almost immediately take it out of loiter and back to stabilize.

Here are a couple of videos:

dropbox.com/s/fv2nwqyhpd453 … _hover.mp4

dropbox.com/s/vdefixpx1rn4b … loiter.mp4

dropbox.com/s/e809807fjv787 … loiter.mp4

Here are a couple of log files:

dropbox.com/s/kyfv6ue72jj3c … ts.bin.log

dropbox.com/s/h941wrk2xstvi … 05.bin.log

In the middle and last videos, you can see when I flip the switch from stabilize to loiter.

  • My corrections up to that point are just little bumps to the left.
  • But once I go to loiter, you can see it bank hard to the right of the frame.
  • I have to switch it back to stabilize and get it on the ground (because I’m not the best pilot)

Update:

Copter freaked out and crashed in a pretty spectacular fashion.

Waiting for two replacement T-Motor U3’s to arrive since it cost me two motors.

I’m five T-motor carbon fiber props down and two T-motor U3 motors now.

My plan is to regroup, move to cheaper nylon 12" props and try to get restarted next week.

I’d check for vibrations.
Just a wag, but that’s what it sounds like to me… Worth a shot.