Pixhawk Octocopter Takeoff Problem - PWM output to some motors lower than others

Hello All,

I am working on a project this summer which was my senior project from this past year. It is a medium octa-quad copter using a Radiolink . I am using Mission Planner to read telemetry data and flying it manually using Taranis X9D Plus and FrSky x8R receiver. There is an error which has now occurred twice to the system which prevents it from taking off the ground. I am very confused about this since I was able to fix it once (which it appears whatever I did was not what truly fixed it).

So, the octocopter will arm and the motors will spin up to the min throttle level. This can be seen on the Mission Planner Screen where ch#out is shown for channels #1-8. This shows the PWM values that the ESCs are being sent from the Pixhawk I believe. The min value is 1084. Then as the motors PWM values increase as the throttle is raised, which matches the ch3in PPM value increasing, the values do not change equally. But the problem is not that simple. I know it is not simple things like the Pixhawk is uneven or ESCs are not calibrated because 1) that has been checked and 2) the PWM values for the channels will sometimes have very large variations, which can be seen and heard when the motors are spinning and the drone is still on the ground (and of course when we throttle higher the drone will ALWAYS flip). Sometimes one channel will get stuck at a value of 1135 (which is not the min throttle value). I do not understand this, it is almost as if the Pixhawk thinks it is uneven but only in the one motor which I guess would mean it things it is rolled and pitched at 45 deg angle (estimating of course).

So I am open to hearing what you guys think, I am pretty certain it isn’t a simple fix as I have mentioned above, but I am willing to go through every process again since I am sort of re-starting the project after a month vacation. I will try to get the data and video footage of the error occurring in the next couple days to add information on the error at hand.

Thanks ahead of time!

Caleb, have you checked motor order and propellers (clockwise, counterclockwise, pusher, etc …) configuration? The fact that it “always flips” if you throttle further, and the earlier signs you describe, could be explained by an incorrect set-up. One way to test this is to slightly increase throttle and at the same time pitching, or rolling, while still on the ground. When you do this you should see the copter (slightly) tipping in the right direction.

Next thing to check if you are absolutely sure of your configuration is to redo esc calibration, also radio calibration, and/or test motors individually to make sure PWM values are in the right range from low to high throttle (1000-ish to 2000-ish). Also posting a log would help.

Oliver, I will work on getting logs on by the end of today or tomorrow which I think explain the problem better than I can describe.
I have flown the drone perfectly fine before, where I did the things as you list which include slowly increasing throttle and check the pitching and rolling on the ground (and it felt great!). But this problem occurred once before it flew, and now again after it flew. And nothing major changed as far as I know, we did rebuild it almost from the ground-up though when it didn’t work the second time. So I will go through all the processes again like you mention, but I am certain the problem is more of a software or parameter issue.
Thanks for the help, I will get more information on the topic ASAP!

If the copter flips while propellers turns the right way the problem usually is that you have connected motors to the wrong Pixhawk outputs.

The solution is simple, go in Mission Planner Motor test and check running each motors once that motors are in the right order.

That is not the problem, as I have checked that many times. Also, like I mentioned, we have had it fly perfectly with the same exact setup.

You might have corrupt your configuration file, it happens once to me when the hard disk on my ground staton had bad sectors.
Save your configuration and reset all parameters starting a complete new config.

I have 3 craft that exibit the same behavior and never got an answer.
On my 450 quad, 550 hex and Tarot 680Pro, two motors spin at slower RPMs than the rest.
Motor logs confirm this and performing an Auto Analysis shows Motor Balance failure every time, every flight on all craft.

The fact that the craft in question “always flips” is not a true indicator of incorrect propeller placement.
On my 450 with my starboard side motors “out of balance” per the test, I have to gun the throttle and compensate roll left in order to take off. It is not as much of a chore to take off…it has become second nature for me since it has occurred on all home made craft since I started using PixHawk 2.6.8

And get this, if I reload the FC firmware by loading a different vehicle, reboot, load the proper vehicle and reboot…two completely different motors will be affected. As I said on the quad now, motors 1 and 4 are affected with firmware 3.4.6.

I test the rc firmware so I will flash to the rc, try it and flash back using the complete wipe method. One time with 3.4.6 I had low RPM motors on 1 and 2…and yet another time, I had low RPM motors on 3 and 4.

I am attaching the log and files. Also, I am including a picture so you can understand the difference in the RCOU values before even searching through my log file.

2017-04-25 17-06-06.log.param (10.1 KB)

2017-04-25 17-06-06.bin (518.3 KB)

I posted my log files and a picture showing my error. Check it out above. Is this the same kind of error you are having?

Yes, that is very close to what I experience.
As I indicated with listing my craft, I do not have a octo but that is what my power output would look like on the max side:
2 motors (out of 4 or 6) at least 100 less on the output.

I have a volunteer presentation tomorrow.
During the practice, I will turn on motor logging and post my current findings.

**Just captured a great sunset and enabled the motor logs:

2017-06-01 20-41-18.zip (1.4 MB)

Are you taking off in GPS or Loiter? Loiter is the best, otherwise it’s trying to chase a GPS signal with the gear still fixed to the ground.