I have a HK pixhawk, ZTW Spider Opto 30A ESCs, and SunnkSky X3108S-720 motors.
I did the all at once ESC calibration following the Arducopter video but am not 100% certain it has worked. I have been flying the quad in multiple modes but feel it is not running at its best.
I wanted to check the RC out on each motor, so I removed the props and connected to Mission Planner so I could monitor the Channel outputs 1-4.
On power up all 4 channels read 982. After arming all 4 channels read 1051.
What you describe is normal behavior. If you’re in stabilize mode, then the motors will throttle up with different values depending on the tilt of the copter.
You can check the calibration by starting the copter in “all-at-once” calibration mode and throttling through the range. If the speeds match throughout the range they’re calibrated. You can usually tell by the sound and can hear it if one is out of sync. Don’t spend much time at full throttle, as you could damage the motors from heat without the props attached.
That being said, I doubt the motor cal is an issue, because you’re using the same ESCs all around they’re probably matched regardless of whether cal worked. What issue in flight are you noticing that you think needs to be improved? Flight logs can help diagnose that.
I saw a motor warning message in the Auto Analyse report in MP which alerted me to the discrepancy. I then checked some logs and could see that motors 1 and 2 are closely matched, but 3 and 4 are a large range apart.
I would expect variations in the outputs when pitching or rolling, but the values I observed were with the quad level on the ground with no props.
Can I rule out differences/problems with the motors? As I understand the channel outputs, this is the PWM value sent to the ESC. Is there any feedback from the motor to the ESC and/or Pixhawk which could influence the Channel output?
Varying the throttle (0-100), it sounds like the motors are matched, but the logs and realtime status show this not to be the case.
Make sure you get the copter in a hover the air when making motor balance evaluations, as they’ll be out of balance on the ground because the copter will be fighting against the ground. Using the dataflash logs or a telemetry log in Mission Planner will be helpful (I don’t recommend flying with a USB cable attached!)
If they’re not matched in hover, it’s usually one of two things. Either the copter isn’t balanced at the geometric center of the motors, or the motors aren’t mounted level (common with round boom frames). Can you check CG and motor leveling and see if you notice anything out of the ordinary? My guess on a 3/4 imbalance is that one of the motors is tilted.
Occasionally, it can be a prop, ESC, or motor that doesn’t behave like the rest. But I find that’s less common than the above.
I checked the CofG, and it was a bit off so I moved the battery back about 0.5". The motors all look square to the arms and at the same height above ground. The arms on the S500 frame are angled up slightly.
I will run hover tests in Stabilize, AltHold, PosHold once the wind co-operates!
Interesting. A CG shift wouldn’t cause what you describe, as it would make two motors high and two motors low. And the S500 arms aren’t likely to be twisted. I’m interested to hear the results of the flight test.
I believe I have found the cause. I took a close look at the #3 & #4 props – they were slightly different lengths!
Some time ago after a minor ground plough, I sanded the chipped rough edges to clean up the prop and continued to fly. What I failed to notice was that after this repair, the props were not the original length but approx 1/4" shorter which I noticed when placing 2 props side-by-side.
I put on brand new props and went for a test flight even though there was a lite wind.
I just ran the Auto Analyze in Mission Planner and the motor test result was “Good” . The RCout graph for channels 1-4 are now much closer together too.
This hobby can be frustrating, but it sure feels good when you get a solution to a problem – especially if it is of your own making!
. The RCout graph for channels 1-4 are now much closer together too.