MSuper V3s+ Violent Shaking

[zlymex]

For the first generation of balancing vehicles such as Segway/Solowheel, the pedals are very 'soft', meaning the pedals will tilt forward and tilt back with very large angle. While this create problems in some ridding situations, it make the firmware/algorithm easy to control.

When the new generation EUC emerged namely Rockwheel 12/16 geared in 2013, the pedals are very 'hard', the respond is very quick, but it suffered from vibration problem right from the very beginning. Just like @MichaelHensen has guessed, the problem lies in the PID algorithm. They may re-write the code or simply adjusted the PID parameters, either way they made the loop unstable, in more professional term, the loop gain became larger, and the phase margin became smaller. Motor control is a very complex matter which involves at least two control loops(one for current the other for speed), and it take a good loop stability analysis in order to create a good algorithm. I've written a very long thread explaining how loop stability is done. Like many have pointed out that Rockwheel/Gotway/KS share the same origin, I'm in doubt GW has the expertise for that except trial and error.

My Msuper2 suffered from high speed vibration and GW says it is normal. My Msuper3s+ suffered from low speed vibration but I tolerate this.

[Christoph Zens]

49 minutes ago, zlymex said:

For the first generation of balancing vehicles such as Segway/Solowheel, the pedals are very 'soft', meaning the pedals will tilt forward and tilt back with very large angle. While this create problems in some ridding situations, it make the firmware/algorithm easy to control.

When the new generation EUC emerged namely Rockwheel 12/16 geared in 2013, the pedals are very 'hard', the respond is very quick, but it suffered from vibration problem right from the very beginning. Just like @MichaelHensen has guessed, the problem lies in the PID algorithm. They may re-write the code or simply adjusted the PID parameters, either way they made the loop unstable, in more professional term, the loop gain became larger, and the phase margin became smaller. Motor control is a very complex matter which involves at least two control loops(one for current the other for speed), and it take a good loop stability analysis in order to create a good algorithm. I've written a very long thread explaining how loop stability is done. Like many have pointed out that Rockwheel/Gotway/KS share the same origin, I'm in doubt GW has the expertise for that except trial and error.

My Msuper2 suffered from high speed vibration and GW says it is normal. My Msuper3s+ suffered from low speed vibration but I tolerate this.

The Ninebot One E+ has one of the hardest response even today (much more so two years ago), but I never had any vibration. On my new KingSong KS16-S, pedals are also very stiff in the hardest setting, and I also don't get any vibration problems, not at low or at high speed. The only thing I noticed is that the KS16 has a larger dead band where you can slightly tilt the pedals back and forth but the motor doesn't do anything. This is not present on the Ninebot. Any move will immediately trigger a response from the wheel.

[Carlos E Rodriguez]

1 hour ago, zlymex said:

For the first generation of balancing vehicles such as Segway/Solowheel, the pedals are very 'soft', meaning the pedals will tilt forward and tilt back with very large angle. While this create problems in some ridding situations, it make the firmware/algorithm easy to control.

When the new generation EUC emerged namely Rockwheel 12/16 geared in 2013, the pedals are very 'hard', the respond is very quick, but it suffered from vibration problem right from the very beginning. Just like @MichaelHensen has guessed, the problem lies in the PID algorithm. They may re-write the code or simply adjusted the PID parameters, either way they made the loop unstable, in more professional term, the loop gain became larger, and the phase margin became smaller. Motor control is a very complex matter which involves at least two control loops(one for current the other for speed), and it take a good loop stability analysis in order to create a good algorithm. I've written a very long thread explaining how loop stability is done. Like many have pointed out that Rockwheel/Gotway/KS share the same origin, I'm in doubt GW has the expertise for that except trial and error.

My Msuper2 suffered from high speed vibration and GW says it is normal. My Msuper3s+ suffered from low speed vibration but I tolerate this.

Do you have a link to your PID. PROPORTINAL INTEGRAL DERIVATIVE Control? It is a well understood algorithm. You can get initial gains from the electronic R-L-C motor characteristics and inertia of stator.  But the final gains need to be tuned manually to achieve no oscillation at steady speed. I did control systems and had to tune PID loops by hand using resistors and capacitors and paper chart recorders. Thank the Engineers for the software PID much easier. From my experience the EUC application should not need a very tight PID. Industrial motion controls use very tight PID to accelare to speed with no overshoots, maintain speed without oscillation and decelerate with no undershoot. The EUC should have a good tuning for the at-speed portion. 

[Carlos E Rodriguez]

http://www.ni.com/white-paper/3782/en/

[Carlos E Rodriguez]

http://courses.csail.mit.edu/6.141/spring2011/pub/lectures/Lec03-MotorControl.pdf