Re: [CAD_CAM_EDM_DRO] Re: Servo Control--Torque vs Velocity
Posted by
Jon Elson
on 2000-11-01 16:05:29 UTC
Mariss Freimanis wrote:
makes that match the input command. A velocity servo has another servo
loop outside the torque loop, which measures velocity from some sensor
and adjusts the torque command to obtain the commanded velocity.
You can't connect a torque servo to a CNC control that is sending a
velocity
command out and expect good results. It may not even be stable.
Velocity servos can use any method to obtain velocity info, including
deriving it from the encoder. The only problem with this is there are
no velocity updates between encoder counts. The system is literally
open-loop between encoder counts, as any change in velocity is
unknown until the next encoder count comes in. This lowers bandwidth,
possibly by a LOT. Consider the case of VERY slow motion on one
axis, perhaps while following a curve almost parallel to another axis.
The encoder counts may only be coming in at a rate of 10 counts/sec.
It would be impossible to have a real world bandwidth over 5 Hz
under this condition, even if the servo loop on the computer is being
processed at 1000 Hz.
A DC tach is not troubled by this, and provides wideband velocity
information no matter how slow the axis is moving.
Jon
> --- In CAD_CAM_EDM_DRO@egroups.com, jmw@c... wrote:No. A torque servo measures the current it is sending to the motor, and
> > Can someone comment briefly on the difference between the torque
> and
> > velocity servo control schemes? I understand that both require
> > encoders but only one must also have a tach. Other things being
> > equal I'd just as soon not bother with tachs--but are other things
> > equal?
makes that match the input command. A velocity servo has another servo
loop outside the torque loop, which measures velocity from some sensor
and adjusts the torque command to obtain the commanded velocity.
You can't connect a torque servo to a CNC control that is sending a
velocity
command out and expect good results. It may not even be stable.
Velocity servos can use any method to obtain velocity info, including
deriving it from the encoder. The only problem with this is there are
no velocity updates between encoder counts. The system is literally
open-loop between encoder counts, as any change in velocity is
unknown until the next encoder count comes in. This lowers bandwidth,
possibly by a LOT. Consider the case of VERY slow motion on one
axis, perhaps while following a curve almost parallel to another axis.
The encoder counts may only be coming in at a rate of 10 counts/sec.
It would be impossible to have a real world bandwidth over 5 Hz
under this condition, even if the servo loop on the computer is being
processed at 1000 Hz.
A DC tach is not troubled by this, and provides wideband velocity
information no matter how slow the axis is moving.
Jon
Discussion Thread
jmw@c...
2000-10-31 16:10:00 UTC
Servo Control--Torque vs Velocity
Mariss Freimanis
2000-10-31 16:59:28 UTC
Re: Servo Control--Torque vs Velocity
jmw@c...
2000-10-31 17:47:29 UTC
Re: Servo Control--Torque vs Velocity
Jon Elson
2000-11-01 16:05:29 UTC
Re: [CAD_CAM_EDM_DRO] Re: Servo Control--Torque vs Velocity
ballendo@y...
2000-11-01 16:39:55 UTC
Re: Re: Servo Control--Torque vs Velocity