Re: Closed Loop PC Based CNC

This probably doesn't satisfy your immediate needs but I am working
on a true closed-loop step motor/encoder system that will be able to
use any step motor drive.

The "magic" will be in the step/direction controller.

Step motors in many ways make ideal CNC motors; they have high torque
at low speeds where work is being done and little torque at high
speeds (rapids) where little torque is demanded of the motor. The
motor stalls when it falls behind by two ordinal steps. Just short of
that it is delivering 100% of its available torque for that speed.

A step motor's "percentage of available torque demanded" is
sinusoidally related to the motor's "lag angle". This simply means
the motor's position falls behind from where it should be as torque
increases. This "lag angle" can be easily measured with a motor-
mounted encoder.

The step/direction controller would adaptively react to an impending
overload by "backing up" the motor's speed-torque curve by decreasing
the step pulse rate sent to the drive until a new stasis point is
established, accomodating the new load-point.

This would not be a problem in 2-D and 3-D vector motion because the
velocity ratios of the 2 or 3 axies would be maintained; only the
vector-sum velocity would change.

The effect would be an "unstallable" step motor system. It would
accelerate at a rate limited only by the mechanicals to speed limited
only by the load or controller parameters. As load would change, the
motors would slow down or speed up as needed to adapt. An increasing
load would ultimately slow the motor down to its holding torque (zero
speed). Past that, it would misposition but the encoder would allow
it to recover position once the overload was relieved.

This past week I cobbled together a strange combination of a G320
(encoder inputs, summing junction and a PID loop), a G210 (a VCO and
step motor drive) and a G2002 (a motion controller step/direction
source). It was enough to demonstrate the feasability of this idea,
feeble and limited as it was.

I am coding up the G2002 to move the encoder inputs, summing node and
PID loop to reside within it. This will allow it to control an
ordinary step drive (G201) external to it when the driven motor is
encoder mounted. The results should be interesting because of the
goal.

Mariss



--- In CAD_CAM_EDM_DRO@yahoogroups.com, "nitewatchman5"
<garyclick1@c...> wrote:

>
>
> I have a Bench Mill currently running with stepper motors in an open
> loop configuration. We are running production parts in fixtures and
> after running several parts to several dozen parts, sizes and
> locations start to drift due to missed steps. Slowing traverse

rates

> down and extending longer acc/dec ramps helps but the problem is

still

> there depending upon cutting forces and the shear number of axis
> reversal moves. Additionally, slowing the machine to fit the
> speed/torque curve of the stepper motors increases the part cycle

time

> and the part cost directly.
>
> As a solution I want to take this machine to a true closed-loop

system

> using AC Servo Motors. While I have retro-fitted several large
> machines (25'+ travel) with Fagor and AB 9-Series controls,
> maintaining a PC based control presents some problems.
>
> A likely scenario would be to use a matched servo drive and motor
> combination such as the Yaskawa DR-1/SGM set for each axis. The
> encoder feedback from the motor for commutation can be redirected

to a

> Controller board in the PC for position error control. The

controller

> would output a +/-10VDC analog signal to the Servo Amps to control
> speed and direction. The calculated following error can then be
> intergrated into a PID Loop to control speed, torque and location.
>
> There are several interesting controllers on the market such as the
> Galil, Delta Tau and Servo to Go. The Galil especially looks very
> powerful and is sometimes found on eBay for a reasonable price. The
> big question however is the Front-End software to drive the
> controller, does for example EMC support these controllers?
>
> Advice would be appreciated, practical experience perferred -
> theoretical suggestion catalogued. If we can find no other solution

we

> will refit the machine with a Fagor 8025M control.
>
> nitewatchman