Re: [CAD_CAM_EDM_DRO] Re: deskcnc servo driver

Gregory Kamysz wrote:

>So these errors are best avoided by using motors that are strong enough
>to keep up with the feeds or feeds low enough that the motors can handle?
>
>

Generally, yes, but it gets a lot more complicated. Larger machines such as
a Bridgeport mill have substantial inertia, and it takes a lot of force
to make
it accelerate rapidly. Yes, of course, the control must never request a
velocity
or acceleration the driver/motor cannot produce, or the error will increase
until the limit is reached.

>Setting the gain and damping on the Gecko 320's will also come into
>play. What is the best way to do that without access to a scope?
>
>Am I correct in thinking that these issues mostly come into play when
>making heavy cuts at high speed?
>
>

A scope is certainly the best way. But, it can also be done to a reasonable
degree by ear. You increase gain until you get oscillation, then back off
a bit. Then, you increase damping and gain together in small increments
while making rapid short moves. What you want is for those moves to
start and stop without wiggling (oscillation). It may be possible to use
a ballpoint pen in the spindle to plot the motion of one axis against
another
to determine this stability margin. Gcode like this could be used :

G01 X0 Y0 Z0.1 (PEN OFF PAPER)
G01 Z0 (LOWER PEN TO PAPER)
G01 X1 Y1 (START MOVE AT 45 DEGREES)
G01 X2 (CONTINUE X MOVE, BUT Y STOPS AT Y=1)

This should give a line at 45 degrees from (0,0) to (1,1),
followed by a straight line to (2,1). At the inflection where
the Y axis was supposed to stop, you'd see any instability as
a wiggle in the line. This is almost as good as a scope, although
the wiggles might be pretty small.

Either high cutting forces or just rapid motion may run the servo drive
out of headroom, and the following error will increase. The worst
case is usually at the maximum rapid traverse speed, which is not used
for cutting, so modest following error is not a problem.

Jon