Re: [CAD_CAM_EDM_DRO] Re: UCMI Interface

Art Fenerty wrote:

> >
> > We demonstrated this capability with EMC and my stepper rate generators
> > at the NAMES show near Detroit this April. We got VERY smooth contouring,
> > and ran a Sherline with MicroKinetics drivers twice as fast as anyone
> > had ever seen them go before.
>
> Jon:
> How do you handle the circumstance where a circle is made up of 4 arc's.
> Do you ramp up and down each line? Does this type of circle move smoothly
> around the entire circle without slowing down? How is the decision made that
> says when to ramp down and when to ramp up each axis?

No, EMC looks ahead and sees each arc sector can be continued into the
next without exceeding the machine's acceleration limits (in fact basically no
acceleration at the 'joints', only in the rest of the curves). So, there is no

need to decelerate for this chain of blocks. Now, lets say we have a bunch of
short vectors that are nearly all in a straight line, then all of a sudden
there's
one the reverses the direction of motion. A surface contouring program might
look like that. It looks ahead, and finds the segment with a discontinuity in
the
movement, then works backward in the program to figure out where it needs
to start decelerating, so that the machine's acceleration limits are not
exceeded.

You never ramp up and down each axis independently, although you have to
calculate each axis' acceleration to determine if you are exceeding the
capacity of the machine. But, if you have to ramp velocity down, you have to
ramp them ALL down proportionately.

Anyway, EMC runs totally smoothly around a circle composed of 4 quadrants.
I also do 'spirals' that are composed of expanding arc quadrants which meet
at a point, but have slight dicontinuities in the radius. I use this to bore
out
round pockets and such, and it works real well!

Jon