Re: [CAD_CAM_EDM_DRO] re:Re: suggested improvement for EMC

Hi Jon,

This was really what I had in mind when I said that perhaps the machine
could back up a bit and 'take a run at the work' after a pause. I hadn't
really appreciated the extent of the look-ahead but, now looking back over
earlier posts it seems it can be quite a way so perhaps the idea isn't
feasible. Is the look-ahead a defined quantity like so many lines or, in
other words, could it be defined in such a way that, if the machine returned
to a point defined by the beginning of the look-ahead buffer and started
from there repeating previous 'cut moves' in thin air, it would eventually
catch up with where it should be in a controlled manner (or would the buffer
be so large that the time involved in such a move would be impractical)?

Ian
--
Ian W. Wright
Sheffield UK
www.iw63.freeserve.co.uk

----- Original Message -----
From: "Jon Elson" <jmelson@...>
> Yes, it sounds simple, but the interpreter was written with forward
> lookahead. So, on each move, it is carefully considering this move's
> effect on future moves until it has looked far enough ahead that
> this move could not affect any moves further ahead in time.
> It considers the ability of the machine to accelerate and decelerate,
> with respect to blending curves on all axes.
>
> That is why it is not easy to back up with EMC. Now, the interpreter
> COULD be reconfigured so that it would rebuild the lookahead
> any time there was a program interruption. This would also make
> it possible to put loops and subroutines in the programs, which it
> does not support now, for that reason. But, nobody who is qualified
> has stepped forward and volunteered to do this rewrite.
>