Re: EMC stability

From: machines@...

<s>I came across one referance to one of Matt Shavers customers who

>noted that "We are still having run aways"
>Can anyone shed light on this<s>

Releases from about two years ago now had a few places where it was
possible to cause a divide by zero. Most of these showed up as NAN
(not a number) in the position windows and would shut the machine down.
Some just caused an axis to head off toward infinity at rapid. Ian Wright
found several of these soon after the backplotter was released.

With the exception of some USA battleship running NT-4, I haven't seen or
heard of a NAN recently. I don't think that the EMC will be tarred by the
ship incident, 'cause without a lot of help, it won't run under NT-4.

>Second thing I picked up was a recent post about the BDI install
>resetting the referance points after a G30, end of program move. Has
>this been dealt with yet?

No. And it isn't likely to be dealt with. It is a planned in feature
rather than a bug. The purpose is to give the EMC interpreter a known world
view for the start of the next run. If you want to preserve the G92 values
use (%) on the first and last line and skip M30 or M2 at the end of the
program.

HTH

Ray