Re: [CAD_CAM_EDM_DRO] EMC stability

machines@... wrote:

> Many thanks to the folks who have answered my questions so far on EMC
> and to Ray Henry and Paul Corner who are looking into the previous
> G44 question I asked.
> Contary to popular myth I read the FAQ and also a lot of what is in
> the Linux/emc files.
> 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, Matt? It seems a little worrying if
> this can happen. I often set my machine running and walk away into
> the other shop to carry on bench work.

I probably shouldn't answer for Matt, but I keep in very close touch
with him about EMC issues. I believe this was his first EMC retrofit,
and was delivered to the customer in 1998. This machine may have
been updated since then, but this comment was definitely from before
the update. So, it is related to a VERY old version of the software,
and probably a very old version of the Real Time patch set, too.

I can say, that on the hardware I have, the Linux 2.0.36 kernel with real
time patch 0.9J, and the 20-Dec-1999 version of EMC is absolutely
stable. I have been running this version for about 18 months, and have
only had one burp. On that occasion, the keyboard locked up, but the
machine finished the program normally. I then had to reboot.
I have never had a "runaway" axis.

> 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?

G30? I suspect you mean M30, which was used to rewind the program paper
tape on Allen Bradley CNC controls. Officially, the RS-274D spec
calls M30 "end of data" and M02 "end of program". The more detailed
description doesn't really indicate much difference between the two
commands.

Jon