Re: A couple newbie questions about EMC

[Will Shackleford <shackle-at-cme.nist.gov>]

I will update the getrcs.html file shortly.  However I would generally 
not expect
EMC developers to use it but rather to get both EMC and rcslib together 
from
www.sourceforge.net either a combined tar or from CVS or to use the 
source code on the BDI.

For  EMC,  yes you need to compile much of the code twice , once to run 
in realtime and
once in nonrealtime.

My guess is the system is hanging because the realtime program is using 
so much of the CPU that nonrealtime services like Xwindows etc never get 
 a chance to run.

Try increasing the value of PERIOD

; Base task period, in seconds
PERIOD =        0.000500

and/or the TRAJ or individual axis  CYCLE_TIME parameters.
PERIOD should always be less than the CYCLE_TIME's
Larger PERIOD's mean you can send out fewer pulses per second but you 
 can run on
a slower computer.

I have an alternative build system based on autoconf/configure/automake 
etc that has
successfully built emc and rcslib on a couple of systems here. Hopefully 
I will be able to release it in a couple of weeks. I am still not sure 
whether it is better than the static Makefiles/scripts we currently use. 
Most likely both sets of build files will be included for some time.




-- Will




 



Simon M. Arthur wrote:

>
> I'm building a hexapod mill and want to use EMC to control it. I 
> downloaded and installed the latest release of the BDI. It seems to 
> run fine with the default configuration.
>
> When I switch to hexapod control, however, the entire machine freezes up.
>
> I'm not even sure if I'm doing this right, however, as the best 
> documentation I found is at 
> <URL:http://www.itl.nist.gov/div896/emaildir/emc/msg01095.html>. I 
> believe that I also have to have the line
> EMCMOT= minitetra.o
> in the [EMCMOT] section of emc.ini. Is this correct?
>
> In an attempt to get around this, I decided that I needed to compile 
> my own code. It's taken me two days, but I finally got everything to 
> build correctly. EMC and the RCS library are some of the most 
> difficult to compile packages I have run across in recent years.
>
> Some of the problems I ran across:
> The links on http://www.isd.mel.nist.gov/projects/rcslib/getrcs.html 
> are out of date.
> The file ftp://isdftp.cme.nist.gov/pub/emc/rcslib/rcs_src.tar.gz 
> doesn't include the necessary scripts to build the RCS Library.
>
> Anyhow, I brought everything up to date, with the latest release of 
> EMC, RCS, and Linux 2.2 and rtlinux. After all this, my machine still 
> locks up and won't respond when I attempt to use minitetra.o. The last 
> line displayed when I execute emc.run is
> starting EMC MOTION PROGRAM -- minitetra.o...
> Then the machine accepts no input from the keyboard or mouse. I hear 
> no disk activity.
>
> I don't really know if I compiled everything right. As near as I can 
> figure, you have to make two passes to compile EMC, one for realtime, 
> and one for non realtime. Am I working along the right lines with this?
>
> I read a few snippets which suggest that the GNU Autotools may be used 
> in the future to build EMC. Has there been any progress made on this 
> front?
>
> EMC looks like a very useful piece of software and I hope to use it on 
> my hexapod someday. Any help you could provide would be greatly 
> appreciated.
>
> Simon Arthur
>


-- 


--------------------------------------------------
Will Shackleford
NIST
100 Bureau Dr. Stop #8230
Gaithersburg, MD 20899
(301) 975-4286
Please do not send me Microsoft formatted documents.
They frequently contain viruses and discourage fair
competition and open standards.
I will ignore any email that contains them unless 
 absolutely necessary.
-------------------------------------------------