Re: [CAD_CAM_EDM_DRO] purpose of the group
Posted by
Jon Elson
on 2000-07-07 15:44:23 UTC
ebower wrote:
equipment in a commercial endeavor, but many of them have their
shop at home. (Some of them, conversely, have their home in the shop,
or
would like to.)
Rhino, Dolphin and similar things
are definitely NOT commercial. Gibbs, Unigraphics, Catia, etc. are the
commercial
packages, going for $30,000 for a complet suite, up to $25,000 for each
module,
and you need dozens of modules before you get the ability to do much!
National Institute
of Science and Technology (NIST). It takes in RS-274D (G-code) commands
and moves
a machine (can be a mill, lathe, robot or hexapod) as directed. It also
has a manual
mode for aligning the axes with a reference point and simple cutting
operations.
then ported
to Windows NT. The "Real time" extensions to Win NT (at least at that
time) were
very poor, in the sense that interrupt latency was VERY high, and
expensive, too.
So, when a real-time extension was created for Linux, they tested it,
and it was
observed that the interrupt latency was MUCH lower than with Win NT, and
so they
ported it to RT-Linux. I can say with confidence that it DOES work, and
very reliably!
I have been running a servo system with a 1 KHz servo update rate for
about 18 months
now, under Linux.
imagine spending
a week working on a complicated piece, and then have it ruined by a
stepper that has
quietly lost a few steps here and there, and the reference location is
off by some random
amount. Speed is a funny thing. I designed my servo system to go to
120 IPM, and
due to not quite enough voltage, I am happy to get about 90 IPM
reliably. That is
completely fine with me. I get the shakes when the machine moves much
faster than
that! Commercial CNC machining centers are now working on breaking the
1000 IPM
barrier. This just makes the inevitable crashes more spectacular!
Instead of a
$5,000 crash, it becomes a $100,000 crash.
you get the complete
source with it, so you can explore problems or add extensions, if you
have the programming
skills. If not, there are a lot of others out there who are working on
various parts of it,
so improvements will come, even if the creators drop interest in it
(which they don't
seem to be doing, either).
There are some problems in getting the full speed possible out of EMC,
but I'm
working on some stuff that will eventually lead to a better step pulse
generation
system, that will also allow closed loop operation. (There are already
some
other setups that allow closed loop operation with steppers.) This way,
even if
a motor stalls, it will not lose the position reference, and may be able
to get out
of the stall and continue.
Jon
> I have been watching and reading the email that was to the group forNot really. There are a number of people who are using some of this
> about a
> month now but I would like to have someone explain several questions I
> have:
>
> 1. Is the group for home shop or commercial people? most of the
> email I
> have read seems to be focused toward commercial machines.
equipment in a commercial endeavor, but many of them have their
shop at home. (Some of them, conversely, have their home in the shop,
or
would like to.)
> 2. Most of the software that is stated in the group seems to beNo, the stuff we are speaking about, like Bobcad/CAM ($500), Vector,
> commercial
> grade.
Rhino, Dolphin and similar things
are definitely NOT commercial. Gibbs, Unigraphics, Catia, etc. are the
commercial
packages, going for $30,000 for a complet suite, up to $25,000 for each
module,
and you need dozens of modules before you get the ability to do much!
> 3. What is "EMC" that a lot of people seem to be using. CommercialEMC is essentially free, a product of a US Governmental agency, the
> or
> freeware?
National Institute
of Science and Technology (NIST). It takes in RS-274D (G-code) commands
and moves
a machine (can be a mill, lathe, robot or hexapod) as directed. It also
has a manual
mode for aligning the axes with a reference point and simple cutting
operations.
> 4. Why is Linux being used instead of Windows or Dos on a PC?EMC was first put on Sun workstations as an intellectual exercise, and
then ported
to Windows NT. The "Real time" extensions to Win NT (at least at that
time) were
very poor, in the sense that interrupt latency was VERY high, and
expensive, too.
So, when a real-time extension was created for Linux, they tested it,
and it was
observed that the interrupt latency was MUCH lower than with Win NT, and
so they
ported it to RT-Linux. I can say with confidence that it DOES work, and
very reliably!
I have been running a servo system with a 1 KHz servo update rate for
about 18 months
now, under Linux.
> 5. Why is everyone interested in Servos and by passing steppers. TheThe problem with simple steppers is that they are open loop. I can't
>
> present mill retrofit that I am using uses steppers and have had no
> problem
> programming or running "G" code on it. The speed is appropiate and
> why is
> every one wanting speed.
imagine spending
a week working on a complicated piece, and then have it ruined by a
stepper that has
quietly lost a few steps here and there, and the reference location is
off by some random
amount. Speed is a funny thing. I designed my servo system to go to
120 IPM, and
due to not quite enough voltage, I am happy to get about 90 IPM
reliably. That is
completely fine with me. I get the shakes when the machine moves much
faster than
that! Commercial CNC machining centers are now working on breaking the
1000 IPM
barrier. This just makes the inevitable crashes more spectacular!
Instead of a
$5,000 crash, it becomes a $100,000 crash.
> I run a small home shop business and am using at present a CNC MillWell, EMC is a pretty flexible piece of software, it is open, meaning
> and am
> working on a retrofit for a lathe. I have Dan's 5 amp 3 axis board
> that I
> am going to use.
you get the complete
source with it, so you can explore problems or add extensions, if you
have the programming
skills. If not, there are a lot of others out there who are working on
various parts of it,
so improvements will come, even if the creators drop interest in it
(which they don't
seem to be doing, either).
There are some problems in getting the full speed possible out of EMC,
but I'm
working on some stuff that will eventually lead to a better step pulse
generation
system, that will also allow closed loop operation. (There are already
some
other setups that allow closed loop operation with steppers.) This way,
even if
a motor stalls, it will not lose the position reference, and may be able
to get out
of the stall and continue.
Jon
Discussion Thread
james owens
2000-07-07 14:54:45 UTC
Re: [CAD_CAM_EDM_DRO] purpose of the group
Jon Elson
2000-07-07 15:44:23 UTC
Re: [CAD_CAM_EDM_DRO] purpose of the group
Tim Goldstein
2000-07-07 15:48:46 UTC
RE: [CAD_CAM_EDM_DRO] purpose of the group
Doug Harrison
2000-07-07 15:55:34 UTC
Re: [CAD_CAM_EDM_DRO] purpose of the group
wanliker@a...
2000-07-07 16:37:29 UTC
Re: [CAD_CAM_EDM_DRO] purpose of the group
mSperry
2000-07-07 18:26:03 UTC
Re: [CAD_CAM_EDM_DRO] purpose of the group
ptengin@a...
2000-07-07 19:13:49 UTC
Re: [CAD_CAM_EDM_DRO] purpose of the group
wanliker@a...
2000-07-07 21:14:59 UTC
Re: [CAD_CAM_EDM_DRO] purpose of the group
Jon Elson
2000-07-07 23:23:59 UTC
Re: [CAD_CAM_EDM_DRO] purpose of the group
John Guenther
2000-07-08 04:42:25 UTC
RE: [CAD_CAM_EDM_DRO] purpose of the group
james owens
2000-07-08 12:27:27 UTC
Re: [CAD_CAM_EDM_DRO] purpose of the group
james owens
2000-07-08 12:49:16 UTC
Re: [CAD_CAM_EDM_DRO] purpose of the group
mSperry
2000-07-08 23:28:53 UTC
Re: [CAD_CAM_EDM_DRO] purpose of the group
wanliker@a...
2000-07-09 05:59:05 UTC
Re: [CAD_CAM_EDM_DRO] purpose of the group
Tim Goldstein
2000-07-09 09:30:39 UTC
RE: [CAD_CAM_EDM_DRO] purpose of the group
james owens
2000-07-09 11:19:02 UTC
Re: [CAD_CAM_EDM_DRO] purpose of the group
Larry Ruebush
2000-07-09 16:53:46 UTC
Re: [CAD_CAM_EDM_DRO] purpose of the group
Tim Goldstein
2000-07-09 17:32:40 UTC
Lost Postings, was: purpose of the group
mSperry
2000-07-09 19:57:45 UTC
Re: [CAD_CAM_EDM_DRO] purpose of the group
wanliker@a...
2000-07-09 22:08:44 UTC
Re: [CAD_CAM_EDM_DRO] purpose of the group
Jon Elson
2000-07-09 22:36:10 UTC
Re: [CAD_CAM_EDM_DRO] purpose of the group
Larry Ruebush
2000-07-10 07:25:04 UTC
Re: [CAD_CAM_EDM_DRO] purpose of the group
Rich
2000-07-10 08:00:05 UTC
Re: [CAD_CAM_EDM_DRO] purpose of the group
John D. Guenther
2000-07-10 09:51:19 UTC
Re: purpose of the group
james owens
2000-07-10 11:38:39 UTC
Re: [CAD_CAM_EDM_DRO] purpose of the group