Re: Linux vs. DOS
Posted by
Jon Elson
on 1999-05-29 21:21:47 UTC
Andrew Werby wrote:
set up the system for the first time. But when you get down to running
the machine, it really works quite reliably, with essentially no further
attention. The kernel is the basic center of the operating system. If
you don't like the way your Win-NT is configured, you are pretty well
out of luck. If you don't like the configuration of your linux kernel,
you can edit the configuration (graphically) and re-compile. It takes
less than 30 minutes on a Pentium. It would be a simple install, except
for the real-time patch. Apparently somebody is selling a CD with a
real-time patched RedHat Linux that will install directly, without you
having to install the patch. I should note there are people running
Linux that have had their machines up for YEARS without rebooting.
I'll also note that once I got the hardware all stable, I have had no
crashes with Linux. I have had earlier versions of the EMC program
that malfunctioned or hung up. The current version appears to be
ROCK SOLID!
they didn't work as well as the people at NIST wanted, and had a $2000
per machine license, I think.
I'm using X windows, and it is fine for what I'm doing. This is all on the
CD.
using the Unix user environment, or specifically Linux. Probably any
Barnes&Noble or B. Dalton's has such books.
out the printer port. For the safety of your PC, if not yourself, it is
pretty wise to use opto isolators. They aren't too expensive. A
Grayhill card and 5 or 6 opto-relays will cost about $50. These are
compatible with the Opto-22 modules, just a lot cheaper. You will
likely cobble together some combination of opto-relays and the existing
relays to run things. The machine tools I've seen seem to run most
of the control stuff off 110 Volts. The opto-relays are for pretty
small loads, up to 5 Amps or so. So, you often use the opto-relays
to control the coils of the bigger relays, especially for the spindle
motor.
smells like natural gas. My wife was always worried the house was going
to blow up. (It wasn't that strong, just detectable.) Well, we finally
had the gas log disconnected, and the smell still shows up under humid
conditions. All I can figure is some rainwater gets inside the thing, and
seeps into crevices, breeding anaerobic bacteria. This is the same thing
that makes coolant sumps stink, especially on Monday morning, after
the plant has been shut down and sealed over the weekend.
Does the machine have a coolant sump? If so, the sump or anyplace
stray coolant may have settled could be causing the smell.
Also, I';ll bet you're smelling H2S, not CH3SH (methyl mercaptan).
They smell somewhat alike (note the chemical similarity) and I think
the mercaptan is only produced under high temp, pressure or both,
like in an oil well.
control under NT or DOS. They do have some DOS demos, but they
are just loops, not real time.
Yes, initially, I didn't want to go through the learning curve of Linux,
but EMC showed REAL promise, even 18 months ago. I certainly
couldn't find any true servo package (from the ground up) that I could
afford. And, the idea of having servos controlled by step pulses from
a program that thinks it is driving steppers soulds almost worse than
just using steppers (which I hate, in this application). I couldn't be
more happy now with EMC. I did go through a LOT of hassle, but
I think the way has been cleared for those that follow.
Jon
> [Oh. I really don't know anything about Linux, but it seems intimidating.Well, there is somewhat of a learning curve, and a fair amount of work to
> All the problems with "kernels", whatever they are, seem like they would
> take over my life.
set up the system for the first time. But when you get down to running
the machine, it really works quite reliably, with essentially no further
attention. The kernel is the basic center of the operating system. If
you don't like the way your Win-NT is configured, you are pretty well
out of luck. If you don't like the configuration of your linux kernel,
you can edit the configuration (graphically) and re-compile. It takes
less than 30 minutes on a Pentium. It would be a simple install, except
for the real-time patch. Apparently somebody is selling a CD with a
real-time patched RedHat Linux that will install directly, without you
having to install the patch. I should note there are people running
Linux that have had their machines up for YEARS without rebooting.
I'll also note that once I got the hardware all stable, I have had no
crashes with Linux. I have had earlier versions of the EMC program
that malfunctioned or hung up. The current version appears to be
ROCK SOLID!
> I read some previous discussion here on the fine pointsNT is not a real time system. There are real-time extensions to NT, but
> of running Linux , and I didn't understand thing one. Is there no
> alternative? Does NT run in real time?
they didn't work as well as the people at NIST wanted, and had a $2000
per machine license, I think.
> Has anybody built a graphical userOf course! There are several, including a windows-like environment.
> interface for Linux that makes it more user-friendly?
I'm using X windows, and it is fine for what I'm doing. This is all on the
CD.
> Is there a Linux forI've heard of such a thing. You certainly need to at least get a book on
> Dummies?]
using the Unix user environment, or specifically Linux. Probably any
Barnes&Noble or B. Dalton's has such books.
> [Would that be better than leaving the old relays in place?The version of EMC I'm using sends all the auxilliary control signals
out the printer port. For the safety of your PC, if not yourself, it is
pretty wise to use opto isolators. They aren't too expensive. A
Grayhill card and 5 or 6 opto-relays will cost about $50. These are
compatible with the Opto-22 modules, just a lot cheaper. You will
likely cobble together some combination of opto-relays and the existing
relays to run things. The machine tools I've seen seem to run most
of the control stuff off 110 Volts. The opto-relays are for pretty
small loads, up to 5 Amps or so. So, you often use the opto-relays
to control the coils of the bigger relays, especially for the spindle
motor.
> Also, can youHmm, sounds like my fireplace. Whenever it gets humid, the fireplace
> tell me why there's a strong smell of methyl mercaptans in the mill's
> control box? If I didn't know better, I'd swear it was leaking natural
> gas...]
smells like natural gas. My wife was always worried the house was going
to blow up. (It wasn't that strong, just detectable.) Well, we finally
had the gas log disconnected, and the smell still shows up under humid
conditions. All I can figure is some rainwater gets inside the thing, and
seeps into crevices, breeding anaerobic bacteria. This is the same thing
that makes coolant sumps stink, especially on Monday morning, after
the plant has been shut down and sealed over the weekend.
Does the machine have a coolant sump? If so, the sump or anyplace
stray coolant may have settled could be causing the smell.
Also, I';ll bet you're smelling H2S, not CH3SH (methyl mercaptan).
They smell somewhat alike (note the chemical similarity) and I think
the mercaptan is only produced under high temp, pressure or both,
like in an oil well.
> [If you'd rather, we could take this off-list, but it seemed we were stillWell, I don't think STG really has any software that does true real-time
> discussing things of general interest, so I'm replying in "public". The 8
> axis Servo-to go card sounds interesting at $888 - it certainly beats
> paying $1500 per axis. They mention running NT in conjunction with their
> card, and something about how it (and DOS) can be induced not to interrupt-
> do you know anything about this? They also talk about writing your own
> algorithms, though- so I'm not sure how hard this would be. ]
control under NT or DOS. They do have some DOS demos, but they
are just loops, not real time.
Yes, initially, I didn't want to go through the learning curve of Linux,
but EMC showed REAL promise, even 18 months ago. I certainly
couldn't find any true servo package (from the ground up) that I could
afford. And, the idea of having servos controlled by step pulses from
a program that thinks it is driving steppers soulds almost worse than
just using steppers (which I hate, in this application). I couldn't be
more happy now with EMC. I did go through a LOT of hassle, but
I think the way has been cleared for those that follow.
Jon
Discussion Thread
Andrew Werby
1999-05-29 04:16:11 UTC
Linux vs. DOS
john@x...
1999-05-29 15:07:09 UTC
Re: Linux vs. DOS
Jon Elson
1999-05-29 21:21:47 UTC
Re: Linux vs. DOS
Matt Shaver
1999-05-29 21:29:06 UTC
Re: Linux vs. DOS
Tim Goldstein
1999-05-29 22:27:47 UTC
Re: Linux vs. DOS
Matt Shaver
1999-05-30 00:12:13 UTC
Re: Linux vs. DOS
Tim Goldstein
1999-05-30 15:39:08 UTC
Re: Linux vs. DOS
Tim Goldstein
1999-05-30 20:37:55 UTC
Re: Linux vs. DOS
Matt Shaver
1999-05-30 20:35:33 UTC
Re: Linux vs. DOS
Matt Shaver
1999-05-30 20:48:44 UTC
Re: Linux vs. DOS
Jon Elson
1999-05-30 23:19:07 UTC
Re: Linux vs. DOS
Tim Goldstein
1999-05-31 11:54:03 UTC
Re: Linux vs. DOS
Tim Goldstein
1999-05-31 11:54:07 UTC
Re: Linux vs. DOS
Matt Shaver
1999-05-31 12:12:57 UTC
Re: Linux vs. DOS
Matt Shaver
1999-05-31 12:17:39 UTC
Re: Linux vs. DOS
Tim Goldstein
1999-05-31 13:30:23 UTC
Re: Linux vs. DOS
Matt Shaver
1999-05-31 14:13:26 UTC
Re: Linux vs. DOS
Tim Goldstein
1999-05-31 16:00:38 UTC
Re: Linux vs. DOS
Matt Shaver
1999-05-31 16:00:51 UTC
Re: Linux vs. DOS
Tim Goldstein
1999-05-31 17:32:03 UTC
Re: Linux vs. DOS
Matt Shaver
1999-05-31 17:33:17 UTC
Re: Linux vs. DOS
Matt Shaver
1999-05-31 17:43:39 UTC
Re: Linux vs. DOS
Dan Falck
1999-05-31 19:55:12 UTC
Re: Linux vs. DOS
Matt Shaver
1999-05-31 22:48:46 UTC
Re: Linux vs. DOS
Jon Elson
1999-05-31 22:58:57 UTC
Re: Linux vs. DOS
Tim Goldstein
1999-05-31 23:32:27 UTC
Re: Linux vs. DOS
Tim Goldstein
1999-05-31 23:32:29 UTC
Re: Linux vs. DOS
Matt Shaver
1999-05-31 23:32:50 UTC
Re: Linux vs. DOS
Matt Shaver
1999-06-01 00:38:11 UTC
Re: Linux vs. DOS
Dan Mauch
1999-06-01 06:32:41 UTC
Re: Linux vs. DOS
Tim Goldstein
1999-06-01 09:45:52 UTC
Re: Linux vs. DOS
Jon Elson
1999-06-01 12:16:50 UTC
Re: Linux vs. DOS
Ian W. Wright
1999-06-01 12:37:43 UTC
Re: Linux vs. DOS
Tim Goldstein
1999-06-01 17:03:26 UTC
Re: Linux vs. DOS
Tim Goldstein
1999-06-01 22:05:54 UTC
Re: Linux vs. DOS
Dan Mauch
1999-06-02 06:30:25 UTC
Re: Linux vs. DOS
Dan Mauch
1999-06-02 06:42:14 UTC
Re: Linux vs. DOS
Buchanan, James (Jim)
1999-06-02 13:01:09 UTC
Re: Linux vs. DOS
Jon Elson
1999-06-02 13:33:07 UTC
Re: Linux vs. DOS
Tim Goldstein
1999-06-02 14:26:20 UTC
Re: Linux vs. DOS
Tim Goldstein
1999-06-02 14:45:01 UTC
Re: Linux vs. DOS
Matt Shaver
1999-06-02 15:01:17 UTC
Re: Linux vs. DOS
Tim Goldstein
1999-06-02 23:19:44 UTC
Re: Linux vs. DOS
Matt Shaver
1999-06-03 00:26:35 UTC
Re: Linux vs. DOS
Ian W. Wright
1999-06-04 13:47:19 UTC
Re: Linux vs. DOS
Tim Goldstein
1999-06-05 16:54:15 UTC
Re: Linux vs. DOS
Tim Goldstein
1999-06-05 17:34:22 UTC
Re: Linux vs. DOS
Jon Elson
1999-06-05 23:39:08 UTC
Re: Linux vs. DOS
Jon Elson
1999-06-05 23:41:51 UTC
Re: Linux vs. DOS
Matt Shaver
1999-06-05 23:42:39 UTC
Re: Linux vs. DOS
Tim Goldstein
1999-06-06 00:03:27 UTC
Re: Linux vs. DOS
Tim Goldstein
1999-06-06 00:30:52 UTC
Re: Linux vs. DOS
Matt Shaver
1999-06-06 01:00:46 UTC
Re: Linux vs. DOS
Matt Shaver
1999-06-06 01:37:57 UTC
Re: Linux vs. DOS
Dan Falck
1999-06-06 05:51:38 UTC
Re: Linux vs. DOS
Jon Elson
1999-06-06 17:16:33 UTC
Re: Linux vs. DOS
Tim Goldstein
1999-06-06 21:05:20 UTC
Re: Linux vs. DOS
Matt Shaver
1999-06-06 22:06:24 UTC
Re: Linux vs. DOS
Matt Shaver
1999-06-06 22:30:17 UTC
Re: Linux vs. DOS
Tim Goldstein
1999-06-06 22:57:18 UTC
Re: Linux vs. DOS
Matt Shaver
1999-06-06 23:16:03 UTC
Re: Linux vs. DOS
Jon Elson
1999-06-06 23:29:41 UTC
Re: Linux vs. DOS
Jon Elson
1999-06-06 23:46:32 UTC
Re: Linux vs. DOS
Tim Goldstein
1999-06-07 21:59:53 UTC
Re: Linux vs. DOS
Tim Goldstein
1999-06-08 22:07:54 UTC
Re: Linux vs. DOS
Dan Mauch
1999-06-09 06:18:23 UTC
Re: Linux vs. DOS