¿ªÔÆÌåÓý

ctrl + shift + ? for shortcuts
© 2025 Groups.io
Date

Re: Linux 6.0 - 5.2

William Scalione
 

Tim Goldstein wrote:

From: "Tim Goldstein" <timg@...>

Not sure about the ability to down grade RedHat from 6.0 to 5.2

The 2 versions use vastly different kernels, but it is worth a try if you
want to run EMC. If it works you are ready to go, if it fails you just
reinstall everything with 5.2 and are only out the few minutes it takes to
copy the kernel and edit LILO.

EMC does run step and direction signals out the parallel port in addition to
running servos from an addin card. I am currently running a Camtronics 5 amp
stepper controller using EMC.

Tim
[Denver, CO]

-----Original Message-----
From: William Scalione [mailto:scalione@...]
Sent: Friday, June 18, 1999 5:25 PM
To: CAD_CAM_EDM_DRO@...
Subject: [CAD_CAM_EDM_DRO] Linux 6.0 - 5.2


From: William Scalione <scalione@...>

I already have Linux 6.0 on a computer here and was wondering if it is
possible to downgrade (is that the proper term?) to linux 5.2 without
having to reload. Can I just copy over the kernal with the proper
version or am I asking for trouble?

Also EMC will work with a step and direction controller won't it?

Bill
Tim;

Thanks, I will try the kernal swap and see what happens.


Re: Linux 6.0 - 5.2

Dennis Mino
 

----- Original Message -----
From: Tim Goldstein <timg@...>
To: <CAD_CAM_EDM_DRO@...>
Sent: Friday, June 18, 1999 9:32 PM
Subject: Re: [CAD_CAM_EDM_DRO] Linux 6.0 - 5.2


From: "Tim Goldstein" <timg@...>

Not sure about the ability to down grade RedHat from 6.0 to 5.2

The 2 versions use vastly different kernels, but it is worth a try if you
want to run EMC. If it works you are ready to go, if it fails you just
reinstall everything with 5.2 and are only out the few minutes it takes to
copy the kernel and edit LILO.

EMC does run step and direction signals out the parallel port in addition
to
running servos from an addin card. I am currently running a Camtronics 5
amp
stepper controller using EMC.
Tim what exactly are you using the Camtronics stepper controller and EMC
on??? I'm sure you had mentioned it in the past....... Also, whose motors
are you using? oz-in.?? So your just coming off the parallel port? What
addin card for the servo's?
Thanx.



Tim
[Denver, CO]

-----Original Message-----
From: William Scalione [mailto:scalione@...]
Sent: Friday, June 18, 1999 5:25 PM
To: CAD_CAM_EDM_DRO@...
Subject: [CAD_CAM_EDM_DRO] Linux 6.0 - 5.2


From: William Scalione <scalione@...>

I already have Linux 6.0 on a computer here and was wondering if it is
possible to downgrade (is that the proper term?) to linux 5.2 without
having to reload. Can I just copy over the kernal with the proper
version or am I asking for trouble?

Also EMC will work with a step and direction controller won't it?

Bill

------------------------------------------------------------------------
What do lizards and rock music have in common?

They both have communities at ONElist. Find yours today!
------------------------------------------------------------------------
welcome to CAD_CAM_EDM_DRO@..., an unmodulated list for
the discussion of shop built systems in the above catagories.
To Unsubscribe:

bill, List Manager

------------------------------------------------------------------------
How many communities do you think join ONElist each week?

More than 5,000! Create yours now!
------------------------------------------------------------------------
welcome to CAD_CAM_EDM_DRO@..., an unmodulated list for the
discussion of shop built systems in the above catagories.
To Unsubscribe:

bill, List Manager


Re: MaxNC modifications

Tim Goldstein
 

I never traced the exact source of the interference. I just shielded all the
wires from my junction board to the controller step, direction and ground
pins and when that worked I moved quickly to makin chips!!

I should have phrased my final statement differently as it was really adding
shielding, not more shielding that cured the problem as I originally had
unshielded hookup wire .

I don't think it was the parallel cable as that has not changed, I really
think it was just the wiring inside the controller case. The way I have it
set up is the case has a centronics connector bolted to it and I plug the
printer cable into it. From the back side of the centronics connector I have
22 ga hookup wire running to the ramped headers on the circuit board.

Tim
[Denver, CO]

-----Original Message-----
From: Mo [mailto:mo@...]
Sent: Thursday, June 17, 1999 4:43 PM
To: CAD_CAM_EDM_DRO@...
Subject: Re: [CAD_CAM_EDM_DRO] MaxNC modifications


From: "Mo" <mo@...>

Tim,
Did you every manage to pin down where the interference was coming from
or what made the most difference?
are all cables now shielded?
What was the "more shielding" that smoothed out the motor and did you find
the victim to be the parallel port cabling?
Mo



I don't know if your problem is the same, but I had the same
thing happen
to
me as I increased the amps I started getting problems with lost steps on
curves. It was particularly bad with about a 1/2" radius and
moving around
7
ipm. The step loss was quite consistent on the parts with curves, but I
could do slow or rapid moves along an axis all day long and the
moves were
perfect with no steps lost. It ended up being a shielding problem. I
thought
the interference was coming from the mosfets & heatsinks, but Jon made a
post and thinks it is from the circuit board (with his knowledge and
understanding I don't have much reason to disagree). Dan Falck posted
about
having a similar problem as he upped the juice on his system also.

The fact that your loss of steps increased as you upped the
amps makes me
think you are experiencing the same type of thing. When I was lossing
steps
I also would notice that the motor was running very ragged.
More shielding
caused my step loss to go away and the motor started sounding much
smoother
running.


------------------------------------------------------------------------
Make your voice heard!

Sign up for the ONElist Weekly Survey. Go to homepage for details.
------------------------------------------------------------------------
welcome to CAD_CAM_EDM_DRO@..., an unmodulated list for
the discussion of shop built systems in the above catagories.
To Unsubscribe:

bill, List Manager


Re: Linux 6.0 - 5.2

Tim Goldstein
 

Not sure about the ability to down grade RedHat from 6.0 to 5.2

The 2 versions use vastly different kernels, but it is worth a try if you
want to run EMC. If it works you are ready to go, if it fails you just
reinstall everything with 5.2 and are only out the few minutes it takes to
copy the kernel and edit LILO.

EMC does run step and direction signals out the parallel port in addition to
running servos from an addin card. I am currently running a Camtronics 5 amp
stepper controller using EMC.

Tim
[Denver, CO]

-----Original Message-----
From: William Scalione [mailto:scalione@...]
Sent: Friday, June 18, 1999 5:25 PM
To: CAD_CAM_EDM_DRO@...
Subject: [CAD_CAM_EDM_DRO] Linux 6.0 - 5.2


From: William Scalione <scalione@...>

I already have Linux 6.0 on a computer here and was wondering if it is
possible to downgrade (is that the proper term?) to linux 5.2 without
having to reload. Can I just copy over the kernal with the proper
version or am I asking for trouble?

Also EMC will work with a step and direction controller won't it?

Bill

------------------------------------------------------------------------
What do lizards and rock music have in common?

They both have communities at ONElist. Find yours today!
------------------------------------------------------------------------
welcome to CAD_CAM_EDM_DRO@..., an unmodulated list for
the discussion of shop built systems in the above catagories.
To Unsubscribe:

bill, List Manager


Linux 6.0 - 5.2

William Scalione
 

I already have Linux 6.0 on a computer here and was wondering if it is
possible to downgrade (is that the proper term?) to linux 5.2 without
having to reload. Can I just copy over the kernal with the proper
version or am I asking for trouble?

Also EMC will work with a step and direction controller won't it?

Bill


Re: Linux for Beginners

Ray Henry
 

From: Bob Bachman <bobach@...>

I have been struggling to learn enough Linux to setup EMC. This is not a
learning >- more like a long uphill climb!!

I know the feeling and I had a little experience with unix. I was there
about two months ago. My most recent solution was to clean the hard drive
and install win95 in a 1 gig or so partition. Leave the rest of the space
on the hard drive alone. Get the computer running just the way you want.
If you have a winmodem you can make it work here.

Now start up the Red Hat 5.2 from a floppy. It should find the cd rom
drive and let you install the whole Linux system. When you get to the disk
druid it will recognize the dos partition that already exists. Leave that
part alone and go ahead and make your Linux partitions the way you want
them. When you come to lilo and the boot partition question use the dos one.

When you reboot, you should see the bios and mem stuff and then the lilo
prompt. If you want win95 you can type in whatever you called it during
the linux install there or {enter} for Linux.

Now use windows to download the tar file from NIST. I save it in the C:/
directory. Boot the linux and you should find the file under the /dos
directory. Tar it to where you want it and away you go.

That's harder than Jon E's way but I don't have a net yet.

I'll go over my notes on the patch stuff -- I had a bitch of a time cause
something was wrong with the file I got -- and get it to you asap.

Ray


Re: newbie question

Jon Elson
 

Andy Olney wrote:

I was surprised to learn that the antibacklash type ballscrews with 2
ballnuts back to back have a much reduced efficiency. Standard ball screws
are in the 90% range and I think the anti backlash type are in the 60% range
which does away with some of the ballscrew's advantage.
This is probably very dependent on pitch, the amount of preload, size
of the balls, whether the screw is rolled or ground, etc. I will say that the
static friction of the NSK ballscrews that I adapted to my Bridgeport
is EXTREMELY low, in the range of 1-2 Oz-In! This is not literally
2 ballnuts back to back, but a single block with two ball tracks ground
into the interior, with a slight offset between the two track's pitch.
It is preloaded by putting the right size balls in the tracks, after VERY
precise grinding to size.

I have a similar ballscrew/nut set, I think by HiWin, but it is a smaller
.1" pitch unit that I use for the Z axis. It was new, surplus (no more of
these available, sorry). It has much higher drag, about 5 Oz-in or so.
That is still quite low. So, it may be that some anti-backlash ballnuts
may have low efficiency, but it is not necessarily true. Anyway, the main
reason for using them is the freedom from wear and backlash.

Jon


Re: Lurking... and learning

Jon Elson
 

Matt Shaver wrote:

From: "Matt Shaver" <mshaver@...>

From: Jon Elson <jmelson@...>
Maybe, someday, we'll even have
conversational programming, but I'm making parts, and loving every
minute of it!
Actually, if we made a button or menu item in xemc that would call your
little gcode generators (maybe even put a GUI dialog box in front of yours),
we'd have it now! Sure, there's no graphical toolpath display...
Hmmm, yes! I think it would be best to have a user-configurable menu,
which calls up these programs with a spawn process, so that if you manage
to crash one of these programs, it doesn't take EMC down, you just get
a message that the little program crashed. It could always use the same
file name, and give you an option to save the file (by EMC). If you
didn't want to save it, then whenever you ran another code generator,
or opened a G-code file, the one you just created would be gone.
That would be fine for lots of one-off little things like boring an odd size
hole.

The user-configurable menu would let you add programs as you thought of
things. It would really only need to know what you called the function,
and the file name of the actual program. So, you could see a menu entry
like 'Bore Through Hole of arbitrary size', and the program would be
'/usr/elson/writeRS274/makebore'.

A more complex way of doing things (and making it user-extendable
would be much tougher) would be for Xemc to have a menu presentation
for each program, take all the menu entries for it, and then pass that in
the command string. The advantage for this would be that all the menu
entries could remain, so you could just change one entry and the re-run
it. But, the simpler approach sounds like it would be a big plus.

Jon


Re: List question

Jon Elson
 

Ted wrote:

From: "Ted" <rtr@...>

rtr@...
maybe I'm missing something, but integrating the movements of a couple of
axes is what Bresham's Algorithms in linear and circular motion is about.
It breaks down the motion into a lot of very small motions and steps the
control through these small motions. It never allows the axes to loose
synchronization by more than the resoplution of the machine, if you choose
to limit the small motions to that size.
Oh, yes, of course. But, that is to be run on a single processor. You
can't have two separate processors each running part of the Bresenham
algorithms. So, a distributed system would run the Bresenham circle
generator, for instance, and send a stream of small, incremental moves
to 2 axis control processors. What if those axis processors get out of
sync by a few of those incremental moves? How does the central
computer keep them in sync? It can be done, but it is a lot more
complicated than just throwing the move commands down a blind
pipeline, and hoping everyone is on the corresponding move
simultaneously.

As for resolution, my machine has resolution of .00005" in the X and Y,
and .000025" in the Z. If I was moving X at 120 IPM (= 2 IPS), that
is 20,000 incremental moves per second, which is beyond the
capability of most systems to handle giving 3 (or more) axes
their marching instructions that fast. So, this sceme is really not practical
for systems that have a relatively fast motion capability.

Jon


Lunix 5.2

 

I purchased LINUX 5.2 , Deluxe and Secure version to day at BEST BUY in Alb.
39.95 with a 20.00 rebate coupon from REDHAT, Final cost 19.95.
For right now I am just going to load LINUX, and play with it to learn a bit
about it.
Question for the list, does anyone know if AOL will work with LINUX??
bill
List Manager


Re: LM628

me
 

Mike Romine wrote:

But given how fast and cheap the 16 and 32 bit microcontrollers are
getting now-a-days, I have come to believe closing the loop in firmware
on a processor is the way to go for a ground up system.
I talked to tech80 about the PMD chipset as they also use it. They seemed to prefer the LM628/629
though. In position mode I usually got trajectory updates at several milliseconds but they insist
that < 1 millisecond can be done with their low level routines. I like the 100 microseconds or
so quoted for EMC much better though! Even 1 millisecond at the high feed rates of a cnc router
could result in several thousanths position error in interpolated moves.

Leslie M. Watts
ITW Advanced Technology Center


Re: Injection Molding Machines

Dan Mauch
 

I was lurking behind the scenes build an injection molding machine designed
by Vince Gingery. If you need a relatively simple machinme to do small run
injection molded parts then I would recommend building it. The book was from
www.lindsaybooks.com and cost $15.
The instructions are good. It required a lathe, a mill, a bandsaw, a welder,
an oxy-aceteleyne torch, a drill press.
The first part that I molded came out perfect.
I may design another one that is much simpler to build and provide the
injection heater housing which would be difficult to make without a lathe
capable of machining a 1-1/2X1X4 steel stock.
Next week the DRO boards should be done and after I assemble one to verify
everything works properly will make the boards available first to the people
that have allready ordered them then to any one else if there are any left.
I'll also check on making a complete kit available.
Dan


Re: Real time operating systems

 

At 07:25 PM 6/17/99 -0700, you wrote:
From: Mike Romine <mromine@...>
Wouldn't it
be fairly simple to create a FIFO memory buffer between the computer and
motor controller such that it buffers motion commands from a
non-realtime computer and issues them in real time? The FIFO would not
even need to be very deep if you had a feedback to the computer to tell
it "I'm full now, stop transmitting".
Interesting idea, Mike. You would also want some direct, non-queued control
of motor on/off for emergencies, hardwired to a panic stop routine. It begs
the question of why real-time control is needed when we're talking to a
servo controller board, which I thought would (or could be made to) handle
the real time requirements.

Phil Plumbo


notice from ONELIST

 

The following is a posting I received from ONELIST today. I have no way of
knowing if it will cause any problems for us, so please be aware.

<
The ONELIST web site will be down for scheduled maintenance Sunday evening,
June 20, 1999. The web site will be off-line from 8pm to 10pm Pacific Time,
while we upgrade the system and add new features.

Email delivery will not be affected during this time.

Thanks for using ONElist! >
bill
List Manager


Re: Linux for Beginners

 


From: Jon Elson <jmelson@...>

From: Bob Bachman <bobach@...>


Jon,

Thanks for the response. How do you connect the two boxes? Serial? Network
cards?
I have some WD network cards which are supported by Linux.
Yes, I got PCI Ethernet cards for most of the machines, but the CNC
machine has an ISA card. I chose thinwire ethernet (10-base-5) over
twisted pair (10-base-T) because it didn't require a hub, and I knew
the technology well. Actually, I have at least 6 machines that could,
potentially, be hooked up.
You can hook two 10baseT cards together without a hub with a special cable that
reverses one pair of wires. On the other hand, you can buy a pocket hub
for $29 these days.

I used to like thinnet, but given today's technology it's more trouble
than it's worth. We ended up ripping it out of our university labs
when I ran them and we recently got rid of the last bits on our
internal net. If you get a problem with one cable it will take down
the entire segment. Not a big deal if you only have 2 computers
connected, but a real pain if there are more and you have to try to
track down the bad cable.

The other advantage with 10baseT is you can upgrade to 100 Mb ethernet
by swapping the hub if you use 10/100 cards in the nodes. (Of course,
you have to use cat5 wiring, but that's not a big deal either)

As for ethernet cards, we've had very good experience with the LinkSys
Etherfast 10/100 cards. They work great in both Linux and Windows, are
reasonably cheap and will work on both 10 Mb and 100 Mb ethernet. In
the FWIW department, they're based on the DEC tulip ethernet chip which
has a long sucessful track record.

Paul

--
Paul Amaranth | Rochester MI, USA
Aurora Group, Inc. | Software Development
paul@... | Unix / C / Tcl-Tk


Re: Pic servos

Andy Olney
 


[This looks interesting, Dan. Have you ever tried these out for
controlling
servos? Has anybody here? I assume the $175 is per axis, but still, it
doesn't look that bad if that's all I'll need to retrofit my old mill to
accept unlimited g-code file sizes. Could it be that simple? Are these the
kind of boards that have their own chips onboard, so they can buffer the
input from the computer and real-timeness isn't an issue? Or is there
something I'm not taking into consideration here? I wonder how well the
control software that goes with these things really works...]


Andrew Werby
These boards have their own control language running over a serial port and
you need some kind of software to generate that from gcode. Kerr has a
stepper adaptor that will fool the board into accepting stepper input but
Dan Mauch says it slow. Since to look is closed in hardware on the board the
real time requirements are not as much an issue. Jeff Kerr says it is
possible to tell 4 axis what to do about 400 times per second max, which
seems like it should be adequate.

Andy Olney


Re: Pic servos

Dan Mauch
 

I have a pic servo controller that I tested. With it I have the rs232 link
and the pic step board for running the pic servo from the parallel port.
Here is how it works.
Using the supplied software from jrkerr you connect the serial port on your
pc to the serial interface board. You set the PID filter parameters and
enable the servo. Then you disconnect the serial cable and can run dos based
g code interpreter that communicate via the pic step to the servo board. As
stated previously, in another message. It runs slow on a 486-33 using the
parallel port. Using the serial port it is fast but you would have to write
you own software for controlling multiple axis. What is need is the rate
multiplier that he is comming out with. That should speed it up a bunch.
Dan

-----Original Message-----
From: Andrew Werby <drewid@...>
To: CAD_CAM_EDM_DRO@... <CAD_CAM_EDM_DRO@...>
Date: Thursday, June 17, 1999 9:30 PM
Subject: [CAD_CAM_EDM_DRO] Pic servos


From: Andrew Werby <drewid@...>

"Dan Mauch" <dmauch@...>wrote:
Subject: Re: List question

see www.jrkerr.com
It is a programmable controller chip call the PIC servo that does the
intelligence for a low cost servo system.
Dan
-----Original Message-----
From: TADGUNINC@...
m <TADGUNINC@...>
To: CAD_CAM_EDM_DRO@... <CAD_CAM_EDM_DRO@...>
Date: Thursday, June 17, 1999 6:40 AM
Subject: Re: [CAD_CAM_EDM_DRO] List question


From: TADGUNINC@...

Sorry newbe question.....what is a pic board/chip?
Tracey
[This looks interesting, Dan. Have you ever tried these out for controlling
servos? Has anybody here? I assume the $175 is per axis, but still, it
doesn't look that bad if that's all I'll need to retrofit my old mill to
accept unlimited g-code file sizes. Could it be that simple? Are these the
kind of boards that have their own chips onboard, so they can buffer the
input from the computer and real-timeness isn't an issue? Or is there
something I'm not taking into consideration here? I wonder how well the
control software that goes with these things really works...]


Andrew Werby

Andrew Werby

Andrew Werby - United Artworks
Sculpture, Jewelry, and Other Art Stuff




------------------------------------------------------------------------
How many communities do you think join ONElist each week?

More than 5,000! Create yours now!
------------------------------------------------------------------------
welcome to CAD_CAM_EDM_DRO@..., an unmodulated list for the
discussion of shop built systems in the above catagories.
To Unsubscribe:

bill, List Manager


Re: newbie question

Andy Olney
 

Yes, ball nut is the proper term. Yes, there are inner 'works'.
The works are ball bearings. So, the nut does not slide on the screw,
the rolling balls make it a rolling action, thereby reducing friction (and
wear) to minute levels. There is a rounded track in the nut, and the
balls roll to one end of the track, so a recirculating tube of some
sort is provided to return the balls back to the other end of the
track. If you put two separate circuits of balls in one nut, and either
make the two tracks slightly off the screw pitch, or put springs
between the two nut sections, you can preload the two nuts against
each other. This makes an anti-backlash screw. Depending on
design, you can get screws that will deliver thousands of pounds
of linear force, with backlash on the order of .0001" or so.
Obviously, for a precision machine, or a CNC where smooth,
complex motions are required, removing backlash is a very
desirable thing. The reduction in wear and friction is an added
benefit.

Jon
I was surprised to learn that the antibacklash type ballscrews with 2
ballnuts back to back have a much reduced efficiency. Standard ball screws
are in the 90% range and I think the anti backlash type are in the 60% range
which does away with some of the ballscrew's advantage.


mouse as dro

 

to whoever asked about using a mouse as a dro device, it is something
I have thought about
before now.

I think the basic mechanics are more useful than the whole unit
though.
at this page

you'll see how a chap used the encoders for dro purposes.

the code wheels them selves are not very high resolution, 200 to 400
dpi begin common, when all the mechanics are taken into account.
certainly there are not many slots in the wheel....

using a friction drive (no backlash, and provides mechanical
muliplication of resolution), I don't see any big reasons why one
should not use these mechanisms if you cannot afford anything else.
by using careful design to replace expensive encoders, I'm sure it can
be done.

steam and wind

--
David Forsyth DaForce A-T Iwr.Ru.Ac.Za
Keeper of the listserver for South African Railways fans _|_ His
Part time gricer, kiter, photographer, father etc etc | Way
| Up


Stepster

Charles Gallo
 

Hi,
Oops - Dave found a serious bug last night - it'll be a while before I
ship out some more copies

Charlie

PGP Key Avalible upon request