discuss@lists.openscad.org

OpenSCAD general discussion Mailing-list

View all threads

has anyone worked on a Laser Design tool similar to the one for 3D printing?

JD
John David
Sat, Jun 28, 2025 11:18 PM

I had looked at klipper long ago, but never looked at it that deeply.
Something else to look into.  Also, that dremel blade sharpener trick is a
nice deal.  Thank you for sharing.

Gene, you probably know/remember.  I was an official developer for
LinuxCNC.  I finally completely gave up on it due to the systematic
bit-rot.  I remember one time I proposed a change (I forget what it was
specifically, this was circa 2003), but I do remember the screams of "DO
NOT TOUCH IT!"  To be fair, a couple of the LinuxCNC members explained
off-list how it took them sometimes a year or more to get a machine
stabilized after a major update.  Basically the entire community would get
something that works for small groups of the membership, and then they
would maintain that on their personal equipment -- until they had to
upgrade, or some feature was so desirable that it was worth spending the
time to upgrade. So, I ended up mostly hacking what I needed for my
personal projects and sending back minor bugs.  I had great hopes with the
advent of MachineKit.  But, over the pandemic, I ended up spending 100's of
hours and was never able to get the versions available at the time to even
fully run!  That is when I completely gave up on LinuxCNC.  Now if I
compare that with something like FluidNC, the code is clean, modern, and
has a more progressive support community.  That said, the main code base
does not support backlash, and is something I would have to bring back in
from a previous branch.  As we all know, with old iron, unmanaged backlash
is death. I would early love if LinuxCNC could be made viable, but my
personal experience has demonstrated repeatedly that it is not past taking
the time to get a single machine configuration working on very
specific hardware...

I cannot remember if I posted about it here, but a buddy of mine and I
hacked a printer driver onto a BeagleBone Black, with some filtering
programs to change the dialect of g-code produced by SolidWorks, and hooked
the BBB into the tape port of a 1989 Cincinnati Milicron lathe.  It also
managed the tape's forward and reverse, so that it could page in blocks of
g-code.  It should like I should be able to do the same thing with my
Creality Falcon-2 laser cutter.  That will be a background project, but one
that would help save some time in the long run...

EBo --

On Sat, Jun 28, 2025 at 5:33 PM gene heskett via Discuss <
discuss@lists.openscad.org> wrote:

On 6/28/25 12:42, John David via Discuss wrote:

William, gcpdxf.py
https://github.com/WillAdams/gcodepreview/blob/main/gcpdxf.py looks
interesting.  I will need to take a closer look at it.  I have been
focusing on SVG because I have other tools that worked, but at the time I
could not get the DXF to work properly.  But, I do not inherently care

if I

use SVG or DXF for the format as long as the various interface option
(LaserGRBL, LightBurn, Inkscape, etc.) all play nice.  I will reach out

to

you off-list about both providing 2D DXF test CAD files. I will probably
start with a quick-release jig fixture for the Falcon-2 - it would be

nice

to have the lockable jig plates also have engraved rulers...

Gene, I'll put shoehorning a workflow into OctoPrint on to my ToDo list.
As far as I know, both LaserGRBL and LightBurn natively uses g-code. I
think it is more an issue of figuring out how to set up the

configuration.
Generally that assessment is good. I'm in favor of klipper, which
replaces both merlin and octoprint, with something that to me is
reconfigurable and a lot easier to extend to make the printer into a
send it a file, go get the part work flow. No baby sitting during the
warmup to clean nozzles and such. I've made the printer do that by
itself. My biggest problem is time, at 90 I'm running out of time to do
the things I can dream up.

EBo --

On Sat, Jun 28, 2025 at 11:18 AM William F. Adams willadams@aol.com

wrote:

On Sat, Jun 28, 2025 at 6:22 AM John David via Discuss <

OpenSCAD has a Design tool for 3D printing.  Has anyone worked on one

for laser cutters/engravers?

On Saturday, June 28, 2025 at 01:48:04 AM EDT, Guenther Sohler via

Discuss

Did you check out

https://github.com/WillAdams/gcodepreview

Its based on openscad and it can help you to display the gcodes

This runs in a couple of modes, and is fully programmatic (but that

should

be okay for folks on this list).

https://github.com/WillAdams/gcodepreview/blob/main/gcpdxf.py

is a basic Python file which will make a multi-colour DXF (uploaded a
converted SVG version --- let me know if you have trouble making the

DXF)

I'm currently deep into a rewrite which flattens out the (currently)

ever

deeply more nested CSG unions and incorporating better support for

colour

should dovetail well with that.

If someone would provide an example DXF file (where different colours

are

coded to different laser operations) and pseudocode for creating that

file,

I'll do my best to at least keep that in mind as I work on the re-write.

William

--
Sphinx of black quartz, judge my vow.
https://designinto3d.com/


OpenSCAD mailing list
To unsubscribe send an email to discuss-leave@lists.openscad.org

Cheers, Gene Heskett, CET.

"There are four boxes to be used in defense of liberty:
soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author, 1940)
If we desire respect for the law, we must first make the law respectable.

  • Louis D. Brandeis

OpenSCAD mailing list
To unsubscribe send an email to discuss-leave@lists.openscad.org

I had looked at klipper long ago, but never looked at it that deeply. Something else to look into. Also, that dremel blade sharpener trick is a nice deal. Thank you for sharing. Gene, you probably know/remember. I was an official developer for LinuxCNC. I finally completely gave up on it due to the systematic bit-rot. I remember one time I proposed a change (I forget what it was specifically, this was circa 2003), but I do remember the screams of "DO NOT TOUCH IT!" To be fair, a couple of the LinuxCNC members explained off-list how it took them sometimes a year or more to get a machine stabilized after a major update. Basically the entire community would get something that works for small groups of the membership, and then they would maintain that on their personal equipment -- until they *had* to upgrade, or some feature was so desirable that it was worth spending the time to upgrade. So, I ended up mostly hacking what I needed for my personal projects and sending back minor bugs. I had great hopes with the advent of MachineKit. But, over the pandemic, I ended up spending 100's of hours and was never able to get the versions available at the time to even fully run! That is when I completely gave up on LinuxCNC. Now if I compare that with something like FluidNC, the code is clean, modern, and has a more progressive support community. That said, the main code base does not support backlash, and is something I would have to bring back in from a previous branch. As we all know, with old iron, unmanaged backlash is death. I would early love if LinuxCNC could be made viable, but my personal experience has demonstrated repeatedly that it is not past taking the time to get a single machine configuration working on very specific hardware... I cannot remember if I posted about it here, but a buddy of mine and I hacked a printer driver onto a BeagleBone Black, with some filtering programs to change the dialect of g-code produced by SolidWorks, and hooked the BBB into the tape port of a 1989 Cincinnati Milicron lathe. It also managed the tape's forward and reverse, so that it could page in blocks of g-code. It should like I should be able to do the same thing with my Creality Falcon-2 laser cutter. That will be a background project, but one that would help save some time in the long run... EBo -- On Sat, Jun 28, 2025 at 5:33 PM gene heskett via Discuss < discuss@lists.openscad.org> wrote: > On 6/28/25 12:42, John David via Discuss wrote: > > William, gcpdxf.py > > <https://github.com/WillAdams/gcodepreview/blob/main/gcpdxf.py> looks > > interesting. I will need to take a closer look at it. I have been > > focusing on SVG because I have other tools that worked, but at the time I > > could not get the DXF to work properly. But, I do not inherently care > if I > > use SVG or DXF for the format as long as the various interface option > > (LaserGRBL, LightBurn, Inkscape, etc.) all play nice. I will reach out > to > > you off-list about both providing 2D DXF test CAD files. I will probably > > start with a quick-release jig fixture for the Falcon-2 - it would be > nice > > to have the lockable jig plates also have engraved rulers... > > > > Gene, I'll put shoehorning a workflow into OctoPrint on to my ToDo list. > > As far as I know, both LaserGRBL and LightBurn natively uses g-code. I > > think it is more an issue of figuring out how to set up the > configuration. > Generally that assessment is good. I'm in favor of klipper, which > replaces both merlin and octoprint, with something that to me is > reconfigurable and a lot easier to extend to make the printer into a > send it a file, go get the part work flow. No baby sitting during the > warmup to clean nozzles and such. I've made the printer do that by > itself. My biggest problem is time, at 90 I'm running out of time to do > the things I can dream up. > > > > EBo -- > > > > On Sat, Jun 28, 2025 at 11:18 AM William F. Adams <willadams@aol.com> > wrote: > > > >>> On Sat, Jun 28, 2025 at 6:22 AM John David via Discuss < > >> discuss@lists.openscad.org> wrote: > >>>> OpenSCAD has a Design tool for 3D printing. Has anyone worked on one > >> for laser cutters/engravers? > >> > >> On Saturday, June 28, 2025 at 01:48:04 AM EDT, Guenther Sohler via > Discuss > >> <discuss@lists.openscad.org> wrote: > >> > >>> Did you check out > >>> > >>> https://github.com/WillAdams/gcodepreview > >>> > >>> Its based on openscad and it can help you to display the gcodes > >> This runs in a couple of modes, and is fully programmatic (but that > should > >> be okay for folks on this list). > >> > >> https://github.com/WillAdams/gcodepreview/blob/main/gcpdxf.py > >> > >> is a basic Python file which will make a multi-colour DXF (uploaded a > >> converted SVG version --- let me know if you have trouble making the > DXF) > >> > >> I'm currently deep into a rewrite which flattens out the (currently) > ever > >> deeply more nested CSG unions and incorporating better support for > colour > >> should dovetail well with that. > >> > >> If someone would provide an example DXF file (where different colours > are > >> coded to different laser operations) and pseudocode for creating that > file, > >> I'll do my best to at least keep that in mind as I work on the re-write. > >> > >> William > >> > >> -- > >> Sphinx of black quartz, judge my vow. > >> https://designinto3d.com/ > >> > > > > _______________________________________________ > > OpenSCAD mailing list > > To unsubscribe send an email to discuss-leave@lists.openscad.org > > Cheers, Gene Heskett, CET. > -- > "There are four boxes to be used in defense of liberty: > soap, ballot, jury, and ammo. Please use in that order." > -Ed Howdershelt (Author, 1940) > If we desire respect for the law, we must first make the law respectable. > - Louis D. Brandeis > _______________________________________________ > OpenSCAD mailing list > To unsubscribe send an email to discuss-leave@lists.openscad.org >
GH
gene heskett
Sun, Jun 29, 2025 3:38 AM

On 6/28/25 19:18, John David wrote:

I had looked at klipper long ago, but never looked at it that deeply.
Something else to look into.  Also, that dremel blade sharpener trick is a
nice deal.  Thank you for sharing.

Gene, you probably know/remember.  I was an official developer for
LinuxCNC.  I finally completely gave up on it due to the systematic
bit-rot.  I remember one time I proposed a change (I forget what it was
specifically, this was circa 2003), but I do remember the screams of "DO
NOT TOUCH IT!"  To be fair, a couple of the LinuxCNC members explained
off-list how it took them sometimes a year or more to get a machine
stabilized after a major update.  Basically the entire community would get
something that works for small groups of the membership, and then they
would maintain that on their personal equipment -- until they had to
upgrade, or some feature was so desirable that it was worth spending the
time to upgrade. So, I ended up mostly hacking what I needed for my
personal projects and sending back minor bugs.  I had great hopes with the
advent of MachineKit.  But, over the pandemic, I ended up spending 100's of
hours and was never able to get the versions available at the time to even
fully run!  That is when I completely gave up on LinuxCNC.  Now if I
compare that with something like FluidNC, the code is clean, modern, and
has a more progressive support community.  That said, the main code base
does not support backlash, and is something I would have to bring back in
from a previous branch.  As we all know, with old iron, unmanaged backlash
is death. I would early love if LinuxCNC could be made viable, but my
personal experience has demonstrated repeatedly that it is not past taking
the time to get a single machine configuration working on very
specific hardware...

I have 4 machines running linuxcnc, started back quite some time before
the name change from emc2.

I long ago learned to write my own gcode, and hal stuff, the only
problem I ever had was something that should have been fixed about the
time the axis gui was written. It did not properly change the focus of
he teeny little axis buttons to correspond to register with the last
axis moved, so touch-offs were often inadvertently applied to the wrong
axis, making one start from scratch rehoming the machine. And it
litterally took years, 3 or 4, from my reporting that bug to getting it
fixed. Other than that, I have been able to make a machine do exactly as
I wanted it to do.  In 20+ years, I haven't noticed any bit rot that
wasn't fixable by correcting an error in my .hal files.  I also am not a
machinist other than making parts to keep a television transmitter on
the air for 50 years beyond its use by date, I am a CET, I can and have
taught papered EE's stuff they should have learned in school, but the
prof skipped it because he didn't understand it himself. Or maybe
miss-understood it and taught it wrong.  I am amazed at the folks who
think Newtons 3rd law is optional, or that relativity doesn't apply to
the real world.

Please understand also that LinuxCNC is a 100% volunteer effort by
people who COULD be doing something generating a paycheck, most have a
$dayjob that might not even be linuxcnc related but that is not how
things started by NIST 50 or 60 years ago works. I have even offered
hardware to expand things, been turned down, there is not I have been
told, a mechanism in place that has a bank account. That means there is
no one an MBA can sue, a limiting factor in its popularity because MBA's
MUST have somebody to sue when their ideas wind up in the toilet. Its no
secret that I consider MBA's in the same category as lawyers when
Shakespear wrote "first we kill all the lawyers".  OTOH that is what you
get from a 90 yo that tested at 147 on the Iowa test in the 7th grade,
made it 2 months into freshman in high school, got bored & quit and went
to work fixing what was then the brand new things we call tv's now.

I'm sorry your experience has not been that rosy.  Mine on balance has
been quite good, and I thank the people that have made it what is is today.

Tech changes, usually for the better, so my machines are gradually being
converted from plain steppers to closed loop stepper/servo's, they can
be run on much higher voltages and the drivers control them MUCH faster
to higher accuracy's. One of my lathes is run by an rpi4b, runs so well
I wish it had been my first conversion, the other 3 running on wintel
stuff use 10x the ac power when idle.

Thank you.

I cannot remember if I posted about it here, but a buddy of mine and I
hacked a printer driver onto a BeagleBone Black, with some filtering
programs to change the dialect of g-code produced by SolidWorks, and hooked
the BBB into the tape port of a 1989 Cincinnati Milicron lathe.  It also
managed the tape's forward and reverse, so that it could page in blocks of
g-code.  It should like I should be able to do the same thing with my
Creality Falcon-2 laser cutter.  That will be a background project, but one
that would help save some time in the long run...

Sounds interesting. Keep us posted.

EBo --

On Sat, Jun 28, 2025 at 5:33 PM gene heskett via Discuss <
discuss@lists.openscad.org> wrote:

On 6/28/25 12:42, John David via Discuss wrote:

William, gcpdxf.py
https://github.com/WillAdams/gcodepreview/blob/main/gcpdxf.py looks
interesting.  I will need to take a closer look at it.  I have been
focusing on SVG because I have other tools that worked, but at the time I
could not get the DXF to work properly.  But, I do not inherently care

if I

use SVG or DXF for the format as long as the various interface option
(LaserGRBL, LightBurn, Inkscape, etc.) all play nice.  I will reach out

to

you off-list about both providing 2D DXF test CAD files. I will probably
start with a quick-release jig fixture for the Falcon-2 - it would be

nice

to have the lockable jig plates also have engraved rulers...

Cheers, Gene Heskett, CET.

"There are four boxes to be used in defense of liberty:
soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author, 1940)
If we desire respect for the law, we must first make the law respectable.

  • Louis D. Brandeis
On 6/28/25 19:18, John David wrote: > I had looked at klipper long ago, but never looked at it that deeply. > Something else to look into. Also, that dremel blade sharpener trick is a > nice deal. Thank you for sharing. > > Gene, you probably know/remember. I was an official developer for > LinuxCNC. I finally completely gave up on it due to the systematic > bit-rot. I remember one time I proposed a change (I forget what it was > specifically, this was circa 2003), but I do remember the screams of "DO > NOT TOUCH IT!" To be fair, a couple of the LinuxCNC members explained > off-list how it took them sometimes a year or more to get a machine > stabilized after a major update. Basically the entire community would get > something that works for small groups of the membership, and then they > would maintain that on their personal equipment -- until they *had* to > upgrade, or some feature was so desirable that it was worth spending the > time to upgrade. So, I ended up mostly hacking what I needed for my > personal projects and sending back minor bugs. I had great hopes with the > advent of MachineKit. But, over the pandemic, I ended up spending 100's of > hours and was never able to get the versions available at the time to even > fully run! That is when I completely gave up on LinuxCNC. Now if I > compare that with something like FluidNC, the code is clean, modern, and > has a more progressive support community. That said, the main code base > does not support backlash, and is something I would have to bring back in > from a previous branch. As we all know, with old iron, unmanaged backlash > is death. I would early love if LinuxCNC could be made viable, but my > personal experience has demonstrated repeatedly that it is not past taking > the time to get a single machine configuration working on very > specific hardware... I have 4 machines running linuxcnc, started back quite some time before the name change from emc2. I long ago learned to write my own gcode, and hal stuff, the only problem I ever had was something that should have been fixed about the time the axis gui was written. It did not properly change the focus of he teeny little axis buttons to correspond to register with the last axis moved, so touch-offs were often inadvertently applied to the wrong axis, making one start from scratch rehoming the machine. And it litterally took years, 3 or 4, from my reporting that bug to getting it fixed. Other than that, I have been able to make a machine do exactly as I wanted it to do.  In 20+ years, I haven't noticed any bit rot that wasn't fixable by correcting an error in my .hal files.  I also am not a machinist other than making parts to keep a television transmitter on the air for 50 years beyond its use by date, I am a CET, I can and have taught papered EE's stuff they should have learned in school, but the prof skipped it because he didn't understand it himself. Or maybe miss-understood it and taught it wrong.  I am amazed at the folks who think Newtons 3rd law is optional, or that relativity doesn't apply to the real world. Please understand also that LinuxCNC is a 100% volunteer effort by people who COULD be doing something generating a paycheck, most have a $dayjob that might not even be linuxcnc related but that is not how things started by NIST 50 or 60 years ago works. I have even offered hardware to expand things, been turned down, there is not I have been told, a mechanism in place that has a bank account. That means there is no one an MBA can sue, a limiting factor in its popularity because MBA's MUST have somebody to sue when their ideas wind up in the toilet. Its no secret that I consider MBA's in the same category as lawyers when Shakespear wrote "first we kill all the lawyers".  OTOH that is what you get from a 90 yo that tested at 147 on the Iowa test in the 7th grade, made it 2 months into freshman in high school, got bored & quit and went to work fixing what was then the brand new things we call tv's now. I'm sorry your experience has not been that rosy.  Mine on balance has been quite good, and I thank the people that have made it what is is today. Tech changes, usually for the better, so my machines are gradually being converted from plain steppers to closed loop stepper/servo's, they can be run on much higher voltages and the drivers control them MUCH faster to higher accuracy's. One of my lathes is run by an rpi4b, runs so well I wish it had been my first conversion, the other 3 running on wintel stuff use 10x the ac power when idle. Thank you. > I cannot remember if I posted about it here, but a buddy of mine and I > hacked a printer driver onto a BeagleBone Black, with some filtering > programs to change the dialect of g-code produced by SolidWorks, and hooked > the BBB into the tape port of a 1989 Cincinnati Milicron lathe. It also > managed the tape's forward and reverse, so that it could page in blocks of > g-code. It should like I should be able to do the same thing with my > Creality Falcon-2 laser cutter. That will be a background project, but one > that would help save some time in the long run... Sounds interesting. Keep us posted. > EBo -- > > On Sat, Jun 28, 2025 at 5:33 PM gene heskett via Discuss < > discuss@lists.openscad.org> wrote: > >> On 6/28/25 12:42, John David via Discuss wrote: >>> William, gcpdxf.py >>> <https://github.com/WillAdams/gcodepreview/blob/main/gcpdxf.py> looks >>> interesting. I will need to take a closer look at it. I have been >>> focusing on SVG because I have other tools that worked, but at the time I >>> could not get the DXF to work properly. But, I do not inherently care >> if I >>> use SVG or DXF for the format as long as the various interface option >>> (LaserGRBL, LightBurn, Inkscape, etc.) all play nice. I will reach out >> to >>> you off-list about both providing 2D DXF test CAD files. I will probably >>> start with a quick-release jig fixture for the Falcon-2 - it would be >> nice >>> to have the lockable jig plates also have engraved rulers... Cheers, Gene Heskett, CET. -- "There are four boxes to be used in defense of liberty: soap, ballot, jury, and ammo. Please use in that order." -Ed Howdershelt (Author, 1940) If we desire respect for the law, we must first make the law respectable. - Louis D. Brandeis
JD
John David
Sun, Jun 29, 2025 4:51 AM

re: LinuxCNC on an RPi4B...

What reversion of LinuxCNC did you use to get it to build and install on an
RPi4B?  As I recall, it was exactly my experience with LCNC on one of the
RPi4's at the beginning of the pandemic is exactly what drove me to such
distraction...  That said, that was 5 years ago.  Hopefully
something changed.  Let me know, and I will see if I can get it installed
on one of my RPi4's I have lying around.  That project as a whole is about
287 down the list, but I do know I have a couple of RPi4's sitting in a
box, and I have a rebuilt Techno Isel router with a blown controller, and a
South Bend Turret Mill that I picked up with a broken controller, and other
machines that need lot's of love...  Getting either of them up and running
would be more than useful.

re: small number of people maintaining LinuxCNC...

Back around 2003 Yodaiken actually hired me as a contractor to update then
EMC2 to run on their latest and greatest version of RTLinux - so that they
could give it away as demo's of all the coolness they were working on with
RTLinux.  The hassles I had with the community at that time over any
changes, clued me into what all was going on. I then learned that 80 to 90%
of all changes were being made by only 4 people, and of those 4, the most
prolific one spent a total of maybe 5hrs/week averaged over an entire year.
When I compared the work I had done for Yodaiken at that time, it was
roughly the equivalent of a full year of volunteer effort of those 4
people!  Yes, you are right that there are not many people to maintain it,
and I doubt there ever will be.  When people, like myself and others, come
along, we get pushed off or eventually get so tired we go find other fun
stuff to do. I would dearly love to be wrong and find that I can spend a
day getting LCNC running on an RPi4, and get the axes moving, even if it
takes me days more to dial in the config.  Back in 2020 I was not even able
to get any of the releases and release candidates running on that RPi.

re: High school experiences...

Sounds like you and I could swap a few stories about yesteryears. I took my
books home all of 4 times during my freshman year of high school (midterms
and finals). At the end of the year, I convinced the principle to grant me
a waver to apply to the early admittance program at the local college.  I
started a month later (paying for it myself with money I made selling my
own pottery throughout the region -- I grew up in a family of professional
potters, and started professionally slinging mud at my mother's knee at the
age of 7).  Now THAT is a stole down amnesia lane (thank you, Robin
Williams, from the Dead Poets Society).

Final note - sorry for coming across so LCNC Downer.  I have just seen so
many people lured in by the promise of production ready interfaces, only to
get turned off by how brittle the code is.  I would love you to prove me
wrong, and if you do, I will publically apologize for my misstep.

EBo --

On Sat, Jun 28, 2025 at 11:38 PM gene heskett gheskett@shentel.net wrote:

On 6/28/25 19:18, John David wrote:

I had looked at klipper long ago, but never looked at it that deeply.
Something else to look into.  Also, that dremel blade sharpener trick is

a

nice deal.  Thank you for sharing.

Gene, you probably know/remember.  I was an official developer for
LinuxCNC.  I finally completely gave up on it due to the systematic
bit-rot.  I remember one time I proposed a change (I forget what it was
specifically, this was circa 2003), but I do remember the screams of "DO
NOT TOUCH IT!"  To be fair, a couple of the LinuxCNC members explained
off-list how it took them sometimes a year or more to get a machine
stabilized after a major update.  Basically the entire community would

get

something that works for small groups of the membership, and then they
would maintain that on their personal equipment -- until they had to
upgrade, or some feature was so desirable that it was worth spending the
time to upgrade. So, I ended up mostly hacking what I needed for my
personal projects and sending back minor bugs.  I had great hopes with

the

advent of MachineKit.  But, over the pandemic, I ended up spending 100's

of

hours and was never able to get the versions available at the time to

even

fully run!  That is when I completely gave up on LinuxCNC.  Now if I
compare that with something like FluidNC, the code is clean, modern, and
has a more progressive support community.  That said, the main code base
does not support backlash, and is something I would have to bring back in
from a previous branch.  As we all know, with old iron, unmanaged

backlash

is death. I would early love if LinuxCNC could be made viable, but my
personal experience has demonstrated repeatedly that it is not past

taking

the time to get a single machine configuration working on very
specific hardware...

I have 4 machines running linuxcnc, started back quite some time before
the name change from emc2.

I long ago learned to write my own gcode, and hal stuff, the only
problem I ever had was something that should have been fixed about the
time the axis gui was written. It did not properly change the focus of
he teeny little axis buttons to correspond to register with the last
axis moved, so touch-offs were often inadvertently applied to the wrong
axis, making one start from scratch rehoming the machine. And it
litterally took years, 3 or 4, from my reporting that bug to getting it
fixed. Other than that, I have been able to make a machine do exactly as
I wanted it to do.  In 20+ years, I haven't noticed any bit rot that
wasn't fixable by correcting an error in my .hal files.  I also am not a
machinist other than making parts to keep a television transmitter on
the air for 50 years beyond its use by date, I am a CET, I can and have
taught papered EE's stuff they should have learned in school, but the
prof skipped it because he didn't understand it himself. Or maybe
miss-understood it and taught it wrong.  I am amazed at the folks who
think Newtons 3rd law is optional, or that relativity doesn't apply to
the real world.

Please understand also that LinuxCNC is a 100% volunteer effort by
people who COULD be doing something generating a paycheck, most have a
$dayjob that might not even be linuxcnc related but that is not how
things started by NIST 50 or 60 years ago works. I have even offered
hardware to expand things, been turned down, there is not I have been
told, a mechanism in place that has a bank account. That means there is
no one an MBA can sue, a limiting factor in its popularity because MBA's
MUST have somebody to sue when their ideas wind up in the toilet. Its no
secret that I consider MBA's in the same category as lawyers when
Shakespear wrote "first we kill all the lawyers".  OTOH that is what you
get from a 90 yo that tested at 147 on the Iowa test in the 7th grade,
made it 2 months into freshman in high school, got bored & quit and went
to work fixing what was then the brand new things we call tv's now.

I'm sorry your experience has not been that rosy.  Mine on balance has
been quite good, and I thank the people that have made it what is is today.

Tech changes, usually for the better, so my machines are gradually being
converted from plain steppers to closed loop stepper/servo's, they can
be run on much higher voltages and the drivers control them MUCH faster
to higher accuracy's. One of my lathes is run by an rpi4b, runs so well
I wish it had been my first conversion, the other 3 running on wintel
stuff use 10x the ac power when idle.

Thank you.

I cannot remember if I posted about it here, but a buddy of mine and I
hacked a printer driver onto a BeagleBone Black, with some filtering
programs to change the dialect of g-code produced by SolidWorks, and

hooked

the BBB into the tape port of a 1989 Cincinnati Milicron lathe.  It also
managed the tape's forward and reverse, so that it could page in blocks

of

g-code.  It should like I should be able to do the same thing with my
Creality Falcon-2 laser cutter.  That will be a background project, but

one

that would help save some time in the long run...

Sounds interesting. Keep us posted.

EBo --

On Sat, Jun 28, 2025 at 5:33 PM gene heskett via Discuss <
discuss@lists.openscad.org> wrote:

On 6/28/25 12:42, John David via Discuss wrote:

William, gcpdxf.py
https://github.com/WillAdams/gcodepreview/blob/main/gcpdxf.py looks
interesting.  I will need to take a closer look at it.  I have been
focusing on SVG because I have other tools that worked, but at the

time I

could not get the DXF to work properly.  But, I do not inherently care

if I

use SVG or DXF for the format as long as the various interface option
(LaserGRBL, LightBurn, Inkscape, etc.) all play nice.  I will reach out

to

you off-list about both providing 2D DXF test CAD files. I will

probably

start with a quick-release jig fixture for the Falcon-2 - it would be

nice

to have the lockable jig plates also have engraved rulers...

Cheers, Gene Heskett, CET.

"There are four boxes to be used in defense of liberty:
soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author, 1940)
If we desire respect for the law, we must first make the law respectable.

  • Louis D. Brandeis
re: LinuxCNC on an RPi4B... What reversion of LinuxCNC did you use to get it to build and install on an RPi4B? As I recall, it was exactly my experience with LCNC on one of the RPi4's at the beginning of the pandemic is exactly what drove me to such distraction... That said, that was 5 years ago. Hopefully something changed. Let me know, and I will see if I can get it installed on one of my RPi4's I have lying around. That project as a whole is about 287 down the list, but I do know I have a couple of RPi4's sitting in a box, and I have a rebuilt Techno Isel router with a blown controller, and a South Bend Turret Mill that I picked up with a broken controller, and other machines that need lot's of love... Getting either of them up and running would be more than useful. re: small number of people maintaining LinuxCNC... Back around 2003 Yodaiken actually hired me as a contractor to update then EMC2 to run on their latest and greatest version of RTLinux - so that they could give it away as demo's of all the coolness they were working on with RTLinux. The hassles I had with the community at that time over any changes, clued me into what all was going on. I then learned that 80 to 90% of *all* changes were being made by only 4 people, and of those 4, the most prolific one spent a total of maybe 5hrs/week averaged over an entire year. When I compared the work I had done for Yodaiken at that time, it was roughly the equivalent of a full year of volunteer effort of those 4 people! Yes, you are right that there are not many people to maintain it, and I doubt there ever will be. When people, like myself and others, come along, we get pushed off or eventually get so tired we go find other fun stuff to do. I would dearly love to be wrong and find that I can spend a day getting LCNC running on an RPi4, and get the axes moving, even if it takes me days more to dial in the config. Back in 2020 I was not even able to get any of the releases and release candidates running on that RPi. re: High school experiences... Sounds like you and I could swap a few stories about yesteryears. I took my books home all of 4 times during my freshman year of high school (midterms and finals). At the end of the year, I convinced the principle to grant me a waver to apply to the early admittance program at the local college. I started a month later (paying for it myself with money I made selling my own pottery throughout the region -- I grew up in a family of professional potters, and started professionally slinging mud at my mother's knee at the age of 7). Now THAT is a stole down amnesia lane (thank you, Robin Williams, from the Dead Poets Society). Final note - sorry for coming across so LCNC Downer. I have just seen so many people lured in by the promise of production ready interfaces, only to get turned off by how brittle the code is. I would love you to prove me wrong, and if you do, I will publically apologize for my misstep. EBo -- On Sat, Jun 28, 2025 at 11:38 PM gene heskett <gheskett@shentel.net> wrote: > On 6/28/25 19:18, John David wrote: > > I had looked at klipper long ago, but never looked at it that deeply. > > Something else to look into. Also, that dremel blade sharpener trick is > a > > nice deal. Thank you for sharing. > > > > Gene, you probably know/remember. I was an official developer for > > LinuxCNC. I finally completely gave up on it due to the systematic > > bit-rot. I remember one time I proposed a change (I forget what it was > > specifically, this was circa 2003), but I do remember the screams of "DO > > NOT TOUCH IT!" To be fair, a couple of the LinuxCNC members explained > > off-list how it took them sometimes a year or more to get a machine > > stabilized after a major update. Basically the entire community would > get > > something that works for small groups of the membership, and then they > > would maintain that on their personal equipment -- until they *had* to > > upgrade, or some feature was so desirable that it was worth spending the > > time to upgrade. So, I ended up mostly hacking what I needed for my > > personal projects and sending back minor bugs. I had great hopes with > the > > advent of MachineKit. But, over the pandemic, I ended up spending 100's > of > > hours and was never able to get the versions available at the time to > even > > fully run! That is when I completely gave up on LinuxCNC. Now if I > > compare that with something like FluidNC, the code is clean, modern, and > > has a more progressive support community. That said, the main code base > > does not support backlash, and is something I would have to bring back in > > from a previous branch. As we all know, with old iron, unmanaged > backlash > > is death. I would early love if LinuxCNC could be made viable, but my > > personal experience has demonstrated repeatedly that it is not past > taking > > the time to get a single machine configuration working on very > > specific hardware... > > I have 4 machines running linuxcnc, started back quite some time before > the name change from emc2. > > I long ago learned to write my own gcode, and hal stuff, the only > problem I ever had was something that should have been fixed about the > time the axis gui was written. It did not properly change the focus of > he teeny little axis buttons to correspond to register with the last > axis moved, so touch-offs were often inadvertently applied to the wrong > axis, making one start from scratch rehoming the machine. And it > litterally took years, 3 or 4, from my reporting that bug to getting it > fixed. Other than that, I have been able to make a machine do exactly as > I wanted it to do. In 20+ years, I haven't noticed any bit rot that > wasn't fixable by correcting an error in my .hal files. I also am not a > machinist other than making parts to keep a television transmitter on > the air for 50 years beyond its use by date, I am a CET, I can and have > taught papered EE's stuff they should have learned in school, but the > prof skipped it because he didn't understand it himself. Or maybe > miss-understood it and taught it wrong. I am amazed at the folks who > think Newtons 3rd law is optional, or that relativity doesn't apply to > the real world. > > Please understand also that LinuxCNC is a 100% volunteer effort by > people who COULD be doing something generating a paycheck, most have a > $dayjob that might not even be linuxcnc related but that is not how > things started by NIST 50 or 60 years ago works. I have even offered > hardware to expand things, been turned down, there is not I have been > told, a mechanism in place that has a bank account. That means there is > no one an MBA can sue, a limiting factor in its popularity because MBA's > MUST have somebody to sue when their ideas wind up in the toilet. Its no > secret that I consider MBA's in the same category as lawyers when > Shakespear wrote "first we kill all the lawyers". OTOH that is what you > get from a 90 yo that tested at 147 on the Iowa test in the 7th grade, > made it 2 months into freshman in high school, got bored & quit and went > to work fixing what was then the brand new things we call tv's now. > > I'm sorry your experience has not been that rosy. Mine on balance has > been quite good, and I thank the people that have made it what is is today. > > Tech changes, usually for the better, so my machines are gradually being > converted from plain steppers to closed loop stepper/servo's, they can > be run on much higher voltages and the drivers control them MUCH faster > to higher accuracy's. One of my lathes is run by an rpi4b, runs so well > I wish it had been my first conversion, the other 3 running on wintel > stuff use 10x the ac power when idle. > > Thank you. > > > I cannot remember if I posted about it here, but a buddy of mine and I > > hacked a printer driver onto a BeagleBone Black, with some filtering > > programs to change the dialect of g-code produced by SolidWorks, and > hooked > > the BBB into the tape port of a 1989 Cincinnati Milicron lathe. It also > > managed the tape's forward and reverse, so that it could page in blocks > of > > g-code. It should like I should be able to do the same thing with my > > Creality Falcon-2 laser cutter. That will be a background project, but > one > > that would help save some time in the long run... > Sounds interesting. Keep us posted. > > EBo -- > > > > On Sat, Jun 28, 2025 at 5:33 PM gene heskett via Discuss < > > discuss@lists.openscad.org> wrote: > > > >> On 6/28/25 12:42, John David via Discuss wrote: > >>> William, gcpdxf.py > >>> <https://github.com/WillAdams/gcodepreview/blob/main/gcpdxf.py> looks > >>> interesting. I will need to take a closer look at it. I have been > >>> focusing on SVG because I have other tools that worked, but at the > time I > >>> could not get the DXF to work properly. But, I do not inherently care > >> if I > >>> use SVG or DXF for the format as long as the various interface option > >>> (LaserGRBL, LightBurn, Inkscape, etc.) all play nice. I will reach out > >> to > >>> you off-list about both providing 2D DXF test CAD files. I will > probably > >>> start with a quick-release jig fixture for the Falcon-2 - it would be > >> nice > >>> to have the lockable jig plates also have engraved rulers... > > Cheers, Gene Heskett, CET. > -- > "There are four boxes to be used in defense of liberty: > soap, ballot, jury, and ammo. Please use in that order." > -Ed Howdershelt (Author, 1940) > If we desire respect for the law, we must first make the law respectable. > - Louis D. Brandeis > >
BC
Bob Carlson
Sun, Jun 29, 2025 5:48 AM

On Jun 28, 2025, at 20:38, gene heskett via Discuss discuss@lists.openscad.org wrote:

Its no secret that I consider MBA's in the same category as lawyers when Shakespear wrote "first we kill all the lawyers”.

Like almost everyone, you misunderstand that quote. It’s the bad guys in the play that say that. Without honest lawyers we are lost.

-Bob
Tucson AZ

> On Jun 28, 2025, at 20:38, gene heskett via Discuss <discuss@lists.openscad.org> wrote: > > Its no secret that I consider MBA's in the same category as lawyers when Shakespear wrote "first we kill all the lawyers”. Like almost everyone, you misunderstand that quote. It’s the bad guys in the play that say that. Without honest lawyers we are lost. -Bob Tucson AZ
GH
gene heskett
Sun, Jun 29, 2025 7:54 AM

On 6/29/25 01:49, Bob Carlson wrote:

On Jun 28, 2025, at 20:38, gene heskett via Discuss discuss@lists.openscad.org wrote:

Its no secret that I consider MBA's in the same category as lawyers when Shakespear wrote "first we kill all the lawyers”.

Like almost everyone, you misunderstand that quote. It’s the bad guys in the play that say that. Without honest lawyers we are lost.

Key word being HONEST.  Generally most, say 90% are.  I've met both
sides of that in my 90 years.  MBA's OTOH will offer $18k/yr for a tv
Chief Engineer worth $90k/yr and think they are doing you a huge favor. 
I don't work for people like that, just one of the reasons I turned that
chair at KTLA-tv, the #1 station in the country, down 30+ years ago when
a head hunter tried to poach me.. That $110k would have been a good
offer if it wasn't in LA.  WV was/is good to me.  Among other things it
gave me my 3rd wife for 31 years.  Never had a fight.  Passed from COPD
in late 2020. Real estate free and clear a/o 26 years ago. Now I'm just
a diabetic and alone old fart, the end in distant sight.

-Bob
Tucson AZ

Cheers, Gene Heskett, CET.

"There are four boxes to be used in defense of liberty:
soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author, 1940)
If we desire respect for the law, we must first make the law respectable.

  • Louis D. Brandeis
On 6/29/25 01:49, Bob Carlson wrote: >> On Jun 28, 2025, at 20:38, gene heskett via Discuss <discuss@lists.openscad.org> wrote: >> >> Its no secret that I consider MBA's in the same category as lawyers when Shakespear wrote "first we kill all the lawyers”. > Like almost everyone, you misunderstand that quote. It’s the bad guys in the play that say that. Without honest lawyers we are lost. Key word being HONEST.  Generally most, say 90% are.  I've met both sides of that in my 90 years.  MBA's OTOH will offer $18k/yr for a tv Chief Engineer worth $90k/yr and think they are doing you a huge favor.  I don't work for people like that, just one of the reasons I turned that chair at KTLA-tv, the #1 station in the country, down 30+ years ago when a head hunter tried to poach me.. That $110k would have been a good offer if it wasn't in LA.  WV was/is good to me.  Among other things it gave me my 3rd wife for 31 years.  Never had a fight.  Passed from COPD in late 2020. Real estate free and clear a/o 26 years ago. Now I'm just a diabetic and alone old fart, the end in distant sight. > > -Bob > Tucson AZ > > Cheers, Gene Heskett, CET. -- "There are four boxes to be used in defense of liberty: soap, ballot, jury, and ammo. Please use in that order." -Ed Howdershelt (Author, 1940) If we desire respect for the law, we must first make the law respectable. - Louis D. Brandeis
JD
John David
Sun, Jun 29, 2025 8:23 AM

Sorry to hear about your loss and other changes.  Yea, I was recently
surprised to run across a skilled machinist with 10+ years of experiences,
and was offering a little over minimum wage.  Yea, some people and places
are not worth working for.

EBo --

On Sun, Jun 29, 2025 at 3:54 AM gene heskett gheskett@shentel.net wrote:

On 6/29/25 01:49, Bob Carlson wrote:

On Jun 28, 2025, at 20:38, gene heskett via Discuss <

Its no secret that I consider MBA's in the same category as lawyers

when Shakespear wrote "first we kill all the lawyers”.

Like almost everyone, you misunderstand that quote. It’s the bad guys in

the play that say that. Without honest lawyers we are lost.
Key word being HONEST.  Generally most, say 90% are.  I've met both
sides of that in my 90 years.  MBA's OTOH will offer $18k/yr for a tv
Chief Engineer worth $90k/yr and think they are doing you a huge favor.
I don't work for people like that, just one of the reasons I turned that
chair at KTLA-tv, the #1 station in the country, down 30+ years ago when
a head hunter tried to poach me.. That $110k would have been a good
offer if it wasn't in LA.  WV was/is good to me.  Among other things it
gave me my 3rd wife for 31 years.  Never had a fight.  Passed from COPD
in late 2020. Real estate free and clear a/o 26 years ago. Now I'm just
a diabetic and alone old fart, the end in distant sight.

-Bob
Tucson AZ

Cheers, Gene Heskett, CET.

"There are four boxes to be used in defense of liberty:
soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author, 1940)
If we desire respect for the law, we must first make the law respectable.

  • Louis D. Brandeis
Sorry to hear about your loss and other changes. Yea, I was recently surprised to run across a skilled machinist with 10+ years of experiences, and was offering a little over minimum wage. Yea, some people and places are not worth working for. EBo -- On Sun, Jun 29, 2025 at 3:54 AM gene heskett <gheskett@shentel.net> wrote: > On 6/29/25 01:49, Bob Carlson wrote: > >> On Jun 28, 2025, at 20:38, gene heskett via Discuss < > discuss@lists.openscad.org> wrote: > >> > >> Its no secret that I consider MBA's in the same category as lawyers > when Shakespear wrote "first we kill all the lawyers”. > > Like almost everyone, you misunderstand that quote. It’s the bad guys in > the play that say that. Without honest lawyers we are lost. > Key word being HONEST. Generally most, say 90% are. I've met both > sides of that in my 90 years. MBA's OTOH will offer $18k/yr for a tv > Chief Engineer worth $90k/yr and think they are doing you a huge favor. > I don't work for people like that, just one of the reasons I turned that > chair at KTLA-tv, the #1 station in the country, down 30+ years ago when > a head hunter tried to poach me.. That $110k would have been a good > offer if it wasn't in LA. WV was/is good to me. Among other things it > gave me my 3rd wife for 31 years. Never had a fight. Passed from COPD > in late 2020. Real estate free and clear a/o 26 years ago. Now I'm just > a diabetic and alone old fart, the end in distant sight. > > > > -Bob > > Tucson AZ > > > > > Cheers, Gene Heskett, CET. > -- > "There are four boxes to be used in defense of liberty: > soap, ballot, jury, and ammo. Please use in that order." > -Ed Howdershelt (Author, 1940) > If we desire respect for the law, we must first make the law respectable. > - Louis D. Brandeis > >
WF
William F. Adams
Sun, Jun 29, 2025 1:43 PM

Okay, I added a command for changing colour:

setcolour("Gray")

(see attached screen grab)

One interesting complication/consideration is that there is a limited intersection in named colours betwixt OpenSCAD's support of the named Web Colors (see article on Wikipedia) and the basic numbered colours in the DXF format:

\item "Black" (0)
\item "Red" (1)
\item "Yellow" (2)
\item "Green" (3)
\item "Cyan" (4)
\item "Blue" (5)
\item "Magenta" (6)
\item "White" (7)
\item "Dark Gray" (8)
\item "Light Gray" (9)

with the common options being:

\item Black
\item Red
\item Yellow
\item Green
\item Blue
\item White
\item (Dark) Gray

Black and White aren't that useful/are potentially confusing since at least one program toggles them so as to make Dark Mode work, so we are down to red, yellow, green, blue, and (dark) gray --- can those be made to work for this?

William

Okay, I added a command for changing colour: setcolour("Gray") (see attached screen grab) One interesting complication/consideration is that there is a limited intersection in named colours betwixt OpenSCAD's support of the named Web Colors (see article on Wikipedia) and the basic numbered colours in the DXF format: \item "Black" (0) \item "Red" (1) \item "Yellow" (2) \item "Green" (3) \item "Cyan" (4) \item "Blue" (5) \item "Magenta" (6) \item "White" (7) \item "Dark Gray" (8) \item "Light Gray" (9) with the common options being: \item Black \item Red \item Yellow \item Green \item Blue \item White \item (Dark) Gray Black and White aren't that useful/are potentially confusing since at least one program toggles them so as to make Dark Mode work, so we are down to red, yellow, green, blue, and (dark) gray --- can those be made to work for this? William
GH
gene heskett
Sun, Jun 29, 2025 2:46 PM

On 6/29/25 04:23, John David wrote:

Sorry to hear about your loss and other changes.  Yea, I was recently
surprised to run across a skilled machinist with 10+ years of experiences,
and was offering a little over minimum wage.  Yea, some people and places
are not worth working for.

10 yrs of experience does not always make a "machinist"...

I was, in the later 50's, involved with the building of the tv cameras
that were on the Navy's Trieste when it made that one trip to the bottom
of the mohole, the deepest place in the pacific, nearly 38,000 feet
deep. We needed pressure cases for the cameras we were making, so
Oceanographic Engineering bought a huge LeBlonde lathe and hired a
machinist.  Since the pressure at that depth, about 18,000 psia, is
great enough to compress water, he followed the navy's math, and carved
two camera cases out of bronze, starting with two 12" diameter and 22"
long billets of bronze. The cameras were 2.5" in diameter and around 13"
long w/o lenses. The navy supplied the solid quartz windows the cameras
looked thru. He knew how much that bronze would compress under that
pressure, doubled it and put elastic pads around the cameras pcb's, same
for the quartz windows.  And designed a bnc connector for the i/o
connection that could withstand that pressure w/o taking on any
seawater. Watching him work was amazing.  Him, I would call a
machinist.  FWIW the cameras worked flawlessly although one of the pan &
tilt assemblies came back up full of water.  The bnc connector was what
we started with but transitioned to a single strand of 16 ga packard
automotive wire, no coax, as we had only 6 conductors into the 6' cast
iron gondola on the bottom of the Trieste.  One of the company founders
was good with relay logic and designed a system that let us use
sequential push buttons to control the lights and both pan & tilt units
with the remaining 4 wires.

Pure serendipity has put me in other places to make what is now
forgotten history. I've had what one might call a good ride thru an
interesting time in human technological history.

Cheers, Gene Heskett, CET.

--
"There are four boxes to be used in defense of liberty:
soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author, 1940)
If we desire respect for the law, we must first make the law respectable.

  • Louis D. Brandeis
On 6/29/25 04:23, John David wrote: > Sorry to hear about your loss and other changes. Yea, I was recently > surprised to run across a skilled machinist with 10+ years of experiences, > and was offering a little over minimum wage. Yea, some people and places > are not worth working for. 10 yrs of experience does not always make a "machinist"... I was, in the later 50's, involved with the building of the tv cameras that were on the Navy's Trieste when it made that one trip to the bottom of the mohole, the deepest place in the pacific, nearly 38,000 feet deep. We needed pressure cases for the cameras we were making, so Oceanographic Engineering bought a huge LeBlonde lathe and hired a machinist.  Since the pressure at that depth, about 18,000 psia, is great enough to compress water, he followed the navy's math, and carved two camera cases out of bronze, starting with two 12" diameter and 22" long billets of bronze. The cameras were 2.5" in diameter and around 13" long w/o lenses. The navy supplied the solid quartz windows the cameras looked thru. He knew how much that bronze would compress under that pressure, doubled it and put elastic pads around the cameras pcb's, same for the quartz windows.  And designed a bnc connector for the i/o connection that could withstand that pressure w/o taking on any seawater. Watching him work was amazing.  Him, I would call a machinist.  FWIW the cameras worked flawlessly although one of the pan & tilt assemblies came back up full of water.  The bnc connector was what we started with but transitioned to a single strand of 16 ga packard automotive wire, no coax, as we had only 6 conductors into the 6' cast iron gondola on the bottom of the Trieste.  One of the company founders was good with relay logic and designed a system that let us use sequential push buttons to control the lights and both pan & tilt units with the remaining 4 wires. Pure serendipity has put me in other places to make what is now forgotten history. I've had what one might call a good ride thru an interesting time in human technological history. Cheers, Gene Heskett, CET. -- "There are four boxes to be used in defense of liberty: soap, ballot, jury, and ammo. Please use in that order." -Ed Howdershelt (Author, 1940) If we desire respect for the law, we must first make the law respectable. - Louis D. Brandeis
GH
gene heskett
Sun, Jun 29, 2025 3:31 PM

On 6/29/25 00:52, John David wrote:

re: LinuxCNC on an RPi4B...

What reversion of LinuxCNC did you use to get it to build and install on an
RPi4B?  As I recall, it was exactly my experience with LCNC on one of the
RPi4's at the beginning of the pandemic is exactly what drove me to such
distraction...  That said, that was 5 years ago.  Hopefully
something changed.  Let me know, and I will see if I can get it installed
on one of my RPi4's I have lying around.

I started out with an rpi3b, with whatever was b4 stretch, but running a
realtime preempt kernel I built. zero help from the foundation as the
3rd time I asked their forum about a realtime kernel I was given a link
to a 4.19 src, and banned forever. So I did get it to compile, then
figured out how to make a 27 megabyte tarball I could unpack to the u-sd
card that installed it. Then I hacked up the hal file quite a bit and
made linuxcnc run with an occasional stutter because even with all the
tricks, the rpi3b was a little too slow. So the rpi3b got replaced with
an rpi4b when the 4b came out. For about the last year its been running
flawlessly on bookworm built for me by Rod Webster. The current
preempt-rt kernel has a poorer latency than mine, but is still adequate.
I replaced the original steppers with closed loop stepper/servo's which
improved the finish while doubling the rapids. So it currently claims
its running debian 12.  Rod may still have that image.  My setup might
not be available from mesa today but if interested I can tarball my
config, but be aware there's lots of candy in my version that isn't std
LinuxCNC.

That project as a whole is about

287 down the list, but I do know I have a couple of RPi4's sitting in a
box, and I have a rebuilt Techno Isel router with a blown controller, and a
South Bend Turret Mill that I picked up with a broken controller, and other
machines that need lot's of love...  Getting either of them up and running
would be more than useful.

re: small number of people maintaining LinuxCNC...

Back around 2003 Yodaiken actually hired me as a contractor to update then
EMC2 to run on their latest and greatest version of RTLinux - so that they
could give it away as demo's of all the coolness they were working on with
RTLinux.  The hassles I had with the community at that time over any
changes, clued me into what all was going on. I then learned that 80 to 90%
of all changes were being made by only 4 people, and of those 4, the most
prolific one spent a total of maybe 5hrs/week averaged over an entire year.
When I compared the work I had done for Yodaiken at that time, it was
roughly the equivalent of a full year of volunteer effort of those 4
people!  Yes, you are right that there are not many people to maintain it,
and I doubt there ever will be.  When people, like myself and others, come
along, we get pushed off or eventually get so tired we go find other fun
stuff to do. I would dearly love to be wrong and find that I can spend a
day getting LCNC running on an RPi4, and get the axes moving, even if it
takes me days more to dial in the config.  Back in 2020 I was not even able
to get any of the releases and release candidates running on that RPi.

re: High school experiences...

Sounds like you and I could swap a few stories about yesteryears. I took my
books home all of 4 times during my freshman year of high school (midterms
and finals). At the end of the year, I convinced the principle to grant me
a waver to apply to the early admittance program at the local college.  I
started a month later (paying for it myself with money I made selling my
own pottery throughout the region -- I grew up in a family of professional
potters, and started professionally slinging mud at my mother's knee at the
age of 7).  Now THAT is a stole down amnesia lane (thank you, Robin
Williams, from the Dead Poets Society).

Yeah, we could trade history's and bore the crowd but much of mine isn't
LinuxCNC related.

Final note - sorry for coming across so LCNC Downer.  I have just seen so
many people lured in by the promise of production ready interfaces, only to
get turned off by how brittle the code is.  I would love you to prove me
wrong, and if you do, I will publically apologize for my misstep.

EBo --

On Sat, Jun 28, 2025 at 11:38 PM gene heskett gheskett@shentel.net wrote:

On 6/28/25 19:18, John David wrote:

I had looked at klipper long ago, but never looked at it that deeply.
Something else to look into.  Also, that dremel blade sharpener trick is

a

nice deal.  Thank you for sharing.

Gene, you probably know/remember.  I was an official developer for
LinuxCNC.  I finally completely gave up on it due to the systematic
bit-rot.  I remember one time I proposed a change (I forget what it was
specifically, this was circa 2003), but I do remember the screams of "DO
NOT TOUCH IT!"  To be fair, a couple of the LinuxCNC members explained
off-list how it took them sometimes a year or more to get a machine
stabilized after a major update.  Basically the entire community would

get

something that works for small groups of the membership, and then they
would maintain that on their personal equipment -- until they had to
upgrade, or some feature was so desirable that it was worth spending the
time to upgrade. So, I ended up mostly hacking what I needed for my
personal projects and sending back minor bugs.  I had great hopes with

the

advent of MachineKit.  But, over the pandemic, I ended up spending 100's

of

hours and was never able to get the versions available at the time to

even

fully run!  That is when I completely gave up on LinuxCNC.  Now if I
compare that with something like FluidNC, the code is clean, modern, and
has a more progressive support community.  That said, the main code base
does not support backlash, and is something I would have to bring back in
from a previous branch.  As we all know, with old iron, unmanaged

backlash

is death. I would early love if LinuxCNC could be made viable, but my
personal experience has demonstrated repeatedly that it is not past

taking

the time to get a single machine configuration working on very
specific hardware...

I have 4 machines running linuxcnc, started back quite some time before
the name change from emc2.

I long ago learned to write my own gcode, and hal stuff, the only
problem I ever had was something that should have been fixed about the
time the axis gui was written. It did not properly change the focus of
he teeny little axis buttons to correspond to register with the last
axis moved, so touch-offs were often inadvertently applied to the wrong
axis, making one start from scratch rehoming the machine. And it
litterally took years, 3 or 4, from my reporting that bug to getting it
fixed. Other than that, I have been able to make a machine do exactly as
I wanted it to do.  In 20+ years, I haven't noticed any bit rot that
wasn't fixable by correcting an error in my .hal files.  I also am not a
machinist other than making parts to keep a television transmitter on
the air for 50 years beyond its use by date, I am a CET, I can and have
taught papered EE's stuff they should have learned in school, but the
prof skipped it because he didn't understand it himself. Or maybe
miss-understood it and taught it wrong.  I am amazed at the folks who
think Newtons 3rd law is optional, or that relativity doesn't apply to
the real world.

Please understand also that LinuxCNC is a 100% volunteer effort by
people who COULD be doing something generating a paycheck, most have a
$dayjob that might not even be linuxcnc related but that is not how
things started by NIST 50 or 60 years ago works. I have even offered
hardware to expand things, been turned down, there is not I have been
told, a mechanism in place that has a bank account. That means there is
no one an MBA can sue, a limiting factor in its popularity because MBA's
MUST have somebody to sue when their ideas wind up in the toilet. Its no
secret that I consider MBA's in the same category as lawyers when
Shakespear wrote "first we kill all the lawyers".  OTOH that is what you
get from a 90 yo that tested at 147 on the Iowa test in the 7th grade,
made it 2 months into freshman in high school, got bored & quit and went
to work fixing what was then the brand new things we call tv's now.

I'm sorry your experience has not been that rosy.  Mine on balance has
been quite good, and I thank the people that have made it what is is today.

Tech changes, usually for the better, so my machines are gradually being
converted from plain steppers to closed loop stepper/servo's, they can
be run on much higher voltages and the drivers control them MUCH faster
to higher accuracy's. One of my lathes is run by an rpi4b, runs so well
I wish it had been my first conversion, the other 3 running on wintel
stuff use 10x the ac power when idle.

Thank you.

I cannot remember if I posted about it here, but a buddy of mine and I
hacked a printer driver onto a BeagleBone Black, with some filtering
programs to change the dialect of g-code produced by SolidWorks, and

hooked

the BBB into the tape port of a 1989 Cincinnati Milicron lathe.  It also
managed the tape's forward and reverse, so that it could page in blocks

of

g-code.  It should like I should be able to do the same thing with my
Creality Falcon-2 laser cutter.  That will be a background project, but

one

that would help save some time in the long run...

Sounds interesting. Keep us posted.

 EBo --

On Sat, Jun 28, 2025 at 5:33 PM gene heskett via Discuss <
discuss@lists.openscad.org> wrote:

On 6/28/25 12:42, John David via Discuss wrote:

William, gcpdxf.py
https://github.com/WillAdams/gcodepreview/blob/main/gcpdxf.py looks
interesting.  I will need to take a closer look at it.  I have been
focusing on SVG because I have other tools that worked, but at the

time I

could not get the DXF to work properly.  But, I do not inherently care

if I

use SVG or DXF for the format as long as the various interface option
(LaserGRBL, LightBurn, Inkscape, etc.) all play nice.  I will reach out

to

you off-list about both providing 2D DXF test CAD files. I will

probably

start with a quick-release jig fixture for the Falcon-2 - it would be

nice

to have the lockable jig plates also have engraved rulers...

Cheers, Gene Heskett, CET.

"There are four boxes to be used in defense of liberty:
soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author, 1940)
If we desire respect for the law, we must first make the law respectable.
- Louis D. Brandeis

Cheers, Gene Heskett, CET.

"There are four boxes to be used in defense of liberty:
soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author, 1940)
If we desire respect for the law, we must first make the law respectable.

  • Louis D. Brandeis
On 6/29/25 00:52, John David wrote: > re: LinuxCNC on an RPi4B... > > What reversion of LinuxCNC did you use to get it to build and install on an > RPi4B? As I recall, it was exactly my experience with LCNC on one of the > RPi4's at the beginning of the pandemic is exactly what drove me to such > distraction... That said, that was 5 years ago. Hopefully > something changed. Let me know, and I will see if I can get it installed > on one of my RPi4's I have lying around. I started out with an rpi3b, with whatever was b4 stretch, but running a realtime preempt kernel I built. zero help from the foundation as the 3rd time I asked their forum about a realtime kernel I was given a link to a 4.19 src, and banned forever. So I did get it to compile, then figured out how to make a 27 megabyte tarball I could unpack to the u-sd card that installed it. Then I hacked up the hal file quite a bit and made linuxcnc run with an occasional stutter because even with all the tricks, the rpi3b was a little too slow. So the rpi3b got replaced with an rpi4b when the 4b came out. For about the last year its been running flawlessly on bookworm built for me by Rod Webster. The current preempt-rt kernel has a poorer latency than mine, but is still adequate. I replaced the original steppers with closed loop stepper/servo's which improved the finish while doubling the rapids. So it currently claims its running debian 12.  Rod may still have that image.  My setup might not be available from mesa today but if interested I can tarball my config, but be aware there's lots of candy in my version that isn't std LinuxCNC. > That project as a whole is about > 287 down the list, but I do know I have a couple of RPi4's sitting in a > box, and I have a rebuilt Techno Isel router with a blown controller, and a > South Bend Turret Mill that I picked up with a broken controller, and other > machines that need lot's of love... Getting either of them up and running > would be more than useful. > > re: small number of people maintaining LinuxCNC... > > Back around 2003 Yodaiken actually hired me as a contractor to update then > EMC2 to run on their latest and greatest version of RTLinux - so that they > could give it away as demo's of all the coolness they were working on with > RTLinux. The hassles I had with the community at that time over any > changes, clued me into what all was going on. I then learned that 80 to 90% > of *all* changes were being made by only 4 people, and of those 4, the most > prolific one spent a total of maybe 5hrs/week averaged over an entire year. > When I compared the work I had done for Yodaiken at that time, it was > roughly the equivalent of a full year of volunteer effort of those 4 > people! Yes, you are right that there are not many people to maintain it, > and I doubt there ever will be. When people, like myself and others, come > along, we get pushed off or eventually get so tired we go find other fun > stuff to do. I would dearly love to be wrong and find that I can spend a > day getting LCNC running on an RPi4, and get the axes moving, even if it > takes me days more to dial in the config. Back in 2020 I was not even able > to get any of the releases and release candidates running on that RPi. > > re: High school experiences... > > Sounds like you and I could swap a few stories about yesteryears. I took my > books home all of 4 times during my freshman year of high school (midterms > and finals). At the end of the year, I convinced the principle to grant me > a waver to apply to the early admittance program at the local college. I > started a month later (paying for it myself with money I made selling my > own pottery throughout the region -- I grew up in a family of professional > potters, and started professionally slinging mud at my mother's knee at the > age of 7). Now THAT is a stole down amnesia lane (thank you, Robin > Williams, from the Dead Poets Society). Yeah, we could trade history's and bore the crowd but much of mine isn't LinuxCNC related. > > Final note - sorry for coming across so LCNC Downer. I have just seen so > many people lured in by the promise of production ready interfaces, only to > get turned off by how brittle the code is. I would love you to prove me > wrong, and if you do, I will publically apologize for my misstep. > > EBo -- > > On Sat, Jun 28, 2025 at 11:38 PM gene heskett <gheskett@shentel.net> wrote: > >> On 6/28/25 19:18, John David wrote: >>> I had looked at klipper long ago, but never looked at it that deeply. >>> Something else to look into. Also, that dremel blade sharpener trick is >> a >>> nice deal. Thank you for sharing. >>> >>> Gene, you probably know/remember. I was an official developer for >>> LinuxCNC. I finally completely gave up on it due to the systematic >>> bit-rot. I remember one time I proposed a change (I forget what it was >>> specifically, this was circa 2003), but I do remember the screams of "DO >>> NOT TOUCH IT!" To be fair, a couple of the LinuxCNC members explained >>> off-list how it took them sometimes a year or more to get a machine >>> stabilized after a major update. Basically the entire community would >> get >>> something that works for small groups of the membership, and then they >>> would maintain that on their personal equipment -- until they *had* to >>> upgrade, or some feature was so desirable that it was worth spending the >>> time to upgrade. So, I ended up mostly hacking what I needed for my >>> personal projects and sending back minor bugs. I had great hopes with >> the >>> advent of MachineKit. But, over the pandemic, I ended up spending 100's >> of >>> hours and was never able to get the versions available at the time to >> even >>> fully run! That is when I completely gave up on LinuxCNC. Now if I >>> compare that with something like FluidNC, the code is clean, modern, and >>> has a more progressive support community. That said, the main code base >>> does not support backlash, and is something I would have to bring back in >>> from a previous branch. As we all know, with old iron, unmanaged >> backlash >>> is death. I would early love if LinuxCNC could be made viable, but my >>> personal experience has demonstrated repeatedly that it is not past >> taking >>> the time to get a single machine configuration working on very >>> specific hardware... >> I have 4 machines running linuxcnc, started back quite some time before >> the name change from emc2. >> >> I long ago learned to write my own gcode, and hal stuff, the only >> problem I ever had was something that should have been fixed about the >> time the axis gui was written. It did not properly change the focus of >> he teeny little axis buttons to correspond to register with the last >> axis moved, so touch-offs were often inadvertently applied to the wrong >> axis, making one start from scratch rehoming the machine. And it >> litterally took years, 3 or 4, from my reporting that bug to getting it >> fixed. Other than that, I have been able to make a machine do exactly as >> I wanted it to do. In 20+ years, I haven't noticed any bit rot that >> wasn't fixable by correcting an error in my .hal files. I also am not a >> machinist other than making parts to keep a television transmitter on >> the air for 50 years beyond its use by date, I am a CET, I can and have >> taught papered EE's stuff they should have learned in school, but the >> prof skipped it because he didn't understand it himself. Or maybe >> miss-understood it and taught it wrong. I am amazed at the folks who >> think Newtons 3rd law is optional, or that relativity doesn't apply to >> the real world. >> >> Please understand also that LinuxCNC is a 100% volunteer effort by >> people who COULD be doing something generating a paycheck, most have a >> $dayjob that might not even be linuxcnc related but that is not how >> things started by NIST 50 or 60 years ago works. I have even offered >> hardware to expand things, been turned down, there is not I have been >> told, a mechanism in place that has a bank account. That means there is >> no one an MBA can sue, a limiting factor in its popularity because MBA's >> MUST have somebody to sue when their ideas wind up in the toilet. Its no >> secret that I consider MBA's in the same category as lawyers when >> Shakespear wrote "first we kill all the lawyers". OTOH that is what you >> get from a 90 yo that tested at 147 on the Iowa test in the 7th grade, >> made it 2 months into freshman in high school, got bored & quit and went >> to work fixing what was then the brand new things we call tv's now. >> >> I'm sorry your experience has not been that rosy. Mine on balance has >> been quite good, and I thank the people that have made it what is is today. >> >> Tech changes, usually for the better, so my machines are gradually being >> converted from plain steppers to closed loop stepper/servo's, they can >> be run on much higher voltages and the drivers control them MUCH faster >> to higher accuracy's. One of my lathes is run by an rpi4b, runs so well >> I wish it had been my first conversion, the other 3 running on wintel >> stuff use 10x the ac power when idle. >> >> Thank you. >> >>> I cannot remember if I posted about it here, but a buddy of mine and I >>> hacked a printer driver onto a BeagleBone Black, with some filtering >>> programs to change the dialect of g-code produced by SolidWorks, and >> hooked >>> the BBB into the tape port of a 1989 Cincinnati Milicron lathe. It also >>> managed the tape's forward and reverse, so that it could page in blocks >> of >>> g-code. It should like I should be able to do the same thing with my >>> Creality Falcon-2 laser cutter. That will be a background project, but >> one >>> that would help save some time in the long run... >> Sounds interesting. Keep us posted. >>> EBo -- >>> >>> On Sat, Jun 28, 2025 at 5:33 PM gene heskett via Discuss < >>> discuss@lists.openscad.org> wrote: >>> >>>> On 6/28/25 12:42, John David via Discuss wrote: >>>>> William, gcpdxf.py >>>>> <https://github.com/WillAdams/gcodepreview/blob/main/gcpdxf.py> looks >>>>> interesting. I will need to take a closer look at it. I have been >>>>> focusing on SVG because I have other tools that worked, but at the >> time I >>>>> could not get the DXF to work properly. But, I do not inherently care >>>> if I >>>>> use SVG or DXF for the format as long as the various interface option >>>>> (LaserGRBL, LightBurn, Inkscape, etc.) all play nice. I will reach out >>>> to >>>>> you off-list about both providing 2D DXF test CAD files. I will >> probably >>>>> start with a quick-release jig fixture for the Falcon-2 - it would be >>>> nice >>>>> to have the lockable jig plates also have engraved rulers... >> Cheers, Gene Heskett, CET. >> -- >> "There are four boxes to be used in defense of liberty: >> soap, ballot, jury, and ammo. Please use in that order." >> -Ed Howdershelt (Author, 1940) >> If we desire respect for the law, we must first make the law respectable. >> - Louis D. Brandeis >> >> Cheers, Gene Heskett, CET. -- "There are four boxes to be used in defense of liberty: soap, ballot, jury, and ammo. Please use in that order." -Ed Howdershelt (Author, 1940) If we desire respect for the law, we must first make the law respectable. - Louis D. Brandeis