Hi Andy
On Montag, 13. April 2020, 16:55:17 CEST andy pugh wrote:
> On Mon, 13 Apr 2020 at 05:26, Reinhard
wrote:
> > May be not broken, but at least weird.
>
> Are you able to do a git bisect to find exactly where the behaviour changed?
Just to let you know:
I had to sort
On Donnerstag, 16. April 2020, 08:33:31 CEST Gene Heskett wrote:
> That's now fixed, and the need to purchase a rider because my place us
> turning into a jungle, plus the ever increasing needs of my bride of 30+
> years as she is close to dying from COPD. That and medical emergencies
> since I
On Donnerstag, 16. April 2020, 05:51:38 CEST Jared McLaughlin wrote:
> it'd be super interesting to see canned cycles for linuxcnc
Just happened to discover ...
Do you know about ./nc_files/ngcgui_lib/qpocket.ngc ?
Looks like it is already done.
Reinh
so may be the Heidenhain way is
easier to adopt.
Reinhard
___
Emc-developers mailing list
Emc-developers@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-developers
ned cycle?
Pocket are well known jobs, that occur quite often. Such canned cycle could
make lots of users happy.
... and in the canned cycle you could distinct between roughing and smoothing
and of cause go for efficient roughing.
Reinhard
___
Emc-d
s would be enuf challenge ;)
... by the way: recently I was testing linuxvariant of PlanetCNC. To me it
looks like a fork of lc - and they already solved several problems of lc
apparently.
Reinhard
___
Emc-developers mailing list
l properties would be useless.
But tests with G43.1 gimme the hope, that linuxcnc already takes the fake
values for path calculations.
I will test that, when I have a new dev-box up and running.
Reinhard
___
Emc-developers mailing list
Em
at's the source of most inconsistencies of linuxcnc.
Reinhard
___
Emc-developers mailing list
Emc-developers@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-developers
than the estimated use-time in job.
We use shift-working, so I can't be sure, that the colleague knows the
sideeffects of my programming. Therefor I'd like to use safe codings.
cheers Reinhard
___
Emc-developers mailing list
Emc-deve
On Dienstag, 14. April 2020, 19:07:32 CEST Chris Morley wrote:
> On 2020-04-14 9:10 a.m., andy pugh wrote:
> > On Tue, 14 Apr 2020 at 03:58, Reinhard
wrote:
> >>> Maybe G43.2 could be tweaked to take either a H number or axis codes.
> >>
> >> That would be
s from tool
table, but I never saw a tooltable from nml with comment fields.
Did you have comment fields supported that time?
Reinhard
___
Emc-developers mailing list
Emc-developers@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-developers
ar-file entries and calculations is not the right syntax for the usual
linuxcnc user.
And after all - there's no reason, why that should be a showstopper for 2.8
cheers Reinhard
___
Emc-developers mailing list
Emc-developers@lists.source
app. But after investigation I was sure, my app
did nothing but showing values from nml.
On Montag, 13. April 2020, 16:55:17 CEST andy pugh wrote:
> On Mon, 13 Apr 2020 at 05:26, Reinhard
wrote:
> > Behaviour may be ok from sight of motion, but it is not from user sight
> >
Hi Andy,
On Montag, 13. April 2020, 16:56:50 CEST andy pugh wrote:
> On Mon, 13 Apr 2020 at 07:20, Reinhard
wrote:
> > I understood documentation that way, that G43.1 adds an offset to existing
> > tool length
>
> No, that is G43.2
I beg your pardon, but I don't underst
Hi Andy,
On Montag, 13. April 2020, 16:55:17 CEST andy pugh wrote:
> Are you able to do a git bisect to find exactly where the behaviour changed?
currently no - never heard that before.
but gonna read about it.
Reinhard
___
Emc-developers mail
not interested in the exact tool offset value.
So what is the usecase of replacing tool offset values in absolute manner?
Reinhard
___
Emc-developers mailing list
Emc-developers@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc
was
I expected - but in that context, things seem to work.
Except with the tool table, which does not show the z-values any more.
Tool-offset of active tool seems to be ok so.
Reinhard
___
Emc-developers mailing list
Emc-developers@lists.source
ime!
The only place where the difference between absolute and relative coordinate
may change is at tool-change or fixture-changes.
Behaviour may be ok from sight of motion, but it is not from user sight (with
me being the user ;) ).
cheers Reinh
Hi,
On Freitag, 10. April 2020, 13:25:04 CEST andy pugh wrote:
> LinuxCNC offers G43.1 to apply a temporary tool offset.
I tried it, and it works as documented.
How can I change this way (temporary) the diameter of a tool?
Reinhard
___
. Tools
have diameter, but no length.
If I remember well, that has been different when I wrote my app.
Have there been changes, that I missed?
Reinhard
___
Emc-developers mailing list
Emc-developers@lists.sourceforge.net
https
Hi
On Samstag, 11. April 2020, 05:54:54 CEST Robert Ellenberg wrote:
> Reinhard, that's a great point. It looks like there's a clear value in
> keeping the select / change steps separate if your ATC needs to pre-load a
> tool. This is what I meant by Fanuc-style tool changes
hobbit.
So I consider it a bad change, remove M6
Reinhard
___
Emc-developers mailing list
Emc-developers@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-developers
3 action which is not in the
> gcode program), I consider this to be an undesirable change.
I agree completely!
Thank you Dave for the detailed post!
Reinhard
___
Emc-developers mailing list
Emc-developers@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-developers
On Freitag, 10. April 2020, 13:25:04 CEST andy pugh wrote:
> LinuxCNC offers G43.1 to apply a temporary tool offset.
Oh, I overlooked that. Thanks for the hint.
Will check asap
Reinhard
___
Emc-developers mailing list
Emc-develop
Hi,
On Freitag, 10. April 2020, 14:19:27 CEST andy pugh wrote:
> On Fri, 10 Apr 2020 at 12:45, Reinhard
wrote:
> > I have. And some of my colleagues do it cause of lazyness for jobs with a
> > single tool only.
>
> Though presumably in that case there wouldn't be a G43
terrupted the job.
For that 3D-jobs we had special tool entries with length and diameter set to
0. Job was a Cam-output generated for that special milling head and tool
length was "measured" manually with high precision gauge blocks.
Reinhard
___
ld work for fade in / fade out too, without any changes.
Output acts completely independant of the command channel.
So decoupling modules increases freedom for system design and different
developers could dedicate on a single volume without having to care about
other modules.
W
should know, what to do, to restart an aborted job.
I don't think, that's a challenge for automation - and its of cause no reason
to blame the interpreter.
Reinhard
___
Emc-developers mailing list
Emc-developers@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-developers
gotten ...
Reinhard
___
Emc-developers mailing list
Emc-developers@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-developers
at tool definition.
Use G43 once is wrong. At least for me.
If you want to retain G43, make it an INI-option and never use it again in
gcode files.
... but than you have to change G40/G41/G42 too.
cheers Reinhard
___
Emc-developers mailing list
in same
tables. Not the way I think.
On Mittwoch, 8. April 2020, 12:20:37 CEST andy pugh wrote:
> To make use of a database, though, the tool data has to be removed
> from motion / NML. All that needs (as far as I can tell, please tell
> me if I am wrong) is th
ile io or database io would quite be
fine.
Reinhard
___
Emc-developers mailing list
Emc-developers@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-developers
able for each job, which then gets transmitted to the
cnc-controller.
Reinhard
___
Emc-developers mailing list
Emc-developers@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-developers
lem, that machine gets confused with tool-numbers and
machine reports a tool in spindle, that is not in the spindle, but stil in
atc. That's a big source for crashes if you don't be aware of. And its a
tedious job to fix that situation.
Reinhard
___
has a nml-queue, so why not use hal-pins to turn of task-
manager, send nml-commands from atc to motion during tool change and then turn
task-manager on again after tool change?
Reinhard
___
Emc-developers mailing list
Emc-dev
limited.
I often have to code jobs for about 20 tools or more and I know my colleagues
have jobs of much more tools used.
I prefer andys aproach of 999 tool entries, or a database with unlimited
entries.
Reinhard
___
Emc-developers mailing
ble as described would improve decoupling of modules and would be
good deed for users too :)
cheers Reinhard
___
Emc-developers mailing list
Emc-developers@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-developers
od old days?
Same may be true for inifile-token "TOOL_CHANNELS".
emcsvr.cc looks like one could specify the number of tool channels in ini-file,
but that token is not documented, nor does the code ask for the value of
TOOL_CHANNELS.
Reinhard
___
On Montag, 6. April 2020, 14:53:48 CEST andy pugh wrote:
> On Mon, 6 Apr 2020 at 13:10, Reinhard
wrote:
> > I tried to work out userspace hal component (which in my way of thinking
> > could be started without the use of halcmd), but all crashed direct after
> > start.
>
ssible, to interact with hal from a program started without
helper from commandline.
Ah - and I never tried to compile RTAI stuff. Had enuf trouble with preempt
kernels provided from debian.
Reinhard
___
Emc-developers mailing list
Emc-deve
the
directories explicitly specified with -I, -iquote, -isystem, and/or -idirafter
options (and the directory of the current file, if appropriate) are searched.
So with that option you have to specify, where the compiler should take
on release and package
testing.
That's not possible for one developer, who spents his sparetime for project
work.
> However I am not an inuxcnc developer ... so mine is only a shadow of
> opinion.
Same is true for me :)
cheers Reinhard
_
Hi,
On Mittwoch, 1. April 2020, 22:05:01 CEST andy pugh wrote:
> On Wed, 1 Apr 2020 at 05:48, Reinhard
wrote:
> > I added a F-word with a slow value (like F100) to the beginning and a
> > faster one (like F400) to the end. Without changing the feed overwrite,
> > the
Hi Rod,
On Mittwoch, 1. April 2020, 06:55:14 CEST Rod Webster wrote:
> Really should not be using the eyeometer here
Nice word :)
I use the combination axis/gremlin together with my app to test my app. The
testcase from Chris was new experience for me.
Reinh
which shows the actual feed rate from status feedback
channel. Status feedback reflects accelerating and decelerating ...
... so its pretty obvious, that the feed values come from motion and not from
interpreter.
cheers Reinhard
___
Emc-develo
Hi,
I'm thrilled that improving synchronization has become so much interest :)
On Dienstag, 31. März 2020, 18:10:37 CEST andy pugh wrote:
> On Tue, 31 Mar 2020 at 06:46, Reinhard
> wrote:
> > The testcase (3Dchips.ngc with different f-words) shows, that f-word is
> > not b
Hi Andy,
On Dienstag, 31. März 2020, 17:59:13 CEST andy pugh wrote:
> On Tue, 31 Mar 2020 at 16:07, Reinhard
> wrote:
> > It's like a C-compiler which does (part) of the linkers job, but so bad,
> > that lot of compilations units fail, as the compiler did not have compiled
>
l change not a queue buster?
> So keeping queue buster to the minimum is a desirable outcome.
Yes, but situation for optimization should be reasonable.
As I understand things, queue busters stand for: stop optimization here, do a
full stop and restart p
he functions
that change that structs data.
Will the changes be published to lc master?
Thanks!
Reinhard
___
Emc-developers mailing list
Emc-developers@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-developers
here is the tool-manager under settings, and the tooltable in
linuxcnc-format at tool-button.
I thought about generalization of atc too. I saw your datamodel and believe
that it can be improved (based on responsibility).
So if you're interested, we could talk about that too.
cheers Reinhard
ource line.
If I have that value accessible, many things could be improved. Even F-word
handling could be verified :)
Of cause, other problems will rise. But I'll head them, when I notice them.
OK?
Reinhard
___
Emc-developers mailing list
Emc
in status feedback.
Reinhard
___
Emc-developers mailing list
Emc-developers@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-developers
er see a global var in your code, I'll cut you a finger for each global =:O
cheers Reinhard
___
Emc-developers mailing list
Emc-developers@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-developers
Hi,
thank you for your attention.
> > I wanted to check other non motion commands and wrote a little gcode file,
> > which I put in the tests/motion-logger/basics directory.
>
> You want to log motion then g-code program is run?
When I discovered motion-logger, I had a misunderstanding and
Hi,
I wanted to check other non motion commands and wrote a little gcode file,
which I put in the tests/motion-logger/basics directory.
Then I started test.sh, but it blocks.
After adding a log message to test-ui.py I realized, that the new test is
started.
I thought, may be toolchange is the
ven at first singlestep the Feed is set
to the last F-Code from file. So there is no delay and result is absolutely
wrong.
Glad you're already working on that problem :)
I noted a bug in single step processing in my app - so have to fix that first.
Che
motion is executed.
That means, feed is zero after Pause or at tool-change.
Cheers Reinhard
___
Emc-developers mailing list
Emc-developers@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-developers
Hi Chris,
On Samstag, 28. März 2020, 14:36:18 CET Chris Morley wrote:
> Sorry I didn't mention th name: feedcode_message
Whow! You're really dedicated to LC.
I'm impressed!
So let me just say thank you - even if I don't understand your work (yet?).
Cheers Reinh
Hi Chris,
On Samstag, 28. März 2020, 14:36:18 CET Chris Morley wrote:
> Sorry I didn't mention th name: feedcode_message
Thanks for the link! Will have a look.
Just rebuilt lc with debugging flags enabled.
Any hint, how to best get into taskmanager?
d some.
So I guess, I have to strip down the taskmanager to get into it.
Grunt work is coming =:O
cheers Reinhard
___
Emc-developers mailing list
Emc-developers@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-developers
delete for over 20 years.
I never have used block delete so far.
In case I want to use only part of existing program, I use "run from line"
together with M0
cheers Reinhard
___
Emc-developers mailing list
Emc-developers@lists.sourc
Hi Andy,
> > I wonder, why motion sets up nml-queues if it is part of the same process
> > than task ...
>
> This is not an area I have studied in depth, but I think that the
> motion queue is not the same as the NML queue.
>
> I think that the motion queue consists only of these commands:
>
d as a filter function of the gcode lister.
GCode-interpreter should support that tag too - and as it is documented, I
suppose, the gcode-interpreter already supports that feature?
If not, it could be implemented like line comments.
cheers Reinhard
__
tion part was quite easy to write.
But then I tried to write the taskplanner / motion optimizer and had to
realize, that I don't know enuf math to do the job. So I gave up.
may be I could help fixing things around taskmanager?
Reinhard
___
Emc-developer
different or do they follow nml-paradigm?
Reinhard
___
Emc-developers mailing list
Emc-developers@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-developers
s.
The valuable time information is the time, of tool usage and time, that
processing of gcode has consumed.
In my fantasy: just a timer that runs as long as gcode processing is active
and pauses at M0 or the like and resumes after user continue action.
Reinhard
__
is emcmot?
Is it a thread of linuxcncsrv or a thread of milltask?
Or is it a module loaded by rtapi_app?
First of all, I'd like to be able to identify all roleplayers understand their
communication channels and find the corresponding code.
Reinhard
___
Hello guys,
I'm Reinhard and I wrote an UI application for linuxcnc which uses nml-
interface only. The app works fine so far.
But on testing it together with axis and linuxcnc-simulation I realized
several quirks and shortcomings of the backend like:
- no process time
- current line number
101 - 168 of 168 matches
Mail list logo
