Trajectory Planner using Ruckig Lib

I have had another go.  First made sure I could compile standard lcnc.  Got that sorted with all the dependencies correct.

Cloned the repo and as advised and ran compile.   I will book a ticket against the repo.  But it is complaining about missing gtk.h .

Thanks.

@JPL,

I may be wrong (or is it just wishful thinking?) but the second video seems to suggest that jog while paused is possible with this new TP ?!?

That's not the case. But i get your point.
It should be possible, but takes time to get it coded.

 

@Hi Joco,

For information, you could build the repository like the original lcnc way. This should work.
Then you can run the build_all script over it.

But it is complaining about missing gtk.h .

This message could be a missing dependency, it's using gtk for halscope etc.
Maybe try:
$ sudo apt-get install libgtk-3-dev
 
 

That missing dependency fixed the build.  Sim not running.  See additional comments on ticket.

hi, i still have erros

current folder:
/home/cnc/linuxcnc


Running equivalent of Sudo Make Setuid.

Start linuxcnc.

LINUXCNC - 2.10.0~pre0
Machine configuration directory is '/home/cnc/linuxcnc/scripts/../configs/sim/axis'
Machine configuration file is 'axis_mm.ini'
Starting LinuxCNC...
linuxcnc TPMOD=tpmod_scurve HOMEMOD=homemod EMCMOT=motmod_scurve
Note: Using POSIX non-realtime
Found file(lib): /home/cnc/linuxcnc/lib/hallib/core_sim.hal
tpCreate, set look_ahead to : 10
tpCreate. set tp->queuesize to: 2000
tpSetCycleTime to: 0.001000
ddt: dlopen: /home/cnc/linuxcnc/rtlib/ddt.so: cannot open shared object file: No such file or directory
/home/cnc/linuxcnc/lib/hallib/core_sim.hal:9: waitpid failed /home/cnc/linuxcnc/bin/rtapi_app ddt
/home/cnc/linuxcnc/lib/hallib/core_sim.hal:9: /home/cnc/linuxcnc/bin/rtapi_app exited without becoming ready
/home/cnc/linuxcnc/lib/hallib/core_sim.hal:9: insmod for ddt failed, returned -1
Shutting down and cleaning up LinuxCNC...
/home/cnc/linuxcnc/scripts/linuxcnc: line 668: /home/cnc/linuxcnc/scripts/axis-remote: cannot execute: required file not found
USRMOT: ERROR: command 31 timeout
emcMotionInit: emcTrajInit failed
Note: Using POSIX non-realtime
LinuxCNC terminated with an error. You can find more information in the log:
/home/cnc/linuxcnc_debug.txt
and
/home/cnc/linuxcnc_print.txt
as well as in the output of the shell command 'dmesg' and in the terminal

regards

I think there is something still missing from cmake.  This worked for me:
1. Build linuxcnc the “normal way” using autgen, configure, make from src directory. 

2. Build using cmake. The cmake script now runs axis successfully.

cheers J. 

Yes and thanks Joco for your testing.

Something is missing indeed. But i think cmake is complete. It's problably in the lcnc python scripts.

If you copy the all the python scripts from the src dir into the bin dir,
this may solve the lcnc startup problem.
I think one of the src python scripts has to be in the bin dir.

But i can't reproduce this startup issue at my pc, here it's starting lcnc.

Grotius wrote:

Yes and thanks Joco for your testing.

Something is missing indeed. But i think cmake is complete. It's problably in the lcnc python scripts.

If you copy the all the python scripts from the src dir into the bin dir,
this may solve the lcnc startup problem.
I think one of the src python scripts has to be in the bin dir.

The compile side the of the build script in the cmake directory probably is done the rest is not. That was my point but I was not specific enough. The more correct remark would be that the script build_all is missing things. I make this comment on the basis that running build_all should compile and successfully run axis with the new module. If it is not achieving that it is missing either setting correct paths or copying needed files to the appropriate locations or whatever the root cause is.

But i can't reproduce this startup issue at my pc, here it's starting lcnc.

 

My suggestion is you spin up a virtual machine (VM). I use vmware player for all my lcnc dev instances, mainly because it supports the VTK plot correctly. Whenever I get an issue that I can't reproduce I spin up a clean VM to check from. That way I know it is not polluted by anything I have done before. Only when I have things behaving in the VM do I move to physical kit connected to hardware.

I will do some digging myself but in honesty it feels like I am taking pot shots in the dark hoping I get lucky.

Cheers - J.

Hi All,

I just finished another scurve motion planner. tpmod_scurve_skynet.so

Features :

1. Look ahead at runtime.
2. Path rules.
3. Motion forward, reverse.
4. Jogging while pause.  (joke
   
   )

Video :


Github is updated with this new planner.
We have to add another few path rules, like tiny arc's that are colineair etc.
But for now i am happy with the result.

@ Joco,
Good idea to use a vm for resolving the startup issue.

Sooo impressed with what is being done here!  I am trying to get a test machine set up to actually try this on real hardware. I had a build fail last night but that could’ve been just getting a cut of the code when it was in flux. At least that is how it feels given I have had a successful build in the past. 

I am very keen to see what difference it can make on my current plasma table. Especially how far beyond my current 1000mm/s/s acceleration limits i can push before jerk/wobble issues return.  

cheers - J.