Trajectory Planner using Ruckig Lib

Hi,

Thanks for testing.

Notice the pauses at various points. 
Yes between the moves the pause could be less.

This can mean a few things :

- Somewhere a delay is given between 2 gcode segments, like Rodw is mentioning.
  But i suspect that is not the case.

- Scurve software need a tiny end velocity between 2 paths, wich is currently not present in the planner.

Todo: check If the scope is not showing the delay between the moves.

S-Curve freezing
This "nan" message is clearly a tiny error and will stop your program and so on.
But is quite easy to solve in next github update.

As for my conclusion, it's not bad for Joco's first attempt.
I would see 4 things changing, like solving the snerping sound in Joco's machine,
remove pauses between moves, solve the nan error, solve the cmake installation.


 

Grotius wrote:

 solving the snerping sound in Joco's machine, 

I had to put headphones on to hear the snerping sound! 
I wonder if its harmonics induced by jerk limited velocity or something loose?

@Joco,

When the isnan occured, i suspect you did a mdi move for a given arc?
Do you remember the given gcode that caused the isnan? Thanks.



 

thanks
sucess build after build orginal linuxcnc 2.10

i will try test with real machine then report

regards

Hi,

I have seen the build problem on the virtual machine "vm", where it's running a debian 12.

It's related to halcompile. Halcompile isn't working properly somehow.
It's not finding the Makefile.modinc somehow. Even the rip-environment is done.

Is now solved by changing the build system procedure.

• compile lcnc the original way.
• overwrite build with cmake.

@Tivoi,

Good luck man. I hope you can "rock and roll a bit" with the scurve.
I am still lerking on a vm.

Github is just updated for changing the build system, also the isnan error is prevent freezing now.
More printf output when isnan occurs.

 

I will pull and run another test today.

Another quirk i was encountering was if i touched off in axis to shift the wcs 0,0 axis would not run the programme. In fact generally became unresponsive. Will see if can reproduce today and trap info from console. I’ll try grabbing that all to file for easier reference.

Cheers - J

rodw wrote:

Grotius wrote:

 solving the snerping sound in Joco's machine, 
I had to put headphones on to hear the snerping sound! 

Or its the squeaky cable chain you guys are hearing.  No idea what snerping even is and I’m in the same room as the thing.  

Did some sim checks this morning as had noticed some off errors appeating.   Two files attached:
1. is a clean Axis mm UI start, move the wcs x,y to a new location in machine space, stop. This on 2.9 linuxcnc
 

2. is the same process but on 2.10pre compiled from the repo.
Observation:
USRMOT error command 31 due to a timeout.
Tinker callback error which results in Axis becoming unresponsive.  This happened after/on touching off Y axis to 0. I tested it the other order (y then x) and the error triggered on the X touch off.  So it seems to be the second touchoff activation triggering things not what axis it is.
 

Not sure if these are tied to a more general issue with 2.10pre or with changes in this particular code line.   I get the same problem under 2.10pre with either s-curve or standard tp's.

Cheers - James.

 

2.10 removes lube and a couple of other pins. check in the docs that what is in your cooling.hal still exist.
 

ccoling hal? What is that? Is that part of Sim?