Closed ghost closed 7 years ago
I'm going to insert a simple post-processing step which strips this out and the lone S value. The "G0 F1500" throws off vector ops when running on Smoothie.
It now strips "G0 F..." and lone S values from lw.raster-to-gcode output.
Found. That's a parameter on https://github.com/lautr3k/lw.raster-to-gcode rapidRate
@lautr3k any way to prevent its use? anyway, G0 could be exposed as a machine setting or raster setting.
Traditionally G0 feed is a setting that gcode can't touch. Reprap people added G0 F because reprap is weird. Smoothie added it because of Reprap. We don't need it and it messes up vector ops which follow.
If we add it as a setting then we'll confuse users with grbl, which does the right thing by ignoring it.
Ok, but when RepRap comes, this issue will arise for sure...
It's a feature, not a bug ;) Can't we just ignore the F value, if rapidRate is set to 0, and for the moment fix rapidRate to 0?
Later on, when/if we integrate 3dp, we can expose this feature to the user.
@tbfleming Isn't adding a postprocessor for fixing something that we could also fix on generation not a bit simplistic?
We can't fix it on generation :(
Unless we fork it and not use the npm package.
If @lautr3k doesn't want to move this repo into the Laserweb organisation (or at least give us push rights on his repo), then we realy should fork it, because we depend on it.
I found we already have a fork in the organisation (21 days old).
Ehm.. I did a fork and contributed some code to his repo. Simple as that.
@jorgerobles Is your fork the one in the laserweb organisation? I would prefer to use that.
He was opposed to solving the lone S problem and claimed it was a Smoothie problem. We need to post process that out because it messes up both Smoothie and grbl-lpc.
@cprezzi will do.
I've tested PR locally and works ok, but have problems building and using our git instead npm. Awating for now.
@tbfleming Postprocessor is a nice idea commented before. How can we make that approach generic so make postprocessors contributed by community to their needs?
I haven't put any thought into that. There are potential security issues with supporting user-contributed code, especially under Electron since it provides direct file system and OS access.
Yes. We could always let them contribute via PR and test before compile.
If @lautr3k doesn't want to move this repo into the Laserweb organisation (or at least give us push rights on his repo), then we realy should fork it, because we depend on it.
@cprezzi You have to do as @jorgerobles just did for the https://github.com/lautr3k/lw.raster-to-gcode/pull/4 PR :
set-rapidRate-optional
, or something that describes the new feature/fix.@tbfleming I will fix the lone S value.
@lautr3k great!
@jorgerobles @tbfleming lw.raster-to-gcode@0.6.3 is out :
@DarklyLabs please test to close :)
Experiencing the same problem with vectors. Seems to occur only after a job is aborted before completion. Once again not repeatable at the moment. I have asked the team to export/examine the gcode next time it occurs.
@DarklyLabs Please especially test for raster. Vector is completely different code.
Since the G0 is being stripped, we have not experienced this slowdown with raster engravings.
@domenic-d that recent issue you told me is related to this one.
I've postprocessed gcode so the first move is prepended to Z move. Also, that fixed an open issue at https://github.com/lautr3k/lw.raster-to-gcode/issues/6
Looks like @jorgerobles changes have resolved this. Thx! Closing for now.
Extracted from a G+ discussion
There is a bug in LW4 which we have not been able to repeat consistently where the G0 feedrate often overrides our firmware settings (smoothie).
From @tbfleming I just repoed it. lw.raster-to-gcode is inserting a rogue "G0 F1500". Please file a bug.