Page 2 of 2
Re: "Coast" To Reduce Stringing?
Posted: 13 Apr 2016, 13:53
by stefan-witbox
Welcome back !! We missed you too.
Re: "Coast" To Reduce Stringing?
Posted: 13 Apr 2016, 16:35
by Dreide
lonesock wrote:It probably won't be a traditional coast. When you stop driving the extruder the plastic oozes out to relieve the back-pressure, and as it oozes the pressure drops. So it should follow a roughly exponential decay. I plan to match that with the head velocity (to a point) then the optional retract & wipe. The curve will be a function of a bunch of things (material, distance from extruder to nozzle, temperature, nozzle diameter, etc). The builtin model will be fairly simple, but I am considering adding a plugin architecture to let you printer-geniuses have the most advanced control over extrusion.
Repetier (and also Marlin, I think) have the ADVANCE feature which tries to account for back-pressure (see for example
https://github.com/repetier/Repetier-Fi ... re-control). But the implementation is flawed, because it is difficult to realize the necessary formulas in the framework of the path planner. So it would be great to have implemented something like ADVANCE at the slicer level.
Re: "Coast" To Reduce Stringing?
Posted: 15 Apr 2016, 13:05
by plexus
lonesock wrote:Jonathan
P.S. It is extremely nice to come back to KISSlicer development! I missed it, and you guys!!
Hi Jonathan! Nice to see you back on the forum! Lots of people asking about updates and fixes... but in the meantime Kisslicer works pretty well. It's still my slicer of choice.
Re: "Coast" To Reduce Stringing?
Posted: 16 Apr 2016, 12:02
by lonesock
Thanks, everybody!
Jonathan
Re: "Coast" To Reduce Stringing?
Posted: 16 Apr 2016, 22:07
by 626Pilot
lonesock wrote:Hello! This is in the top-10 on my ToDo list. I'm in feature freeze right now to actually get the 1.5 release out the door (so the 1.5 will look mostly like the betas you've been using but with some internal bug fixes and a few small features). So coast will go in the 2.0 branch.
It probably won't be a traditional coast. When you stop driving the extruder the plastic oozes out to relieve the back-pressure, and as it oozes the pressure drops. So it should follow a roughly exponential decay. I plan to match that with the head velocity (to a point) then the optional retract & wipe. The curve will be a function of a bunch of things (material, distance from extruder to nozzle, temperature, nozzle diameter, etc). The builtin model will be fairly simple, but I am considering adding a plugin architecture to let you printer-geniuses have the most advanced control over extrusion.
Cool!
I didn't realize you were back and working on this, so a few days ago I implemented a simple (no exponential decay) coasting algorithm as a postprocessor:
https://github.com/626Pilot/KISSCoast
It takes a few seconds to process a ~15MB file, and early results are promising. WAY less stringing, so I would say it's definitely worth it to implement this feature.
One of the important things I did with my script was to allow different coasting values for regular and prime pillar paths. The pillar paths are significantly overextruded, leading to blobs of plastic that get dragged from the wall to the object. With a longer coast value on the pillars, this effect is greatly reduced.
Re: "Coast" To Reduce Stringing?
Posted: 18 Apr 2016, 10:41
by lonesock
Nice! Yes, the coasting will have to be a function of the extrusion volumetric flow rate. I am in feature freeze right now, trying to get the 1.5 build out (I just posted a new one to the Dropbox folder), so your post-processor will be super-useful. Would you be in favor of the plugin architecture I mentioned? (Of course, it would be nice if KISSlicer just automatically Did The Right Thing. [8^)
thanks,
Jonathan
P.S. The new build will hopefully help with the stringing that you are seeing, though.
Re: "Coast" To Reduce Stringing?
Posted: 01 Jun 2016, 12:12
by calinb
That sounds great, Jonathan! I would also like to second Penskeguy's idea, above, to "further enhance the Seam Hiding by effectively slowing the flow rate at the end of the layer." I think more can be achieved in seam hiding with his suggestion and it certainly warrants investigation.