[feature request] separate extrusion multipliers for infills and perimeters

[curiouspl2]

excuse me if tone of my posts is not polite/proper, i'm not native english speaker :)
i do appreciate your work a lot and i try to follow development.

just bumped on : #2441
and concluded it's not really good thing to calibrate either up to solid fill nor to single wall .

either way one ends up in some kind of compromise, which is not really producing ideal prints.
either walls and perimeter get not enough material (and also slic3r 'thinks' it is extruding with larger width it is extruding in reality , causing artifacts like wrong overhangs or incomplete covering of steep surfaces, plus one looses resolution) ... or solid fills are overfilled.

also this raises problems with anchoring of bridges , solid infill overlap with perimeter issues etc.

simple solution would be to introduce flow rate correction (extrusion multiplier) for features like solid infill and perimeters. you do write that it's variable with material - so perhaps it should go where it belongs, so into filament tab, just as extra extrusion multiplier for various features.

i think it's disputable which features require corrections.
personally i think it's very printer/print/material dependent - like bridge flow ratio.

[simonkuehling]

Just a quick thought regarding the calibration of extruder steps (leaving the overlap issues out for the moment): From my understanding the only really precise method for this would be to measure the diameter of one single bridged extrusion. We can safely assume that the cross section of a bridged extrusion is round - and die swell does not count in due to the filament beeing pulled from one anchoring point to the other.

Do you agree?

Could this possibly boil down further configuration to only the overlap factor whereever two extrusion lines are touching each other (perimeters, solid infill, infill-perimeter-contact-points, etc)?

[nophead]

Since what comes out of the extruder equals what goes in, extruder steps
are normally calibrated by extruding 100mm of filament and measuring how
much goes in. Together with the filament diameter that determines the
volume extruded.

On 22 December 2014 at 09:34, simonkuehling notifications@github.com
wrote:

Just a quick thought regarding the calibration of extruder steps (leaving
the overlap issues out for the moment): From my understanding the only
really precise method for this would be to measure the diameter of one
single bridged extrusion. We can safely assume that the cross section of a
bridged extrusion is round - and die swell does not count in due to the
filament beeing pulled from one anchoring point to the other.

Do you agree?

Could this possibly boil down further configuration to only the overlap
factor whereever two extrusion lines are touching each other (perimeters,
solid infill, infill-perimeter-contact-points, etc)?

Reply to this email directly or view it on GitHub
#2443 (comment).

[alranel]

@curiouspl2, you're confusing single walls with perimeters. They're not the same thing. Adjacent perimeters are spaced just like adjacent solid infill lines. If you calibrate properly for solid infill you'll get correct perimeter overlap as well.

@simonkuehling, free-air diameter only depends on nozzle diameter and tells nothing about extruder steps... (Well, it also depends on how much you stretch the filament but that still tells nothing useful about steps)

[simonkuehling]

@alexrj, but provided that the extrusion is stretched as a bridge and does not sag (which is true for ABS at least due to tensioning from shrinkage during cooldown), the diameter must be directly dependant on the extruder steps. It cannot be thicker than nozzle diameter of course, but if it is just so slightly stretched, it should yield accurate measurements... Or am i missing something else? My idea was to have a user-friendly "print a test-piece and measure" calibration method like it is done with single walled cubes...

@nophead, while measuring the input-side of the extruder for calibration works as well, it is much less comfortable to implement as a UI guided procedure than a calibration-cube-print method.

Maybe i did not get the full picture why the single-walled-calibration-piece procedure is unreliable. Is the theory that the cross section of a single wall extrusion is not always exactly a rectangle with two half circles at each end like described in http://manual.slic3r.org/advanced/flow-math, and therefor the wall thickness cannot be predicted accurately?

[nophead]

I only ever measured the single wall thickness a few times many years ago,
to check the formula was accurate. I don't use slic3r but with Skienforge I
only need to calibrate the E steps to get dimensions and fill correct. So I
have no reason to doubt the formula.

The ends will be semicircular due to surface tension as that is the minimum
surface area. Maybe if your plastic is very liquid it might slump a bit due
to gravity but I don't think the error due to not being a perfect
semicircle will amount to much.

For the infill I use the 100% rectangle formula. That might be a little
overstuffed for some peoples taste as you can get a nicer looking surface
with it a bit lower for viscous plastics.

On 22 December 2014 at 12:34, simonkuehling notifications@github.com
wrote:

@alexrj https://github.com/alexrj, but provided that the extrusion is
stretched as a bridge and does not sag (which is true for ABS at least due
to tensioning from shrinkage during cooldown), the diameter must be
directly dependant on the extruder steps. It cannot be thicker than nozzle
diameter of course, but if it is just so slightly stretched, it should
yield accurate measurements... Or am i missing something else? My idea was
to have a user-friendly "print a test-piece and measure" calibration method
like it is done with single walled cubes...

@nophead https://github.com/nophead, while measuring the input-side of
the extruder for calibration works as well, it is much less comfortable to
implement as a UI guided procedure than a calibration-cube-print method.

Maybe i did not get the full picture why the
single-walled-calibration-piece procedure is unreliable. Is the theory that
the cross section of a single wall extrusion is not always exactly a
rectangle with two half circles at each end like described in
http://manual.slic3r.org/advanced/flow-math, and therefor the wall
thickness cannot be predicted accurately?

Reply to this email directly or view it on GitHub
#2443 (comment).

[curiouspl2]

alexrj : not really, because overlap for perimeters, which are constrained only from one side is different than overlap in completely solid fill. single-or-few perimeter objects, measured for dimension can have more or less material and still be printed correct - with just slight dimension shift , while solid infill will overflow with same settings.
I am adressing this issue - there is no way for user to fine-tune that.

This applies also for case in which someone wants solid fill or perimeters 'detuned' for some purpose, i.e. when printing with 2nd extruder using much larger nozzle.
Perhaps solution/hack would be to somehow allow 'fake' multiple extruder setup,
so user could set up multiple extruders using multiple filament settings even if having only one physical extruder. this could allow to also change i.e. temperature settings and flow rates for solid infill , perimeters, etc.

[alranel]

@simonkuehling, there's general consensus that the width of a single extrusion is predictable using the known formula. However, the single-wall procedure only helps with calibrating E steps (i.e. amount of flow). That's not the only thing that would need to be calibrated: there's also the overlap amount between adjacent extrusions (like in perimeters or solid infill). This is explained in http://manual.slic3r.org/advanced/flow-math in the Spacing paths paragraph. The default spacing/overlap is consevative, as it guarantees good bonding and no gaps. Depending on plastics, such overlap might be too much due to viscosity. That's why even if the single-wall calibration was used, users often need to reduce the Extrusion Multiplier according to visual inspection of solid infill.
Of course one might think that we need to expose the overlap factor so that users can customize it. Maybe... but exposing that stuff is very Skeinforge-y and thus not very friendly. First I'd like to design a recommended procedure...

[curiouspl2]

i think that is why there are 'basic' and 'expert' levels of settings. apperance of extra tuning in 'expert' mode is not a 'style' decision imho ;)
i'll add that solid infill being not 100% solid is usefull for other purposes, like printing light-duty objects (toys, figurines, etc) not requiring to be watertight, though personally i would like 'virtual extruder' solution as it allows to tune actual flow rate to produce (and tune) ideal solid infill while not influencing perimeters (and their dimensional stability!) while tuning solid infill.