Calibration Objects, Temperature and Retraction Towers
This is related to the @hyiger version of Prusaslicer with calibration, but it's also generic, so I'm giving it its own thread. Caveat, I'm a test and metrology geek and can't help myself. IMO, a good test gives a clear result that's easy to interpret and that you have high confidence in. That's been my experience with chevron PA objects and thin-wall boxes for extrusion multiplier values. You look at it, measure it and job done. Temperature towers are so-so, at least with PLA, as it can print well over a wide temperature range. Still, the towers do the job. Fan towers are similar. A wide range of acceptability is probably a good thing, lest we never successfully print anything.
Less satisfactory have been YOLO plates for extrusion multiplier and retraction towers. IMO, the YOLO method is for the person who doesn't own micrometers or a decent vernier. It doesn't measure what you're really interested in and I'm not going to waste another brain cell on it, unless you want to. My main source of frustration is retraction towers. Seeing no difference in stringing from top to bottom, I started looking at images from a Google search. Out of all the images, there were only one or two that I thought showed anything of significance. All the images showed stringing that looked more like wet filament, than something connected to the retraction setting. My conclusion was that not many people are successfully using retraction towers. So, to start off, I have two questions:
1) Do your retraction towers show a clear demarcation between stringing and not, allowing for precise measurement and setting of retraction distance?
2) Are there some better objects and procedures for setting retraction distance (and speeds) than the tower?
I ask #2 because I've seen significant differences in the smoothness of vertical walls on objects where I had zero retraction set. That suggests there may be a better way.
RE: Calibration Objects, Temperature and Retraction Towers
I never bothered to print retraction towers. I just follow Bondtech's advice for their direct drive extruders : a retraction distance equivalent to the diameter of the nozzle ( or a bit more). Let's say 0.4 to 0.6mm. for 0.4 nozzles. This is for rigid filaments. For flexibles my settings are 1 to 1.5mm. I assume that if there's too much stringing probably it isn't related to the retraction distance but to other factors ( moist filament in 90% of cases).
RE: Calibration Objects, Temperature and Retraction Towers
This is related to the @hyiger version of Prusaslicer with calibration, but it's also generic, so I'm giving it its own thread. Caveat, I'm a test and metrology geek and can't help myself. IMO, a good test gives a clear result that's easy to interpret and that you have high confidence in. That's been my experience with chevron PA objects and thin-wall boxes for extrusion multiplier values. You look at it, measure it and job done. Temperature towers are so-so, at least with PLA, as it can print well over a wide temperature range. Still, the towers do the job. Fan towers are similar. A wide range of acceptability is probably a good thing, lest we never successfully print anything.
Less satisfactory have been YOLO plates for extrusion multiplier and retraction towers. IMO, the YOLO method is for the person who doesn't own micrometers or a decent vernier. It doesn't measure what you're really interested in and I'm not going to waste another brain cell on it, unless you want to. My main source of frustration is retraction towers. Seeing no difference in stringing from top to bottom, I started looking at images from a Google search. Out of all the images, there were only one or two that I thought showed anything of significance. All the images showed stringing that looked more like wet filament, than something connected to the retraction setting. My conclusion was that not many people are successfully using retraction towers. So, to start off, I have two questions:
1) Do your retraction towers show a clear demarcation between stringing and not, allowing for precise measurement and setting of retraction distance?
2) Are there some better objects and procedures for setting retraction distance (and speeds) than the tower?
I ask #2 because I've seen significant differences in the smoothness of vertical walls on objects where I had zero retraction set. That suggests there may be a better way.
I haven't really put a lot of time in the calibration tests beyond the initial implementation so there are bound to be bugs. I plan to review everything this weekend. I see you've created an issue in: https://github.com/hyiger/PrusaSlicer/issues/14 this is perfect. It's easier for me to respond to it than here. The YOLO test is another story. It's very difficult to implement without post g-code modification. It has to do with the fact that PrusaSlicer doesn't support ironing with an Archimedean spiral and I'll need to figure out how to get it ported over from Orca.
RE:
@hyiger, my hope was for a more general discussion of objects and usage, though hopefully you'll consider anything new or interesting that turns up. My hope is somebody will show a retraction tower that looks the way one would expect it to look, as none of the examples online are very confidence inspiring. I know there's a lot you still need to look at as it's really early days for what I know can't be a full-time project! I'm using your program of course, because my abilities are still a bit iffy when it comes to inserting custom gcode (and probably always will be).
Just for fun, here are a couple things that work really well-
The chevron for PA gives an easy to interpret result. It's a bit bulged at the bottom, things get under control 1/4-1/2 way up, then the edge goes defective after that. The internal angle is a bit harder to interpret, but follows in the same way.
I ran the cube and measured it, then divided for the new extrusion width, target being 0.45 mm. I ran another cube and measured it...
The silly thing was within about 2 microns everywhere! Not bad for a "consumer grade" Core One.
RE: Calibration Objects, Temperature and Retraction Towers
@hyiger, my hope was for a more general discussion of objects and usage, though hopefully you'll consider anything new or interesting that turns up.
I've got Saturday and Sunday blocked off to refine all of the additions I've made to PrusaSlicer. Currently finishing up full OpenPrintTag integration. I hope some of this stuff makes it's way into the mythical PS 3.0
The YOLO test are annoying in that they are somewhat subjective in my opinion. The retraction tower as well. I do think though I need to look at the fan speeds a bit more closely in the model.
RE: Calibration Objects, Temperature and Retraction Towers
Is the caliper method for extrusion multiplier really meaningful? There is surface roughness and compression under force, many moving parts in the equation. Maybe it hits the right spot on the distribution but here's at least someone saying the method is not reliable.
I don't really have an opinion: my extruders are ballpark-accurate (100 mm of extruded filament measure ~ 98..102 mm - close enough given the effective wheel diameter depends on how deep the teeth dig into the filament) and everything beyond that is "cooking" without any claim to scientific accuracy. Usually I leave it at 100%.
I've run quite a few of the Orca calibration tests, but frankly, they didn't yet make me change a single parameter over the defaults for PETG (which is an "easy" material - for engineering filaments this may look differently). Print hot for layer adhesion and not faster than the hotend can melt, but that's just common sense confirmed by the tests I ran.
RE: Calibration Objects, Temperature and Retraction Towers
I'm not smart enough about this stuff yet to have strongly held opinions, but I'm a bit smarter than the average bear with regards to measurements of all types, and very fussy about print quality. IMO, the final arbitrator of the calibration methods is print quality, otherwise, what's the point of all the effort? I've found the Prusa defaults and profiles work very well, but if I can find an improvement, so much the better.