Well Jim...you got me to give Cura another try yesterday. I've never been a big fan...but there are things about it I want to like (how quickly it slices, the live preview, and more efficient gcode...oh and the ability to slice the 3d lab prints planes.)
So I fired it up again and finally tried my idea on how to apply my calibration method. I changed the settings to do a 0.48mm extrusion width (Cura defaults to 0.4 on a 0.4 nozzle even though any number of sources will say that's not a good idea due to die expansion and how extrusion works - but I won't go into the details here.) I also changed it to do only 1 wall thickness and no top/bottom. Then I fired up openscad and created a 10mm cube which I brought in and sliced. Sent the gcode to my printer...and...it only printed about 2-3mm and then just stopped extruding. Not a clog - the extruder just wasn't even trying anymore. But the wall thickness on what did print was close. (within about 0.04mm)
And 10mm was smaller than I meant to use. I scaled it up to 40mm and tried again...same thing, got a few mm up and then stopped extruding. I finally measured one of my usual test prints and realized 20mm is what I usually use so re-scaled to that and tested again...and once again it just stopped extruding a dozen or so layers up.
Grrr. About here I remembered this was the problem I ran into the last time I tried doing a 3D lab prints print. It would start out well and look good...then just suddenly stop extruding.
I looked at the gcode and nothing jumped out at me as wrong...but something must be.
Hmm...what version of Cura is this anyway....2.3.1 Well, that's not the latest version, in fact it's a full major version off...so let's grab the latest and give it another try.
Well, with the latest 3.0.4 my test worked! It was a tiny bit overextruded...my 0.48mm walls came in at 0.54mm (or so...I didn't write it down and don't remember exactly.) Whatever it was I did the math and it looked like about .88 - .89 would be my adjustment if I was using slic3r...but Cura wants a percentage...so put in 89% for Flow and tried again. Now I'm getting an average of 0.48 and am happy!
It's getting late though...not enough time to run a Benchy. So I'll run off one of my favorite
make magazine tests the negative space tolerance test. This is an important test if you're doing parts that need to fit together accurately or are trying to do print in place designs with moving parts that don't have to be assembled after printing. Used almost stock settings...just changed to 0.25mm layers since that's what I know is efficient on my machine and put in the 89% flow rate.
Note - this is something I'm really not big about on Cura...the flow rate adjustment is part of the print config not part of the filament config. This is somewhat infuriating since it's something that varies from roll to roll of filament so really should be a filament setting. Having to manually track it and change it for each filament is really backwards.
Anyway...I ran the print...as it was printing I noticed Cura's infill is totally different than slicers and I'm not a fan....there are probably deeper hidden settings (I just checked...there are) but the default is simple rectilinear which isn't very good IMHO. I usually use a 3D honeycomb in slic3r which uses less filament, prints quicker, but gives a more solid part.
Let's compare:
Cura 20% infill:
Slic3r 15% 3D honeycomb infill (My usual default):
Slic3r 15% rectilinear to be more fair:
Slic3r 20% rectilinear to be totally fair:
Few things to note here.
Slic3r colors perimeters different than infill so it's easier to see what's going on. Not sure if there's a way to change that in Cura but if there is it's hidden well enough I didn't find it.
Also, I'm using 3 perimeters in slic3r but cura's default (wall line count) is 2. On most parts I prefer using more perimeters and less infill. A number of people have tested this and shown that most of the strength in a part comes from the perimeters and more infill doesn't make a huge difference. However infill does help keep the geometery of the part accurate by giving a framework for later layers to rest on. So I generally use about 15-20% for that reason. My MPCNC parts I did print at very high infill because those need to be very dense parts - they get done at 55% infill which results in a very solid part - but they're the exception rather than the rule.
Looking at the infill though. Two things jump out at me.
1) Cura does infill across the entire STL not individual parts. You can see the grid just goes right across the pegs. I tried a few of the other cura infill patterns and it seems to be consistent...the "Cross 3D" pattern may be different...but hard to tell on something this small. Slic3r by comparison evaluates each separate piece of a STL individually...which means...
2) Slic3r does the pins as solid parts due to how small they are. Cura just treats them as part of the whole and gives them sloppy looking infill.
I'm not big on how Cura treats an entire STL as one part and doesn't try to evaluate it more intelligently like how Slic3r does. Whether the pins need to be solid or not is debateable, and I could turn down the setting for when to go to solid in Slic3r if I wanted....but I don't see any option in Cura to get it to do the same.
Also of note - 20% in Cura results in a lot less infill than 20% in Slic3r.
But the big question is...how did the print do? Well...not bad at all:
In fact, most of those pins fell out by themselves. I only had to push one of them out with my fingers. Yes the .2mm pin is stuck...it almost always is, only occasionally am I able to remove it and then only by cheating and using the table or a tool to assist. I can see light around most of it though so it's very close which is good.
There is some ringing visible in the print....but that's not surprising....the 20mm test shells have less ringing showing than I usually get so that's nice...and the corner where the Z changes on those test shells IS cleaner than I usually get (Cura is setup with more retraction by default than what I usually run. (I run 1mm in slic3r but Cura defaults to a whopping 6.5mm! Probably because ultimaker uses bowden extruders by default...I used their "Prusa i3" settings but since Cura puts retraction on the material instead of on the printer like Slic3r does it probably doesn't change that based on the printer you select...something else I'm not big on about Cura.)
So overall print quality is ok...but there are still a lot of things about Cura I'm just not a fan of. But it is nice to know I've finally got it working decently so I'll try a few more tests with it and maybe some 3D lab test parts to see how it does.
I think the big "problem" with cura is that it's so linked to ultimachine and their printers - the way things are setup are great for them...but a bit of a hack for other designs. Slic3r on the other hand has always be it's own independent project...now I have switched to using the Prusa edition of Slic3r which has added some very nice improvements - and some stuff designed just for the Prusa printers...but nothing Prusa has done has made it any less effective on other printers. And the way Slic3r puts properties on filament vs. print vs. printer really makes a LOT more sense to me than how Cura does.
But I'm always willing to try new and different things so I'll give Cura a workout for a week and see if it wins me over at all
