My custom Cartesian

Phaedrux · 6 Sept 2018, 14:55

The time estimates may only be valid for actual Prusa printers running Marlin as they use the Marlin motion planner. The two speed values are for the MK2 and MK3 which have high power and silent mode. The silent mode speeds were just added in order to make their speeds more accurate.

The slicer defaults are probably 6 years old now and come from original slic3r. The Prusa profiles are a bit fresher take on it.

wilriker · 9 Jun 2018, 15:03

@whosrdaddy I only looked at the source code part where they would be overwritten if the flavor is Marlin. That you can find at
xs/src/libslic3r/GCode.cpp:489
the defaults seem to be either in
xs/src/libslic3r/PrintConfig.cpp:961
or
xs/src/libslic3r/GCodeTimeEstimator.cpp:19

EDIT: since I have my acceleration very low (not tuned yet, still using the values from Marlin from times before Duet) and Slic3r has much higher defaults it will estimate print times considerably shorter.

Phaedrux · 6 Sept 2018, 15:56

Just curious, but how long after installing slic3r before you were into the source code? Was that before or after you did a test print?

OBELIKS · 6 Sept 2018, 16:03

I'm betting it was before.

wilriker · 6 Sept 2018, 17:48

@phaedrux Man, what do you think of me? Of course it was before the first print. But I have to admit it was after slicing and uploading the generated Gcode to my Duet.

But seriously, I would not have done/needed that if there would have been any clue on what these two different values for every speed-related property in the .ini file meant.

Phaedrux · 6 Sept 2018, 17:52

What gcode flavour do you have configured in the printer settings? Marlin or reprap?

wilriker · 6 Sept 2018, 18:02

@phaedrux RepRap of course(?).

Phaedrux · 6 Sept 2018, 18:33

It's starting to seem like the time estimates are only remotely accurate if using Marlin, otherwise it reverts to using the old way of estimating, which doesn't take into account any acceleration values.

wilriker · 6 Sept 2018, 18:47

@phaedrux As I said earlier if it is not Marlin it will consider jerk, acceleration and max feedrate values but they are not configurable but only the hardcoded defaults.

Phaedrux · 6 Sept 2018, 19:17

I suppose that's one advantage Cura has, since you can specify all speed settings and the path planner takes everything into account for very accurate time estimates.

wilriker · 6 Sept 2018, 19:27

@phaedrux Yeah, in Cura my estimates are usually within single digits of what Duet gives as simulation.

Funnily I found a comment in Slic3r's time estimator that it is partially inspired by Cura's time estimator.

I am not in the mood of doing yet another patch-your-slicer approach (as I currently do with Cura until the next release when the Bugfix I require will finally be included) just to get those estimates right (just wait a few days and see me change my mind about that - or implement a UI part to set these values ).

Phaedrux · 6 Sept 2018, 19:32

@wilriker Well if print time estimates are your biggest printing issue, you're definitely ahead of the game.

Prusa has been putting a lot of work into Slic3r, and it sounds like the near future will incorporate the modern interface of the Prusa Control basic slicer with the back end slicing engine of Slic3r. So it won't look like it's from the 90's anymore.

It's too bad they are so heavily focusing on Marlin to the exclusion of others. I understand that they are highly integrated with their own version of Marlin firmware, but it would be nice if the gcode flavour switch actually allowed their features to work with other flavours.

At any rate, Slic3r does produce good gcode and the interface is definitely a lot faster than Cura for me at least. It lacks some features and has some others. But generally if I just want to quickly slice something and print it, Slic3r is my go to.

dc42 · 7 Sept 2018, 07:54

If you want an accurate estimate of print time before printing, try right clicking on the file in the DWC GCode Files list and click Simulate.

wilriker · 9 Jul 2018, 08:32

All joking and sarcasm aside:

Me wanting to have accurate print time estimates already in the slicer (no matter which one) is just me being pedantic and not important. It happens quite often that I tune slicer settings to get a print being faster without compromising the requirements for the printed part. But once I get a feeling of how much Slic3r is off it will be just a relative offset I will get used to.

What I meant about the patch-your-slicer approach and not being in the mood for it was maybe poorly worded. It was not supposed to mean that I will not try out Slic3r just because it cannot give me the same level of print time estimation accuracy that I am used to from Cura. It just meant that I could fix this by patching Slic3r but at least for now I won't.

I really do like the speed of Slic3r. It is vastly faster than Cura both in terms of UI and slicing speed. Especially the latter.

Also I don't really care for the look of the UI as long as it gets the job done. I am a minimalist using a tiling window manager on Linux - no fancy animations anywhere, everything full screen all the time.
I can say that I like the look of the UIs of Cura or ideaMaker but it is just not important to me in the first place.

There are two things I miss in Slic3r so far (but still no reason to not try it for some time):

I like to see the coordinates of an object on the build plate. It helps me to arrange for a new print if the previous object is still on the build plate (printing from remote) but I can guess my way around that at least
I would prefer to see the print time estimates without having to save the Gcode first

EDIT: Probably later today I will print my Slic3r sliced Benchy.

wilriker · 7 Sept 2018, 09:21

Also I just now ordered 4x Wantai 42BYGHW811 steppers.

1.8°, 47Ncm, 2.5A, 1.25 Ohm, 1.8mH

I had a very hard time to decide between this model and the 0.9° version because I primarily wanted to have the more silent version but could not get any information on how these two compare in this respect.
In the end it was that I could get the 0.9° steppers only as a set of 5 whereas I only need 4 motors (X, Y, 2xZ) and the 1.8° I could get in arbitrary amounts.

OBELIKS · 7 Sept 2018, 10:41

May I ask, why not on extruder?

wilriker · 7 Sept 2018, 11:03

@obeliks These motors are awfully heavy (370g) and I am using a direct extruder.
Also as soon as I switch to my Titan (also in direct extruder setup) they also would be totally overpowered when considering the 3:1 gearing.

I will keep the Anet stepper on the extruder for now and see how far I get with the (rather weak) 12Ncm pancake stepper that was part of the Titan clone I bought. If this turns out to be insufficient I will order something with around 20Ncm. This will definitely have enough power but weigh a lot less.

wilriker · 9 Jul 2018, 11:42

Slic3r'd Benchy printed. From what I can tell over the webcam it looks quite good but has awful image quality:

A better image and an evaluation of the quality will follow tonight.

OBELIKS · 7 Sept 2018, 12:44

@wilriker Ah,OK. Makes sense.
Check the Titan, especially if the idler bearing, teeth that drive filament and guide holes align. Some clones have this off, and you need to get a new hobbed gear.

wilriker · 7 Sept 2018, 13:14

@obeliks I got mine from Trianglelabs. They supposedly make better quality clones. But I will check this nevertheless. Thanks!