0.05mm layer height is possible and I'm sharing
I've seen claims of MMUs producing layer height and using nozzle configurations not shipped with PrusaSlicer, but they never seem to share their configs and just provide vague answers about how to do it. So as well as showing off an uninspiring print, I'm sharing my custom print profile as well. For Linux (and I assume Mac) users, download the INI file and put it in your ~/.PrusaSlicer/print, then restart PrusaSlicer and it should be available when you select the MMU2 printer profile. I think there is more to work on here as there are two areas of concern, but it worked.
The first issue is the size of the wipe tower. I actually had to drop the purge values down to 50 across the board from 70 just to get the tower to fit on the plate! Still it was massive and seemed to wipe far more than needed or expected. On the first layer, for example, the wipe when switching to the 5th filament went on for half the wipe tower. I didn't see any other tool changes waste that much filament, but they still seemed more than needed.
The second issue was the pre-print purge. Rather than a single double pass you can see both in PS as well as the actual print that it ran a purge of each filament. I actually kinda like that approach, but it was unexpected.
The print was not flawless, but out of 308 tool changes I only had to help 6. One was because one of the filaments decided to break between the MMU and extruder and the reset were due to tool 1 leaving hairs behind that would eventually accumulate and jam the FINDA. Still, a 2% failure rate on an untested profile isn't bad!
Here is what PS looked like:
Post print images:
And a side by side comparison of the finished product and one with more rows printed with a 0.2mm layer height. Each color is 5x5x20mm and has a 15% gyroid infill. All 5 filaments are PLA (from L2R: Inland PLA+ Black, Amazon Green, Inland Natural, Inland Light Brown, and Superfila Brown). Other than the previously mentioned purge values there were no other tweaks:
RE: 0.05mm layer height is possible and I'm sharing
Hmmm... I just noticed that the printer says the run time was about 8 hours. Not sure what is going on there as it gave the same estimate as PrusaSlicer and when you took out the time for those failures PS had it right.
Also not sure why this print got the elephant foot and the other (printed Sat) didn't. It was in a different spot so it could be a bed leveling issue (though it was dead on when I rebuilt the Y bed a month ago).
RE: 0.05mm layer height is possible and I'm sharing
Nice work, seems like a purge bucket/chute mod might be useful at this point when this much build space is getting sucked up by the purge block... :-/
RE: 0.05mm layer height is possible and I'm sharing
Nice work, seems like a purge bucket/chute mod might be useful at this point when this much build space is getting sucked up by the purge block... :-/
Yeah I hope there is something I've missed that can bring that monster down a notch (or 10), but this is also an aggressive print meant to test lots of tool changes in a short span.
I have heard that Prusa is also working on multiple purge blocks for a future PS release. That would let you break a monster like this down and better use your space.
The purge bucket might not be a bad idea though. Have to look at the gcode to see if the tower ops is easily identifiable. If so some post processing could play some games. With our printers, however, you have to give up part of your build space since you can't move the extruder a significant amount outside the area.
RE: 0.05mm layer height is possible and I'm sharing
Yeah - best of luck but you may be stuck - you need to purge volume so if you drop the height of your layer by 1/4, one of the other dimensions is going to have to grow by the inverse...
Either way, losing maybe 5mm of build space on the left or right is a reasonable compromise; I don't think you'd actually need to sacrifice much since I'd envision something that attaches to either of the X ends so you move the nozzle all the way to the right or left. It doesn't have to be a bucket, for space efficiency a chute might be best to just guide the purge stream out to the side. The other upside here is you could potentially purge at a higher rate than on a tower to aid getting the old colour out and have less mixing.
I have some dev/modelling skills that might be useful in this area but based on your activities here it sounds like you do too. I don't have an MMU (yet - thinking of getting one in a few months) but I'm interested in helping develop the purge chute idea further for when I do... let me know if I can help.
RE: 0.05mm layer height is possible and I'm sharing
To some degree I agree on the volume, but this one is easily 5x the footprint of the one for the 20x20x20 example even though I reduced the volume by around a third. Watching the tower in progress, it continued purging long after the color had changed. If that is simply a factor of the purge volume I estimate I could drop those values to the 20-30 range and still not see any bleed in the final product.
On the bucket, yes attaching it to the X axis would be the most efficient for extra movements. Ideally not only could you save build space, but you could also strip out "purge" operations where no tool change is happening...
Damn you for giving me more projects to think about!
RE: 0.05mm layer height is possible and I'm sharing
well, in all fairness I'd been thinking about it too; far easier to make progress with two minds than one 🙂
RE: 0.05mm layer height is possible and I'm sharing
I just realized the gcode implementation here may not be too tricky; There's a user contribution for doing MMU activities without an MMU (e.g. manual filament changes). Seems that could be fairly easy to adapt to these purposes with a slight mod to replace M600 with a "move to purge bucket", MMU tool change & prime, and then resume.
RE: 0.05mm layer height is possible and I'm sharing
Yeah there was recently a bit of discussion about how to use the MMU to "ColorPrint" by using it's filaments rather than still forcing filament changes in a single tool. Definitely some similarities that could be shared between the two needs.
RE: 0.05mm layer height is possible and I'm sharing
Hmmm... Just took a look at my printer while waiting for an 88GB file to copy.
Prusa might have already been thinking along these lines to some degree.
At least from the settings menu you can move the X axis to 255 which has the nozzle outside the print area and actually right at (if not off) the edge. If that can be replicated with gcode it could reduce (remove?) any lost space from the print area. The trick, however, is that at Z0 & X255 there is no space for a bucket to fit in there.
The story is similar with X0, though that is at the edge of the build area so you by definition have to loose space there.
I wonder if the solution might be a folding design. At Z heights less than a TBD value (say 10mm for discussion sake) it would simply look like a normal "box". When Z is raised beyond that, however, the side near the bed would slide out to where the nozzle could reach it and deposit the purge. As the Z was lowered it would force it to collapse again as it touches the bed. So the idea would be for purges below that critical Z value the process would have to raise the Z height enough for the chute to open, perform the purge, then lower again to resume the print.
If that worked you could actually end up losing zero bed space.
The main concern I can see is how is having something constantly dragging on the bed early in prints going to impact Y movement and things like ghosting. Also how is the chute going to hold up to such constant abrasion.
RE: 0.05mm layer height is possible and I'm sharing
Well, G1 X255 should do it- if the knob allows it then so should the gcode. Note that the purge line is technically outside the print area, at Y=-3 or somesuch. It should be possible to move to e.g. X=-3 (or maybe X=257) as well (or whatever the X value is when it homes on the left side. Dunno whether that puts the nozzle off the build surface or not.
OTOH.. if we're willing to take the plunge on firmware mods, one could have a little solenoid actuator driven by one of the PWM pins from the X or Y unused limit switches on the einsy... add a new M command that moves Z up if too close to the bed, over the chute, pop it over, purge, and resume. Ideas and options abound, I'd wager we could make it happen. I don't doubt this would be a high value mod.
RE: 0.05mm layer height is possible and I'm sharing
Well, G1 X255 should do it- if the knob allows it then so should the gcode. Note that the purge line is technically outside the print area, at Y=-3 or somesuch. It should be possible to move to e.g. X=-3 (or maybe X=257) as well (or whatever the X value is when it homes on the left side. Dunno whether that puts the nozzle off the build surface or not.
I forgot about the Y axis going negative, but that wouldn't work as you would need a way for the bucket to raise with the Z or you are going to end up with a giant spaghetti mess as the purge misses falling into the bucket. Additionally that would be undesirable for people with enclosures as it would be a PITA to clean out as needed while a print is running.
X only goes from 0 (still in the build area) to 255 and both are touching the Z carriages, so no room to play there.
OTOH.. if we're willing to take the plunge on firmware mods, one could have a little solenoid actuator driven by one of the PWM pins from the X or Y unused limit switches on the einsy... add a new M command that moves Z up if too close to the bed, over the chute, pop it over, purge, and resume. Ideas and options abound, I'd wager we could make it happen. I don't doubt this would be a high value mod.
Yeah there is a lot that could be done with firmware mods. Heck, through in a frame mod and you could even have the extra space that is needed. I myself prefer to remain as stock as possible though so updates are easy.
RE: 0.05mm layer height is possible and I'm sharing
Yeah, wasn't suggesting we do it on the Y, just providing context for my suggestions on trying to go outside the X range.
You may be on to something though - I just had an idea for a spring-loaded mechanism that pops out the purge chute but is triggered by the extruder pushing into it on the right. Of course you'd still lose a few mm of build space as you can't print in that area and accidentally trigger it at z=0, but with a simple lever one could turn a mm or two of bump from the X carriage into quite a few more for the purposes of putting the chute under the nozzle. Then when the extruder moves away, the spring causes it to retract again. I suspect it could be done with either a bolt-on of some kind or a modified X end. Let me mull on it some more and maybe put something together as a concept design.
RE: 0.05mm layer height is possible and I'm sharing
Actually you could have the lever catch the PINDA mount or fan as they cross the plane well before the nozzle does. That's actually an excellent idea as it removes the concerns of my idea about rubbing against the Y bed.
RE: 0.05mm layer height is possible and I'm sharing
Good point, there's any number of bits that protrude over the edge of the bed before the nozzle gets close. PINDA is probably not the best for fear of shifting the Z value over time.
RE: 0.05mm layer height is possible and I'm sharing
Especially with it's propensity for the mount to soften and move with the R4 parts...
RE: 0.05mm layer height is possible and I'm sharing
So I got curious and just generated 2 minimal prints. One using the Single profile and the second using two colors. Both are 2 1x1x0.2mm squares with skirt loops and height both set to zero. Here is what I found for the purge and change process:
;--------------------
; CP TOOLCHANGE START
; toolchange #1
; material : PLA
;--------------------
M220 B
M220 S100
; CP TOOLCHANGE UNLOAD
G1 X229.250 Y134.920
G1 F1341
G1 X223.671 E0.2816
G1 F1452
G1 X217.626 E0.3051
G1 F1644
G1 X210.785 E0.3453
G1 F1899
G1 X202.881 E0.3990
G1 F2266
G1 X193.449 E0.4761
G1 F2745
G1 X182.025 E0.5767
G1 F3224
G1 X170.750 E0.5691
G1 Y134.140 F7200
G1 F3224
G1 X172.892 E0.1081
G1 F3639
G1 X188.036 E0.7644
G1 F3926
G1 X204.375 E0.8248
G1 F4038
G1 X221.180 E0.8483
G1 E-15.0000 F6000
G1 E-24.5000 F1200
G1 E-7.0000 F600
G1 E-3.5000 F360
M104 S215
G1 Y133.360
G1 X170.750 E20.0000 F488
G1 X221.180 E-20.0000 F326
M73 Q41 S2
M73 P41 R2
G1 E-35.0000 F2000
G1 Y133.500 F2400
G4 S0
T1
G4 S0
; CP TOOLCHANGE LOAD
G1 E12.0000 F1140
G1 X170.625 E42.0000 F1315
G1 X221.180 E6.0000 F713
G1 Y134.000
; CP TOOLCHANGE WIPE
G1 X229.000 E0.3507 F800
G1 X229.750 Y133.500 E0.0404
G1 X171.000 E2.6348 F900
G1 X170.250 Y133.000 E0.0404
G1 X229.000 E2.6348 F1100
G1 Y132.500 E0.0224
G1 X171.000 E2.6012 F2100
G1 Y132.000 E0.0224
G1 X229.000 E2.6012 F2125
G1 X229.750 Y131.500 E0.0404
G1 X171.000 E2.6348 F2150
G1 X170.250 Y131.000 E0.0404
G1 X229.000 E2.6348 F2175
G1 Y130.500 E0.0224
G1 X171.000 E2.6012 F2200
G1 Y130.000 E0.0224
G1 X229.000 E2.6012 F2225
G1 X229.750 Y129.500 E0.0404
G1 X171.000 E2.6348 F2250
G1 X170.250 Y129.000 E0.0404
G1 X229.000 E2.6348 F2275
G1 Y128.500 E0.0224
G1 X171.000 E2.6012 F2300
G1 Y128.000 E0.0224
G1 X229.000 E2.6012 F2325
G1 X229.750 Y127.500 E0.0404
G1 X171.000 E2.6348 F2350
M73 Q72 S1
M73 P72 R1
G1 X170.250 Y127.000 E0.0404
G1 X229.000 E2.6348 F2375
G1 Y126.500 E0.0224
G1 X171.000 E2.6012 F2400
G1 Y126.000 E0.0224
G1 X229.000 E2.6012
G1 X229.750 Y125.500 E0.0404
G1 X171.000 E2.6348
G1 X170.000
G1 Y125.000
G1 X230.000 E2.2804
G1 Y135.560 E0.4013
G1 X170.000 E2.2804
G1 Y125.000 E0.4013
G1 X230.000
M220 R
G1 F6000
G4 S0
G92 E0
; CP TOOLCHANGE END
;------------------
There are also a bunch of other differences, but this appears to be the important part to identify, strip, and replace with bucket code.
RE: 0.05mm layer height is possible and I'm sharing
yeah, that's pretty straightforward to handle in post with a bit of regex, I'd bet.
I drafted something together for a pivot mechanism to swing out the chute (bottom of "scoop" is open - trying to keep it small ATM with the intent to guide the purge into a separate container that is attached to the Y extrusion, for example) that shouldn't interfere with anything; still need to ponder on how to get it to pivot out. No doubt it could be refined and built into the X end but at this stage I'd prefer not to have to disassemble my X or have to print huge honking parts until we get a solid design and prototype.
RE: 0.05mm layer height is possible and I'm sharing
(and yes, I realize a part like that would be a bear to print in one shot; this is a *very* early draft and all I have time for today)
RE: 0.05mm layer height is possible and I'm sharing
Yeah I agree with not modding the X carriage. I was thinking along similar lines. My thought was that it would use the smooth rod for extra stability and use the existing bolt setup (with longer bolts) on top of the carriage. My thought was to build it so that you only need to remove the rod guide at the top and slide the bucket holder on.