MMU2 - Real Multi Material  

RE: MMU2 - Real Multi Material

Looks like your gcode is different to what original PS is generating. That's why you're getting different results. Which PS version are you using?

RE: MMU2 - Real Multi Material

I am using 2.1.0+win64-201909160915

The previous files uploaded I only added in the gcode to allow lower extrusion temps into the printer profile.

Stock MMU2s prnter profile, stock prusament PLA filaments in this gcode example. Only change is the print settings which have your PP script.

Attached are the original output gcode and the adjusted gcode running your script which I changed temp_diff to 30 & wait_for_stabilize = True

5th tool change including temp drop:

; CP TOOLCHANGE START
; toolchange #5
; material : PLA -> PLA
;--------------------
M220 B
M220 S100
; CP TOOLCHANGE UNLOAD
M104 S185 ; MMUGCODEPARSER v0.2
G1 X170.750 Y125.640
M900 K0
G1 F1339

6th tool change for some reason excluding temp drop:

; CP TOOLCHANGE START
; toolchange #6
; material : PLA -> PLA
;--------------------
M220 B
M220 S100
; CP TOOLCHANGE UNLOAD
G1 X170.750 Y140.700
M900 K0
G1 F1339

RE: MMU2 - Real Multi Material

@nikolai-r

A very interesting concept!

How can i use your python script (not mmuGcodeParser.exe) in W7 x64. Which python evironment for Windows (if supported) i should install?

Thanks!

RE: MMU2 - Real Multi Material

@karl-herbert

Any python3 environment.

@duncan-s

Your original gcode and gcode_adjusted doesn't match. The script doesn't change that much :). It would be easier if you could provide your PS project file.

Original:

;--------------------
; CP TOOLCHANGE START
; toolchange #5
; material : PLA -> PLA
;--------------------
M220 B
M220 S100
; CP TOOLCHANGE UNLOAD

Adjusted:

;--------------------
; CP TOOLCHANGE START
; toolchange #5
; material : PVA -> PET
;--------------------
M220 B
M220 S100
; CP TOOLCHANGE UNLOAD

RE: MMU2 - Real Multi Material

Excellent work @nikolai-r. Thank you. The logic makes a lot of sense and worked on a PLA+PETG test print for me, which otherwise was impossible to print. It's very unfortunate Prusa (who claimed the MMU2S was truly multi material, which it was not without these kinds of mods or actual hardware mods I've seen done) does not recognize their own shortcomings with PrusaSlicer and incorporate something like this in there.

My next test, like above, is to use PVA supports. I imagine that will be a challenge all on its own.

RE: MMU2 - Real Multi Material

@nimaim

Thanks for the kind words. I don't blame PR anymore. They are a small company and it's just a matter of resource and business planning. If you look what they have done in the previous years and what's coming up in 2020 then it's clear to me that they're focusing on growth. Making something new/fancy to a competitive price. Let the community polish the product and reinvest the saved time/money in new development.

So the new features and new hardware development will be always preferred. The improvements are usually coming with the new bigger release. And you will have an upgrade path from the old hardware, if you want to.

For me it's still a good deal and OpenSource mentality of Mr. Prusa helps a lot. I'm running custom Prusa firmware (slightly modified) with a post processing script. And this solution is working fine for me to an unbeatable price tag. Honestly, without MMU2 option I wouldn't be able to print multi material at all. Because all good dual head printer are not in my budget.

That being said, as long as PR keeps everything open and don't punish the community too much (like with the firmware lock/breakout in the mini), I still prefer MMU2 over Palette2+. Even if this solution is not ideal. Maybe they will introduce PrusaSlicer Plugin interface and such things could be easily be implemented as plugin (.... I know, ... dreaming should be allowed ... )

RE: MMU2 - Real Multi Material

Posted by: @nikolai-r

Since PR developers still don't consider to merge the provided tiny pull request into main branch, I still have to recompile it every time.

In case someone have similar issues during filament unload, here is a compiled firmware version with this patch applied. It helps a lot in case you don't have perfect filament tips (which is very common by mixing different materials).

 

Attachment removed

Well said above ... although I'm not sure I'd consider Prusa "small" anymore with the huge success they've had in 2019: https://blog.prusaprinters.org/2019-recap-plans-for-2020-and-original-prusa-mini-update/ ... ~20000% revenue growth. Yes they may not be at the $5-10K price bracket of say an Ultimaker (or LulzBot, which was just acquired, thank god I didn't go with them), but it's still the best prosumer printer you can get for in-home without breaking the bank. Their support is also unmatched.

As to what I quoted, I saw the commits for this patch that affects filament unloading time. Have you seen better tips from all filament types or just PLA (which is what most people do multi material prints with)? I seem to get horrible stringing when unloading any PETG that's not as perfect as Prusament (~0.01mm tolerance).

RE: MMU2 - Real Multi Material

Posted by: @nimaim

Have you seen better tips from all filament types or just PLA (which is what most people do multi material prints with)? I seem to get horrible stringing when unloading any PETG that's not as perfect as Prusament (~0.01mm tolerance).

The firmware patch doesn't affect the filament tip. I have an issue on my printer where the extruder pulls on the filament (sometimes) before the MMU is able to pull out the filament out of the extruder. Causing the Unload error more often. With the patch it went away and I don't have this issue anymore even with not perfect filament tip.

RE: MMU2 - Real Multi Material

Ahh ok makes sense. Thanks for the clarification! I will test this as well.

RE: MMU2 - Real Multi Material

Just an update ... 

I tried a quick calibration test with this neat source ( https://www.thingiverse.com/thing:3282265 ) using 4 filaments. Colors 1/2 are PLA, 3/4 are PETG. There's about 75 tool changes and good way to also test adhesion between the different materials with each layer.

Without the MMU2 postprocessing script, I got horrible stringing and therefore no finished print as it constantly required manual intervention. With the script, stringing was basically gone and extrusion was a lot more predictable, as old material purged properly. The only minor issue is the purge tower was not as "clean"; sometimes I get blobs here and there, sometimes there will be some missing layers. Luckily, the actual object printed just fine. I will have to do some more testing with temps / filaments / etc.

Mind you, I've also done countless other mods to the MMU2S including 2.5mm ID tubing, selector with magnet and open FINDA canal for debris to fall out, MMU2 PTFE holder M10 passthrough, M10 passthrough plates at the selector output and extruder input, etc. I will post links to all these if others are interested but it probably warrants a new thread.

As I didn't want to test too many new things at the same time, I did not test the modified firmware posted here. My filament unloads were just fine, but I've only tried with this short test print thus far. For now, I finally have a fairly reliable set up now and can mix these different filament types. Thanks @nikolai-r.

RE: MMU2 - Real Multi Material

Also thought I'd mention ... my extruder tension screw is threaded just barely on (only one full turn into the spring). After all my MMU2S upgrades, this ended up being my main issue. Any tighter and any strings left over from previous unloaded filament being fed in would cause sensor to trip and printer to eject the filament. This constantly happened to me but YMMV since the manual says to tighten the screw until its flush with the nut (did not work for me whatsoever and had to constantly intervene by cutting tip of filament and reloading). With this "loose" set up though you can also print TPU without readjusting again, a huge plus.

So those who FINALLY have your MMU2S adjusted properly, work your way over to the extruder: make sure IR sensor in chimney is reliable (you can take the tube out, manually feed in filament, and make sure it trips with loading/unloading) and extruder screw is tensioned properly. Good luck, this thing is quite the journey.

RE: MMU2 - Real Multi Material

Hello,

I worked on the software side, by creating a special version of PrusaSlicer (code name "Dribbling") that avoids strings unloading the filament and creates a perfect shape tip.

It works well, but I noticed that when the filament is not ejected by the slicer gcode (e.g. unload filament from LCD, or First Layer calibration) the filament has an high probability to jam in the filament eject. For this reason I decided that when I have some time, I will "patch/improve" the Prusa Firmware with the same technique I use in my PrusaSlicer custom version so that also ejecting by LCD panel will work well.

Regards

RE: MMU2 - Real Multi Material

Posted by: @antimix

Hello,

I worked on the software side, by creating a special version of PrusaSlicer (code name "Dribbling") that avoids strings unloading the filament and creates a perfect shape tip.

It works well, but I noticed that when the filament is not ejected by the slicer gcode (e.g. unload filament from LCD, or First Layer calibration) the filament has an high probability to jam in the filament eject. For this reason I decided that when I have some time, I will "patch/improve" the Prusa Firmware with the same technique I use in my PrusaSlicer custom version so that also ejecting by LCD panel will work well.

Regards

Why not just write a simple gcode script with the same commands rather than patching it into the firmware? That way it's portable and not tied to a particular build/version. Run the script instead of using the menu commands.

RE: MMU2 - Real Multi Material

Posted by: @vintagepc

Posted by: @antimix

Hello,

I worked on the software side, by creating a special version of PrusaSlicer (code name "Dribbling") that avoids strings unloading the filament and creates a perfect shape tip.

It works well, but I noticed that when the filament is not ejected by the slicer gcode (e.g. unload filament from LCD, or First Layer calibration) the filament has an high probability to jam in the filament eject. For this reason I decided that when I have some time, I will "patch/improve" the Prusa Firmware with the same technique I use in my PrusaSlicer custom version so that also ejecting by LCD panel will work well.

Regards

Why not just write a simple gcode script with the same commands rather than patching it into the firmware? That way it's portable and not tied to a particular build/version. Run the script instead of using the menu commands.

Agree, this would be easier and allow users to easily modify to their liking on the fly. Prusa Slicer makes it painless to run multiple pp scripts.

RE: MMU2 - Real Multi Material

I did get warping on my print printing PETG and PLA. Is the best way around this to tweak all filament bed temps to the max temp you'll be printing with? I'm guessing it didnt like going from 80C for PETG to 60C for PLA although extruder temps behaved correctly with the script. Does the script manipulate bed temps at all?

RE: MMU2 - Real Multi Material

I am interested in trying this script for mixing PLA and PETG.

But first I am wondering if I can just use it to improve multicolor PETG prints?

The default settings use 230 degrees for PETG to improve filament tips, even though generally PETG prints better at higher temperatures.   (It is ridiculous, isn't it, that the entire print is done at a lower temperature just for the sake of the tips???)

So what I want is to print at say, 240, but use 230 for filament changes.  I don't want to change anything else; the existing MMU2S behavior (at the default, lower temperatures) works well for me.  Is your script easily adaptable for this?

RE: MMU2 - Real Multi Material

@wpegden: It's actually doing it default. Between filament changes it drops -10C. Just make sure different colors have slightly different temperatures (for example 200, 201).  Otherwise it's not changing anything. You can adjust it to whatever you want. I'm printing PETG at 250C without bigger issues. But it always depends on the filament used.

 ram_temp_diff = 10

@nimaim: Bed temperature is left untouched. In case of PLA+PETG I would set for both 70C in the settings. 

RE: MMU2 - Real Multi Material

Posted by: @antimix

Hello,

I worked on the software side, by creating a special version of PrusaSlicer (code name "Dribbling") that avoids strings unloading the filament and creates a perfect shape tip.

It works well, but I noticed that when the filament is not ejected by the slicer gcode (e.g. unload filament from LCD, or First Layer calibration) the filament has an high probability to jam in the filament eject. For this reason I decided that when I have some time, I will "patch/improve" the Prusa Firmware with the same technique I use in my PrusaSlicer custom version so that also ejecting by LCD panel will work well.

Regards

I saw this as well as Skinnydip. Is this an alternative to this script? Enhancement? Can you use both together? Does it perform better with stringy materials such as PETG?

Diff temp PLA materials work 100% with this script as well as some ABS I tested (though bed temp needs to be hot to get it to stick well and not warp the rest of the print).

PETG on the other hand is difficult to get dialed in properly. It seems there is always some stringiness present no matter what is adjusted in the slicer (temps, cooling moves, etc.). Even with Prusament PETG, which is essentially perfect (~0.01mm tolerance), I get it.

RE: MMU2 - Real Multi Material

@nimaim

I made the same experience with PETG. The good thing is it's causing issues maybe something around 200-500 tool changes. You can also make it less an issue if you choose the inputs further away (i.e. 1-5, 1-3-5). Of course only if you're not printing all the 5 colors at once.

RE: MMU2 - Real Multi Material

Posted by: @nikolai-r

@nimaim

I made the same experience with PETG. The good thing is it's causing issues maybe something around 200-500 tool changes. You can also make it less an issue if you choose the inputs further away (i.e. 1-5, 1-3-5). Of course only if you're not printing all the 5 colors at once.

Interesting approach. That could potentially lessen the issue. I think I'm in a good spot for true multi material with this script though. My next test is mixing TPU + PLA and then PVA / BVOH + PLA. Going to first try an all night PETG + PLA print with 200-300+ tool changes and see how she goes. It's unfortunate PETG is so stringy because it's my go to material and gives you the best of PLA (in terms of easy printing, at least in single filament mode) and ABS (in terms of strength).