PRUSA MMU3 - A Robot That Changes Filament. An In Depth Look At How It Works.
Вставка
- Опубліковано 25 вер 2024
- Video sponsored by PCBWay - www.pcbway.com - PCB Manufacturing, 3d Printing, CNC parts, and more...l
Also get $5 of credit if you sign up to PCBWay using this link www.pcbway.com...
Check out the Prusa MMU3: www.prusa3d.co...
Prusa XL: www.prusa3d.co...
General Link to Prusa: www.prusa3d.com...
Today we take a look at PRUSA's MMU3 which has just been released for the MK4 printer (it's been out a while for the MK3s but there's some changes for the MK4). Let's see how good it is!
Links to models used: www.printables...
Fair Warning: As always with large brands I'm expecting some obnoxious comments from some of you who seem to think 3d printer brand tribalism is somehow worth your mental energy. Nobody wants to hear it, and you'll be banned from commenting / strict auto moderation is on, so keep it civil and polite if you want your comment to post :)
If you're interested - I bought the MMU3 myself without Prusa's involvement or any discussion with them aside from casual mention, but they did provide the MK4 previously to me at no cost. They are not sponsoring this video.
Would be interesting to see a similar treatment of the Enraged Rabbit Carrot Feeder because it is in theory compatible with any printer running on Klipper and can expand to more filaments than the MMU. Plus it would also require assembly and could be an interesting comparison to the MMU.
Noted. I've wanted to make one for a while
or the tradrack, which a lot of people say is cheaper/simpler to build. annex engineering is also known for good design.
@@LostInTech3D Do Tradrack AND ERCF, and compare the two. A lot of people taking sides depending on where you hang out. ERCF is getting close to getting yet another update to a new revision which should improve things further. I built the new V2 from a Siboor certified kit and it's been trucking along really quite well. I was sure I'd run into some kind of problem, but the thing works very very well. There's a wealth of mods for it, which I absolutely believe is one big benefit of the ERCF over the Tradrack. That, and the large community behind it. It's easy to get help.
@@nvrprfct9176can only second the tradrack, sooo much easier to build and significantly cheaper to self source. With the encoder mod id say beats the ercf in every category except loading flexibles as its a single drive gear system rather than dual gear
Also interesting since the prusa MMU seems conceptually very similar to the ERCF. (though obviously they are different in implementation)
As far as desk space goes: That's just for the suggested setup. You can of course use your own spool holders or whatever you want to hold to spools, and there are alternative designs for the buffer on Printables, and you can even set up your own system to take care of slack. I believe you can even find designs for spool holders that can rewind a bit, negating the need for a buffer at all.
Or use suitable size ikea furniture to house the buffers in it and put the spools (or dryer box) on top of that, shortening the first ptfe tubes a lot?
Yeah, but the fact that this is the setup that Prusa recommends, whereas their competitor kept everything neat and out of the way, make the sprawl/DIY solution a tough sell.
That's imo one big benefit over the bambulab system or other systems like it (Don't want to hate against bambulab etc.). You're more or less free to choose what ever spool system, size or holder you use. You want a 100g spool? feel free. You want a 10kg spool? If the mmu's gonna pull hard enough I think it would be possible.
I use the original buffer and also printed one auto rewind spool holder for testing, both works without a problem.
Note for materials used for the MMU3:
* The printed parts that you are thinking are CF-PETG are probably CF-PC; that is what they are recommending if you want to reprint them. Source: MMU3 printable parts page on Printables.
* The main plate is printed using MJF, not SLS. Don't know the material, but could be Nylon as you suggested. Source: Assembly guide for the MMU3.
I wish I could post links, but UA-cam likes to delete posts with links in them.
I can approve links manually if I know they're coming, go ahead
@@LostInTech3D ok, just sent the reply
Sls and mjf both use nylon pa12 and tpu, mjf is just a different method of fusing the powder
I use no buffer at all and my 2 MMUs are printing almost non stop. Just have 10cm ptfe tubes out on the back so filaments don’t fall off. Works perfectly and no more hassle with that darn buffer and the tubes…
Are you happy with how it’s working so far ?
My MMU three is pretty reliable. I do 60 hour prints two or three times a week with five colors and go a week or two without any errors over the last two months none of the errors have resulted in loss of a print.
Good info
Nice. Are you printing on a Mk4 printer? Prusa had some trouble to get the MMU3 to work on that model, but it seems they solved whatever the problem was.
@@kaasmeester5903 mk3+
@@kaasmeester5903 sorry MK3S+
@@kaasmeester5903 I have been been considering 3.9 upgrade but it is working so well I am afraid to.
Remember that the Voron community also has a MMU like system, the Enraged Rabbit Carrot Feeder. Still not 'officially' out, but there are many mods and since Bambu introduced the filament cutter, it's also available as a mod in the ERCF V2. It seems like a good system, and you can make it as big as you want. Not sure how reliable it is. I know that dialing it in requires some patience... It may be like the MMU2, where some people had 0 or little problems and others have nothing but trouble. It may be the self building part of it, where you can never know if the printed parts are of good quality and/or the building process was not as accurate as it should be.
I must say that I have had little problems with my AMS. The only time that it gave me headache was with a batch of filament that was brittle like dried spaghetti. That was the first time I had filament breaking off on me. And sometimes I have tangled filament, but I can't blame the AMS for that. I sometimes let go of the filament when swapping in the AMS and then it unwinds a bit. This may cause tangled filament later on.
Yeah I'd like to give the ERCF an episode like this at some point
Exactly zero of the "kids these days" caught the "series of tubes" reference.
I did. I am old.
Same here 😅
kill-any-of-ems' (sorry MILLENNIALS, I have been corrected) suspected that I was window lickin batshit crazy when I tried explaining how everybody had to manually TAPE the songs we liked off a HiFi separates stack from the radio and it was pretty regularly followed by like a second of ADVERTS, too....
genuinely don't know they're born...
@@su3Do Wait, you guys didn't have radios with built-in tape decks? I had one, it was also good for tape duplication as you could insert two cassettes in.
YOU-tube? BOOB-tube? TUBE-acca (say it out loud star wars fans, and I do apologise lol)?
pretty much everything tube-related, according to basically the entirety of this planets' future custodians...
reassuring.....
Is that the word....??
...CONCERNING, that's the one, sorry, my bad
That was uttered in 2006, if you think this is old...... My god brother you are young.
The Wera allen key at 1:18 caught my eye - because I've done the exact same modification to some of mine. That plastic sleeve always gets in the way! :-D
Haha yeah I finally had enough of that sleeve 🤣
I have a Bondtech MK3S+ with an MMU3* and the RMU mounted on the back left side of my printer. Faster and easier to load/unload, takes up less space, much shorter and neater tubing runs. I built the Prusa buffer as well but I definitely prefer the RMU.
Nice that they finally made some actual improvments on the MMU. The MMU2 was probably the worst customer experience I've had in my life, I don't think I ever got even one model printed before I just gave up.
I'm curious, what were the failure modes - and do you see ways that the MMU3 has dealt with them?
It's a neat idea, but pretty complex in implementation, a niche DIY group within a niche, so I can see big challenges getting the data needed to really optimise, looks like they did an awful lot of tuning in house before releasing MMU3.
If you're interested in looking at more filament switcher addons, there's an interesting one called Tradrack by Annex Engineering that I'd love to see an in-depth look at. It's open-source like the ERCF, but with a design philosophy centered around making it easy to add more colors; even the small builds are eight or nine colors, and it can supposedly go quite large. It's pretty new so I haven't seen much about it yet.
I just finished hooking mine up and calibrating it yesterday and I have had a few failure to push filament into the extruder errors. That was on my first print. On my second multi color print, it was flawless feeding.
Your videos and the girl at TADA 3D Printing are both covering the MMU3 install and setup and together your channel and hers have been very valuable. The Prusa MMU3 is quite a design and engineering marvel. It's not nearly as easy or plug n play as an AMS and AMS light, but it's very cool to see the different design.
One thing I find from my Bambu labs AMS, is 95% of the time, a multicolour system is super handy, even for printing single colours since it means not swapping filament. The other 5% is when a filament snaps off *somewhere* and dismantling the thing becomes a huge pain.
This is the major pain point I have with my AMS. MMU3 is better in this regard as any brittle filament that breaks in the PTFE merely requires you to use a non-brittle filament like PETG/ASA and push out the broken bits.
2 years into AMS, 1 snap
I have ams lite and it's super handy most if the time. No pushing filamemt through tube like he does in the video, just put the end into hole and the rest is automatic. Never had filament snap since I got it.
The only problem I see - it could have more powerful motors. Sometimes when printing matte pla (more friction) filament and when there is not much left filament on the spool (filament is coming into feeder at greater angle) motors have not enough power to feed filament and I get error that filament might be tangled.
I had massive Problems with snapping and exploding* filament on my AMS, then it suddenly stopped. For one simple Reason: I used up all my old Filament, which couldn't stand the forces of the AMS. And the room i printed in had a 60% humidity. For me the AMS is a great, but expensive, Filament managing Drybox. You can have your open Filament in it, have it dry. And i had no Snap after 300 hours of printing, because the old Filament was used up. 700 hours printing and no Snaps. *Yes i mean exploding, i had one old Filament Spool with 400g on it, after it i had two 120-140 parts and lot of broken Filament parts. Respooled the rest and could print it.
because I don't always properly store filament I've had it snap a few times. It's VERY easy in the MMU3 to fix. Most of my failures were caused by brittle filament.
I would have loved to see the assembly process. The MMU3 assembly was ok but the mk3.5 was a pain. I almost gave up a couple of times.
I have ERCF v2 + ERCT + ERF, best of both world IMO. ERF is what I wanted from ERCF (v1) but didn't commit because of tip forming issues, glad I waited.
I'm glad Bambu came along to shake things up. This time a few years ago all there was to talk about was the 9,000th ender 3 clone coming to market. Now they are actually having to innovate.
The MMU was there before Bambu existed...
@@marcusst8809 Not in wide use it wasn't.
@@marcusst8809 someone is going to insist that Bambu invented CoreXY. 🙄
@@marcusst8809 true, but since it only works on a Prusa machine, and you had to build it yourself and the MMU2 seemed to be quite unreliable (according to the stories). Bambu succeeded in making the 3D printers more accesible for the masses by offering a 3D printer with a MMU like system for an affordable price. A Prusa Mk4 kit with MMU 3 kit will cost me nearly €1200. A Bambu Lab A1 combo only costs €600, and a P1S Combo a bit over €1000. No assembling required. That's why everybody's talking about Bambu Lab. More people get into the hobby, and that's a good thing.
@@marcusst8809 The MMU was there _way_ before Bambu existed
Also, the ender 3 is a clone of the i3 design, which in turn was a heavily modified version of the mendel.
I have an AMS and I've never done a multicolour print. For me it's just a fancy way to store a few rolls of filament without having to do the manual swaps. The poop has always put me off and seeing this solution just makes me realise even more that I don't want to do multicolour printing on Bambu printers.
I get that.
Same here. I have the AMS and after trying it twice for multicolor, i dont use it for that anymore. So much waste and the time to print is ridiculous
@@anthonylong5870 You can use it strategically. For example labels on prints which result in a handful of swaps per print but otherwise the printtime isn't worth it... I do mainly functional prints and people around me with an ams don't understand why I am not "rushing" to get one. The few swaps I need to do I can do manually.
Yeah, decided that maybe its time to end PLA and use anything else. The Bambu works great with PETG with my PRUSA never really has.
@@anthonylong5870 I need multi material for TPU. cheapest is 4k.
You should look into the Annex engineering tradrack mmu. Like the ERCF it can run on any klipper printer. But it is way cheaper as it only has one set of extruder gears. I think people said like $4 per extra color vs $16.
Thanks for sharing!
I have a upgraded mk4 from mk3, and bought my MMU3 (waiting it to come from fedex)
But I was thinking I haven't made a good option, because the waste material would be the same as the Bambu Lab AMS
Now after the video, I discovered I've made the best decision, because in my country, the filaments aren't so cheap like they should be
So, thank you!
Bring more videos of the MMU if possible
Glad to help, and yes I think more videos in the future on this
There is a way to purge to the infill the model. I haven't tried (I don't have an MMU of any version) so I don't know how well it works. Might be a good topic for a follow up video to this one.
FYI: The MMU3 is compatible with 95A TPU. Prusa just doesn't endorse it. I keep the MMU around just for this reason. It wont do 100s of flawless TPU changes, but it will do 10s.
really? with or without buffer?
@@LostInTech3D I've been running it without a buffer.
Great color choice. I printed my MMU in the same color. Great minds...
One mayor change of the mmu3 + mk4 compared to the more promlematic mk3 + mmu2 is the integrated nozzle/heatbreak combo. Many mmu2 users reported much more reliability after changing the hoted to a revo. Theese integrated heatbreak/nozzle combos allow much better ramping on onload.
For TPU on an MMU you might try Vario Flex. This is a quite hard TPU, nearly as hard as PETG, before melting. It´s final shore hardness depends on nozzle temperature.
Makes sense!
The Rock on the PCB will make it STRONG !
will you be releasing the mmu3 build video?
fingers crossed
I guess given that this vid is more popular than I expected, I'll maybe try to compile it into a 2nd channel vid when I get time
the mmu design looks like it could be parameterized and expanded to many more colors. you'd need a bigger idler drum as you add more colors though
I can say I have had an issue where the roll tied it self and prevented the filament to be pulled through. It kept printing and since it thought the filament was still there.. causing my 12 hour print fail. It isn't failure of the MMU3.. but I wish it could have told me sooner that something went wrong.
You can ditch the buffer and directly feed from a set of filament driers, which I think doesn't take more space than BambuLAB's AMS other than having an extra spool.
I'm basically sold on getting a MK4 with an MMU3. I've been thinking a lot about getting a Bambu Lab printer, as I want a good and fast printer. But i've slowly realized that I would love to be able to multi color printing, but my main type of prints are single color, and I just want a super reliable printer, that is easily repairable.
And from what I can gather is the Prusa printer is simply the king here. I can also see that most print farms rock Prusa printers, as they are simply more stable.
Cannot wait to be able to switch from a Creality CR-6 SE to a MK4 :)
Most of the print farms are rocking P1P
@@seanzhouz3478 depends if you need a filter, the bambu filter isn"t enough
@@seanzhouz3478 It's a bit of a mix. Some farms add Mk4 printers with the MMU3 for reasons spelt out in this video. There are a couple of videos of print farmers doing a comparison between the mix of printers they have, on criteria that are important for them, such as speed, uptime, waste (failed prints, or poop from multicolor prints), and required maintenance. Between the Mk4 and the P1P, the latter came out ahead slightly on speed, uptime and failed prints, though for color printing the score is reversed. I think what matters most in their decision is simply the price, given that the performance of both printers is not all that far apart.
The purge tower you show looks like it's purging more filament than necessary. If the colour switching is fast and clean, can the purge tower be reduced much more? Also, what about purging into the infill?
Ok, could you show now more of the assembly process?
I would love to see those hour of assembly footage. You should have maybe uploaded unlisted, with a link here
I voted that I wanted to see how it works, but I would also would like to see the process of making it! (I just preferred the first option)😂
As I have a MMU3 I haven't yet finished...I'd be interested in the build process.
chris riley has it, but he only did the mk3 not the mk4.
@@awilliams1701 I'll have to look for that...I'm also putting the MMU2 on a MK3.
@@InspGadgt oh ok I can't remember if it was 3 or 3.5, but there is almost no difference except the cable. The extruder requires the same mods. There is a different cable for each because einsy vs buddy.
It seems to be a lot quieter than an AMS? And according to a bar graph it's also a lot faster even with that much ptfe tubes?
Correct, because it doesn't spend as much time feeding purge, which seems to be a very slow process on the ams
Great video.
I'd like to see the Prusa MMU3 used on an alternative 3D printer!
And, maybe an expanded look at the buffer and how that works/helps.
I for one would love to see the assembly process please!!
You should design and get PCBway to make you an "Operation" PCB like the seventies toy that lights up the patients eyes/nose and sounds a buzzer if you touch the sides while trying to remove the organs from him with a pair of metal tweezers. For extra points model the patient on yourself instead of the cartoon figure originally used.
Make this into a short, and it will get 10x the view. People want that mechanical action.
Good idea
Can you bypass the MMU and feed TPU directly into the head of the MK4 after the MMU has been installed once? I have heard different answers to this question.
The easiest way to do that is:
--> go in to the hardware menu, disable MMU
--> remove PTFE tube from the extruder
Now it's basically a standard Prusa Mk4 and works exactly as before.
@@RobinCernyMitSuffix thanks for the answer. That sound fine and simple.
Something I would love to see you test is CHT nozzles and their effect on color switching. A comparison on how much waste you need between colorchanges when using cht, cht clones, cht inserts, revo, revo hf. I heard revo is better for MMUs so that would be something interesting to test out.
Hmm I think cht would be a very bad idea with the mmu because it would totally mess up the tip shaping.
It would be interesting to see though.
I considered the MMU3 but I chose not too. The complexity of the design, the hassle of loading new filament, the desk space (which I don't) have and the cost all factored into my choosing to get a Bambu A1Mini. I own a MK4, but it was cheaper for me to get the A1 Mini with the AMS then the MMU3. It takes up less space, its a simpler design and feeding filament is dead simple.
You didn't say anything about speed -- it is the same amount of time to color swap as an AMS?
Changes are faster on the mmu
Ditch the prusa contraption and get an RMU MK3. I’ve run miles of filament thru them without any problems. Great product!
As someone who struggled with the MMU2, it's nice to see that they managed to get it sorted. The mechanism looks the same, and it's not clear to me what has changed.
My MMU2 kept jamming when retracting filament, regardless of how many cooling moves it did. I believe the root cause was the E3D V6 simply not being designed for it, filament tips coming out too wide. I would expect that the nextruder was designed with this consideration, but I wonder what they did to make MMU3 work with the MK3S?
It uses wider PTFE tubing. Even if the MK3 extruder made the "fat tip" it won't jam in the tube like the tight MMU2 did. It's noticeably easier to load and unload as well.
Will this fit inside a Prusa enclosure for the MK4? I am thinking of buying this , now I change the filament manually in the middle of a print.
If printing two models at the same time, I assume it would be possible to use one of the models as a purge model which will get more random colors but for some parts it really doesn't matter. I think that's a cool idea. Could also probably purge on infill only and keep the visible wall parts the correct color. Interesting possibilities.
Yes absolutely
PrusaSlicer has both of those very options built in. :) You can opt to purge into infill, or print a second object with the purge waste.
This is the way future multicolor printing should be done. If Prusa can get this pre assembled into a 1 box unit then the MMU would be awesome. Its a far superior system then the Bambu AMS
You are getting something wrong, let me explain it to you. The mmu3 and ams are exactly the same from a functional standpoint, one just has a rewinder instead of a buffer and sits in a box. Both retract and load filament. The biggest difference is the attached printer. The prusa uses tip forming, which is less wasteful yet requires more tuning. The cutter approach of the bambu requires no tuning, but wastes more filament. You can also do a mix between those methods, i for example do this on my voron. Ram out the old filament, so there is basically not much left, retract and cut 3 to 5mm of the filament, so the still soft part doesn't get mangled in the extruder. Colour changes take around 30 seconds, but i wont compare waste of my system to the prusa or bambu as my used hotend has a significantly longer melt zone, thus more purge is needed, which also contributes to the increased waste of the bambu. Current trend in the diy scene is to use cutters for the added reliability
@@kilianlindlbauer8277 Bottom line on what matters: 1. Less waste, 2. Quality, 3. Speed , All things the MMU does better than the Bambu AMS
I'm slowly becoming more interested in getting into an automated multi-filament solution. Can this MMU thing be used on an Ender 3 V2/Neo?
Question you said you took into account the new bambu feature of long retract. Did you actually change the flushing volume or left then at 100% because you wont see much difference in flush volume or time unless you do.
Also you can shrink the prime tower as well.
found a lot looking up Bambu Lab Filament Waste True Cost
They got down to 30% flush volume when comparing a 150g waste down to 40ishg.
There is also The Enraged Rabbit Carrot Feed that works with any Klipper based printer, very similar to the MMU in terms of how it works.
The MMU3 can definitely print with flexible filaments.
lmk your setup!
I mean, this isn't really any different from Bambu's AMS, other than a slightly different way of doing it and (seemingly) extracting all filament from the nozzle instead of cutting and leaving part of it. The main difference is that it doesn't use 1 motor per spool, but it uses 1 motor to select the filament and one motor to feed it. Theoretically that's better, because it's infinitely scalable.
Bambu focusses on the end result, making sure that prints have as high a chance of being error free as possible, while this focusses more on reducing waste material. Which produces less waste all depends on purge settings and how much filament is left in the hotend when switching.
Still a pretty cool idea though, but I feel like this is more of a high end hobby option, which would have been amazing anywhere between 10 to 5 years ago, whereas Bambu's solution is more of an "Apple-style" idiot-proof way of doing it.
The "hobby option" is still amazing! I had lot of fun with MMU and MK4 kit.
I wouldn't call it idiot proof. Anytime the filament breaks inside it's a huge hassle to get it out. What I would say is that it's plug and play, which makes it much more accessible.
I've only ever used one filament switcher which is the Bambulabs AMS. It works and it's very good generally. But "generally" doesn't mean "always" since I've had to take it apart twice to clear blockages in the 6 months I've owned it. That means unloading all the filament, detaching rear tube from inside, unscrewing the the tray, wiggling the tray out without avoiding yanking cables, undoing tubes and avoiding damaging PCBs and putting it all back together which is actually more fiddly than taking it apart. I think the design could do with an update that simplifies this process. Maintenance is part and parcel of these things, so make it as easy as possible. There is also the matter of poop that some people have mitigated somewhat with G code but will always be an issue with the AMS design and probably all filament switchers.
I've not used the MMU3 and I'm sure it has its own issues but on the face of it looks a little simpler than the AMS. It's a lot more expensive though. In a way I wonder why a filament switcher even has to be tied to a brand of printer. The switcher is a dumb device that is just told to change spools, advance, cut, signal runout etc. I expect the communication protocol is so simple that somebody could probably make an MMU emulate an AMS or vice versa.
As an owner of both, the MMU3 is better in all ways except footprint. MMU3 is faster, wastes less filament, can take any spool, easier to maintain and troubleshoot.
@@SentientTeapot2444 I don't personally have an AMS (yet...). How does the filament break inside? Like in the Bowden tubes? It's definitely plug and play, but with idiot proof I mainly meant that it's very hard to mess something up while using it. Sure, things go wrong, but that's a different matter.
I have the MMU3. I don't use the buffer at all. I didn't even print it. I just feed my dryer (sunlu s4) right into the mmu itself directly. When filament isn't in use, it just hangs out to the side. I've done something like 10K toolchanges already. I've had 2 legit failures. Maybe 4 in total. One of them was the line unloaded itself. I just had to push it back in. I don't know why that happened. The other one was white filament not loading properly after it had been used in a previous layer. It could be white is more problematic or it could be random. The user error issues were things like using wet filament that wasn't stored properly and became brittle and snapped off (happened at least twice, but that particular filament seems fine now after a lot of drying). One of them was accidentally setting the sunlu s4 to 60 degrees. This caused the line to bend and form kinks. There might have been others, but it's all stuff like that. I don't know if I should count running out of filament as a failure or not. In one case The filament was low enough that it was out of the idler completely, but hadn't tripped the runout sensor. The unit couldn't put it back in. (I did the beveled PTFE backwards because it makes loading easier). So technically it didn't run out, but it was close. Because it couldn't pull it back in it stopped until I could clear it out. And I had to wait a couple days to get that spool replaced on top of it. I had that happen a couple other times, but those times it actually did trip the runout sensor. Anyway it's completely flawless as far as I'm concerned. color is a requirement for me. So I love it.
Great video. You got a subscriber.
I would list No TPU on the cons. You said iy but didnt list it at the end. Great channel and dedication yo build that monstrosity.
yeah I would but literally no AMS/MMU systems support TPU.
I'd like to see more toolchangers out there :)
How feasible is it to make it talk to Clipper based system?
Very.
Im surprised nobody has
@@LostInTech3D I love the concept of less-wasteful multi-material printing - it would be great to see someone make it happen, maybe I could save for MMU then
Oh how lovely. Earrape and fine font captions with no voice overdubbed narration during the demonstrative informational segment of the video, for those viewers who have impaired vision and might be watching somewhere they can't just put on their glasses. *Nice*.
If you have any useful feedback on how I can help improve the experience, instead of sarcastically complaining then please send it my way.
👋🏻Hi, can you compare the noise level of the AMS/MMU3 + filament change inside the printer? I have currently X1C and thinking about switching to Prusa MK4S/MMU3 as the printer itself is quiet and for others reasons.
Does the painting process have a way to use graphics files so I can do logos and such?
So I’ve seen cardboard spools with dents fall off the rollers in other videos, so it’s the same situation in that regard as BBL AMS non-lite anyway, same is only an issue for the AMS when they’re damaged in the first place.
You can center mount them I guess if they're particularly bad. The AMS v1 has problems with cardboard dust fouling the rollers
great video! I wanted to ask if they could make this device more complicated, but I already know the answer lol
🤣
I wonder if we'll ever get the ability to hook one of these up to a Prusa XL...
struggling to find the appropriate term to google - I'm guessing there must be a design out there that just uses white filament but will just spray+mix the filament the colour wanted? (e.g. how a printer works) the heated head will help mix the ink through to make it less patchy?
I've seen rudimentary examples out there but yes, it does work!
Some guy used felt marker pens and a servo to smear a chosen color onto clear filament.
Downside is the difficulty getting full opaque color. For that you'd need to forcefully inject pigment into the melt zone. I imagine machining the intricate fluid channels would be cost prohibitive.
@@Metapharsical yeah I was imagining just some printer head but butchering so you could get the cymk + nozzle
The thing I'm most curious about for these open source MMU solutions is compatibility with 3rd party.
This plugs into the mk4 main control board right? Does it even support the previous prusas? How much work is it to hook it onto a voron? Could we use it one an A1 instead of bambu's own design if we prefered this?
Does it only work with prusa slicer?
good questions, it does support the mk3 or some versions at least. I don't know about a voron, I couldn't easily see the pinout when I looked, but this I think is for the mmu2 github.com/prusa3d/MM-control-2.0
someone would have to code it to work with a voron but it's all documented on the mmu-firmware repository.
Zero chance of it working with an A1 unless bambu decide to not be closed source.
Should work on other slicers if you copy the start, end, tool change etc code sections.
@@LostInTech3D I mean... you can already flash different firmware on Bambu. I'm not expecting it to work out of the box, but if it only takes a tiny bit of reverse engineering the pinout and communication between mmu and printer vs the AMS communication, it's probably highly likely to happen.
I don't really care if a project is open or closed source, there are reasons for both, as long as it's not apple or samsung style of products, doing everything to prevent any and all changes/repairs.
I don't endorse their behaviour towards many parts of the community, but I'd be very happy to see compatibility hacks that lets you use what ever you want along with the printer you bought.
Especially because Bambu got very popular on social media. many people bought them not necessarily knowing in details what they got into. I don't think those people should have to get rid of their printer and buy a whole other system just to benefit from the improvements in the industry once they've learned more and saw both sides of the coin.
@@SquintyGears One doesn't flash different *firmware* / printer operating software on any Bambu Lab printer, only *some parts of Linux* - and only on X1.
Bambu Lab is closed source (outside Linux and slicer) and even though we'd like to see it change, don't hold your breath 🤭
@@kimmotoivanen yeah I mean it is happening in some capacity even though everyone was crying like the world was ending and they could never personally recover at time of release.
So effectively Bambu got called out enough that things have shifted.
I don't think the calling out is going away.
@@SquintyGears Bambu Lab admitted that their slicer is based on open source (not shifting) and allowed changing some open source Linux components (but also made people to sign off BL responsibility for further issues with their printers).
Printing software, the one talking with AMS and handling motion etc is not open source and BL can legally kill/harm any attempts to reverse engineer that. P1 and A1 are simpler machines without Linux, so there is nothing that can be changed.
Prusa MK4 and MMU source files are in github - they don't make adapting MMU3 to other printers or other material systems to MK4 easy, but they make it possible. Where can I find AMS and Bambu Lab source files?
Sure, logic analyzer or USB sniffer can detect communications between AMS and BL printer, good luck with that, and then one could make translator between MMU3 and BL communication.
Could it be used as a multimaterial system?
I know conventional wisdom says no for single nozzle setups because of clogging concerns;
but the filament clearing seems really efficient (would be interesting if it could be replicated for, say, ERCF);
and for common material combinations (pva,petg,pla,hips,asa...) pretty much all materials can be printed at around 235-240 (even if not at best quality or healthily) so...that should not really be a concern i dont think
Ps. Wondering how much the waste tower be shrunk since filament purging isnt needed anymore
I don't think it would be a huge problem tbh, aside from the tendency of some of these to not stick to each other
I asked Prusa support about this before buying it. Buttom line: They think something like PLA+BVOH would probably work, but so far they seem to see it more as a multicolour device.
Something that's on my to-do list is to check out PLA+PETG. They are advertised to work individually and unlike BVOH, they are both cheap and easy and don't stick to each other. The main issue is probably going to be that layer adhesion is bad for quite some time after switching between PLA and PETG, so it would require substantial purges, probably even more than what the AMS does.
@@markus30000 Biggest problem with PLA+PETG is also the reason you want to use them. They don't stick to each other. Think what that does to your purge tower. We did get it to work on mmu2 though. Sorta. On occasion.
@@peterleblanc661 I'm hoping the purge tower issue can be solved by forcing the slicer to use one specific filament for the tower structure. That option already exists and I suspect this is what it's for.
Would be interesting to see the assembly process. :)
@10:30 your face, in the middle of one of those led indicating analog style clocks
😂 nope. I don't want to lose subscribers
What I do not get here is that since PCBWay is sponsoring this video why didn't you have them print some of the parts in metal?
Haha I guess I could
@@LostInTech3D You should really. Make a great video.
Finda?.. Have you named it that :)
Looks so pretty in that blue. Can we have a longer ASMR/working section the next time?
No they called it the finda.
Filament pinda?
@@LostInTech3D Haha, they are consistent with funny names then :)
Looks awesome ,wonder how hard it is to put it on an ender 3
I thought that wouldn't be a thing, but there is a video on Chris Riley's channel of putting the MMU2 on an SKR...so...it seems totally possible!
@@LostInTech3D seems like it would make for a great budget multi material /hueforge machine
can you weight the waste ? I have the feeling this system wastes just as much with that large purge brick
I weighed the waste?
@@LostInTech3D oops. You did. Sorry. I don't know how I skipped that graph ! Thanks X )
I like this a lot but I would insist on colour blending as well as colour changing. Essentially if you had 5 spools loaded: black, white, blue, red, yellow you could conceivably print just about any colour you want. Having a stark contrast between colours is nice but being anle to transition or blend would be better. Maybe I'll just have to wait for the MMU4. :)
Colour blending is hard because of the laminar tendencies....see my older vids on the geetech a20m, it nearly never blends properly
@@LostInTech3D But....but...I want it.
😂 let's hope the polyjet tech gets cheaper
i also want that, but have a hard time imagining a geometry which blends material properly without taking a lot of material to purge out
I do like the mmu but it’s space that I don’t have for the machine so a Bambu is what I have and a second AMS is what I have ordered and it still doesn’t take up as much space
The MMU3 is way more reliable vs the MMU2s. I had to remove the MMU2s inother to to be able to print things again. The only errors I have with the MMU3 was because I used filament that fell victim to the MMU2s who took a bite out of them, preventing the MMU3 from retracting and loading.
Waste is kinda hard to compare 1 to 1 in my opinion as the bambu has a longer melt zone, so it has to purge longer. The cutter approach is still more wasteful than the filament ramming/tip forming of the mk4, although mostly because how conservative bambulab is with retracting the filament, there is still a lot room for improvement.
I personally do a mix between tip forming and cutting. Ram out the filament to minimise colour contamination for the colour change, retract and cut 3 to 5mm of filament, so the still soft tip stays in the hotend and doesn't get mangled in the extruder gears, which was my only failure mode before i had the cutter. So its still a little more wasteful than just tip forming, but quicker as i dont have to wait for the tip to cool down and requires basically no tuning.
Btw, did you know that the tip forming procedure for the mk3 and mk4 is different? The mk3 uses classic tip forming with a short initial ramming period and than trying to cool down the filament tip, the mk4 more or less empties its nozzle with a rapid burst like extrusion above its hotend flow capability and the retracts the filament then. The mk4 procedure results in less overall waste as the majority of purge volume is generated by the colour transition, but when you have basically no old filament left, you just need to do a little prime and you are good to go
Yeah, and it took them like 9 months to perfect this. I wasn't entirely sure they'd get there tbh!
I can't say I'm on that impressed with that shuttle.
Moving back and forth seems to offer a number of possible mechanical failures.
My design would take the five tubes converging in a single point over a small funnel.
Each material goes down into the funnel to the hot end, eliminating the need for that little shuttle thing.
Furthermore, the narrow most point to the funnel has a cutter which allows you to cut away material as it is being fed in.
Allowing a new material being fed directly behind it for a seamless non-stop procedure.
You're describing something like the MMU v1.0
@@LostInTech3D and there is a reason why the MMU2, MMU2s and MMU3 was created.
Two thing I would like to see with this:
1, automatic spool change when filament runout.
2, changing the filament before color change so it does not ned a waste block, or do it in the inside of the print. Why not have stronger parts if we need to use more material anyways.
Oh, I think 1 is a thing already. I'll test it out
U can do both, u can even use a purge model I always do this for gears or models that I'm gonna paint
you can purge to model yes, but I think it still does a block?..I guess maybe I have more videos to make and check some of this stuff out! :D
@@LostInTech3D yes and no, if the model that you choose as a purge "block" has enough volume to purge the requested amount u can print without the purge block
Let me correct myself, the purge block it's always present but it's used only for the filament change, all the purge it's in the infill or model
I had to make a choice, Bamboo X1 Carbon or Prusa mk4 + MMU3 + self made/bought enclosure. I had chosen prusa, but I somehow regret it now. Prusa might seems cheaper, but when you add costs it's roughly the same, even with self build enclosure. I had mk4 before MMU3 launch for mk4, and mk4 was pretty good, although it was shaking so much while using input shaper. I made ikea lack enclosure for MMU. Then MMU3 arrived, it's cool when it works, but feels terrible when something goes wrong. Your filament is a litte brittle? It's likely it will break inside the tube so you will have problems, or the MMU3 pulley will grind it making it unable to move. It's really frustrating, the same spool that worked fine with only mk4 doesn't work well with MMU3. It isn't easy to disconnect tubes from back MMU3 while the whole printer is in enclosure. And it still takes a lot of space. Bamboo would have been smaller and wouldn't have so many long bowden tubes. When loading filament it sometimes can not pass throught second hole adter a pulley (when the end is a little bend then it's more likely), this would result in jam. Pulley in MMU3 has a small contact with filament that makes it easy to grind it.
Without MMU3 the manual color change was super easy to do in prusaslicer, you would thing it would be just as easy to do automatically with MMU3 right? Wrong. While using MMU3 profile that option doesn't work. Color tool also doesn't allow you to change color by layer. There is a way, but quite hidden and Prusa doesn't even mention it on prusa web help. I had to ask in prusa forums. You have to right click on model (on the right list) and choose hight range modifier and then add extruder change based on hight. I was doing one color change in the whole print (to make a sign with protruding text) and it still made a wipe tower. Prusasilcer also doesn't remember the color I set (just to see visally wich color is in which extruder).
When printing with single color but with MMU3 on (without MMU3 profile) the printer does the small pruge line, which is usually not enough (because with MMU3, the mk4 will unload the filament every single time when print ends). The printer automatically shhould purge a little more when MMU3 is installed.
When starting a print with normal/IS profile (without MMU3 profile) the extruders heats up first to 230 C, loads PLA, oozes a litte on build plate, then cool down to 170 C and if the bed is heated, then it starts mesh bed leveling. Then heat up to 230 C again and the print starts. It's makes a little sense to heat up then wait to cool it down to 170, it should heat to 170, do the mesh leveling, and then heat to 230 and load filament.
To sum up, it's nice when it works, but when there are problems, I wonder if I would have been better off with bamboo. I hope they will update prusaslicer with better color change by layer option with MMU3 (it took them one year to add touch support for mk4 so it can take some time)
can u post assembly as a separate video?
Since i heard of the 3D Chameleon this is definitely my got to solution because it basically gives the option to print with 1000 or more colors 😅🙈
I'm 1/3 into the video and still have no idea what you are talking about. Is it a printer? Just a color switcher? Is it compatible with other printers? A short introduction would have been great for understanding the video.
Haha fair point
Why the channel is called " lost in tech"
It's prusa's multi material unit. It's a color/material changer for prusa printers
Only for Prusa?😔
@@3dPrintingMillennialonly prusa
Boi I've wondered for a while how did the MMU system worked, it seems such a well engineered idea!
It's strange seeing all the praise for the Bambu when the AMS is so slow and wasteful, it looks neat and it's reliable, but at a cost that I personally find completely unreasonable. This seems a much better implementation of the idea of multiplexing the extruder, but i've gotta say I now understand why it seems to be such a niche thing in the 3d printind world, the fact you have to put all of this together yourself is ridicolous.
Had Prusa gone for the "completely assembled product" type of route they probably would've smoked the AMS before it even had a chance to impress anyone, but also the costs and work necessary for making that happen probably didn't make much sense for them considering how many units they would sell.
Angry rabbit carrot feeder for Prusa. Does it work on non prusa?
Is this only compatible with prusa printers? Or can I put one onto my ender 5 plus?
It won't work with an ender out of the box, even though it's theoretically possible
I use auto rewind spool holders and it takes very little extra space.
can you ...er...link it? I'll try to allow the post
A series of tubes!
4:18 No, they are all retracting the filament. You talk about cleaning out the nozzle. There, bambu cut‘s the filament and pushes out (maybe) more in a much faster process. To prove the difference you would need to measure the purge tower (and poop) of both.
I did. When I say waste, I mean both of those.
@@LostInTech3D They both retract the filament. Your argument is wrong. Think about the real difference and how to prove which is wasting more.
I'm not sure at all what you mean.
I would have preferred to see some quantitative analysis of the waste, number of grams per same print from MMU and AMS.
Did you miss that part?
Is the setup and build part of this prohibitive for most?
I didn't find it hard personally, but I can see why some might not wish to
I would love to see printers that can mix melted filament the way inkjets mix ink, so full color CMYKW (we need additional color for white as it can't be inferred from the paper) 3d prints could be done.
Inkjet printers do not mix the color. Our brain does that. If you look at the paper with a microscope, you can see many small individual dots of color in cyan, magenta, yellow or black. Our eyes just can't resolve the individual dots and so our brain mixes the colors and creates the impression of these many colors that don't actually exist on the paper.
There are full-color printers such as those from Mimaki. As far as I know, a layer of resin is applied and cured in one go. Then ink is applied on top of the cured resin as with an inkjet printer. You can see the result in the latest video from 3D Printing Nerd, for example.
the design reminds me of pick and place cartridges.
I build ercf for my voron and i had only 50 unsuccessful changes and one successful and now i think about trying to use mmu3 on my voron
It's nice to see that this works well, but loading looks like a massive pain compared to the AMS
Yeah it is quite a bit more involved
Mine gets delivered tomorrow. Excited to start using it ^-^
Interested in trying out things like PLA supports for PETG prints, or vice-versa. (Or TPU too!)
Could even get soluble supports for complex prints! But I doubt I’ll ever need that xD
I’ve already got a few designs floating around in my head for multi-color and multi-material prints.
Yes I know you said TPU isn’t supported, but I’d assume high durometer TPUs loaded and printed at slower speeds might work fine.
I think you might get away with it with some modification and removing the buffer. I might try some things!
So TPU Doesn't work with it...but maybe still try it :P Maybe a mod to make it work(like you can do on the AMS)? We need to make multicolor flexible printing mainstream :P
I'll see what I can do lol
would it be possible to install it in mt neptune4 plus?
Not gonna lie, I didn't know there was a significant amount of assembly required. If I'd known then I 100% would've voted for that
🤣 oh well. I might do something with that footage, maybe second channel, if I get some free time
You did NOT delete that ?!
🤣
it only working with prusas is the big down side for me. if i cant make it work on a ender then does it really even exist?
No, but there are other multi-color solutions such as the Enraged Rabbit Carrot Feeder or 3D Chameleon. The 3D Chameleon can be used with pretty much any printer.
@andreas.grundler right. But you asked for comments on what we think about this. So it not being universal compatible makes it kinda pointless to me. Unless I can hack my printer into using just enough prusa parts to make it work. Like what do I need to change out, just the toolhead? The electrics and code should be able to be put on anything and changed to make it work with other configurations, retract, feeds, ect.
It's open source, you can make it work with your printer (at least in theory, though restrictions (open source printer fw, filament sensor) apply) :)
@@kimmotoivanen so that's what I was originally thinking but then he said it only works with prusa. Must be the extruder part that needs to be moved onto other printers for them to work.
Thanks for great video. Must say that quality of print is not god as Bambu lab, Can see the difference in the colour area