Is MMU2S the best approach for multimuterial?
I have an mk3s with mmu2s since last May. After hours and hours of tweaking, carefully trying to find the optimized position for FINDA and IR, using skinnydip, fast skinnydip, dribbling approaches on prusaslicer, changing temps, filaments. Printing mods, eliminating the buffer doing everything i could find i managed to get a few ok prints. By OK i mean minimal intervention. Sometimes the filaments gets a little more stringy and sometimes with larger tip. sometimes FINDA or IR misfires. My point is that is takes a lot (for me anyway) to have a good print. And if that failures is at the beginning you can restart. But if you get a filament failure after 5hrs it may give you a layer skip. Or if it is persistent (i.e. something with the filament or a new roll) you get from a 10 hours print to 24 hours. Unless you don't go anywhere an babysit it. I know that MMU2s is not for beginners but that much tinkering is absurd. I bought it to print with soluble materials. I dont find multicolor option really a must.
So i was wondering if maybe prusa should redesign the whole concept. Maybe with the new MK4 in far future and while in design state they would considered the multihead or exchangeable head approach. Something that will eliminate filament coming and going and all the variables that come with it (temp, tip control, FINDA etc). And two aspects that i would find beneficial would be the ability to use different nozzle sizes on each head and less time. No more waiting 30 sec per filament change (if succesfull). I know there are aspects like oozing and maybe a kind of filament bucket but i thing there will less of a hassle from all the fine tuning that current mmu needs. Concurrent heads instead of exchangeable limit the working space. But his can be said for the wipe tower also.
I dont know i thing prusa at some point have to decide how to procced with this. I know that are a lot of people that made this working. But the point is it should be that so many people that can't. I love my printer and i'm actually excited with a new printer (MK4) with larger size and maybe geared extruder (i currently have the bondtech bmg since at 0.25 nozzle and 0.5 layer more often that not i was getting heat creep ) but i hope they offer with it a multihead instead of another MMU.
RE: Is MMU2S the best approach for multimuterial?
e3d has their toolchanger design out and open sourced
there was a toolchanger machine at ERRF that had 19 tool heads - not sure what toolchanger mechanism they were using
RE: Is MMU2S the best approach for multimuterial?
Is it the best? By far not. Can it get better? Of course, and it should. There is no need for MK4 for that. Better firmware (communication between printer and MMU unit) and slightly redesigned extruder head (current lever system is horrible to setup and maintain). I've seen a comment that PR is already working on better firmware and I hope they are also spending time on the extruder head redesign. But even the current solution is already working for me and I think it's a good base which can be build on.
I still prefer the current solution over the Palette one. It's the only competitor in this price range. Everything else with dual head like BCN3D or Ultimaker is completely different story (over $2k).
Often linked posts:
Going small with MMU2
Real Multi Material
My prints on Instagram
RE: Is MMU2S the best approach for multimuterial?
@nikolai-r
Of course it can get better. My point is though that, even that it works for most people, from the comments in the forums -and personal experience- there is a significant number that either stopped trying, or gets problems quite often and this cannot characterize a product as a successful one. I'm still trying to make it work but at some point is going to become a lost cause. For me a palette is a similar approach and personally i dont thing i would consider it as an alternative (even if i didn't try it). I'm not sure what can be done in a firmware level to improve on MMU2s that much. Some of the problems are due to the way it works. Taking filament out and reinserting new filament. Too many gears, paths and changes can be add to failure. In a dual or tool-changer we have your filament already preloaded. No unnecessary movements which are also time consuming. Ultimaker is 2K+ but i dont thing that has to two with the dual head rather with the company's policy. In a mmu2 we have two extra motors a second board extra sensors buffers etc. I dont see how material wise is going to be that much more expensive with a second extruder.
The MK4 idea is there. Announced by prusa and supposed to be a whole different approach (core xy if i'm not mistaken). They could easily support their current mmu unit up to a point for their current line (and if it improves believe i'll be the one shouting with joy), but since they are in design stages for MK4 maybe it could be better to take a different path for mmu into consideration.
In the end of the day, at some point, a product has to be get out of the beta state, or better the "not for everyone" and either become a stable product or be redesigned/replaced. MMU is trapped there. I think prusa is a great company with an awesome support and their MK3s is the best you can get value for money. But i personally dont see many people going down the MMU2S route. And with the need for soluble/breackable supports for complex designs there is a market for a proper multimaterial unit.
I saw that and i find it very interesting and promising. The only downside is see is that it may be rather expensive due to the exchange mechanism and additional rails. But i would love the concept of exchanging different heads. i.e. a cutter an 0.25 nozzle a marker for pcbs etc.
RE: Is MMU2S the best approach for multimuterial?
@andreas-t10
I don’t think CoreXY printer will follow the MK* line. Most likely it will be Prusa XL. And once they got enough experience with the new 32bit board (Prusa Mini, Prusa XL) they will upgrade MK3 to MK4.
There are actually more people using MMU as you think. In FB groups people are showing stuff from time to time and there are more then 5 which you could think based on this support forum. It will be great if they will continue the MMU line and make it better. Because all user of existing printers (MK2.5/MK3) would profit from it.
Often linked posts:
Going small with MMU2
Real Multi Material
My prints on Instagram
RE: Is MMU2S the best approach for multimuterial?
Well maybe is called XL and not MK4. The point is they are designing something new. Maybe it could be a good approach to consider offering an alternative -maybe for just for the new XL printer- to those users that never got decent results from MMU2. An later they can give it as an alternative option to MK*. Maybe its my fault and maybe the actual consumer base of MMU2 is quite large so as the percentage that is not happy with MMU2s is small. From a small search i did concerning ultimaker and dual head printers, there are not clear of problems but the number of people that have constant issues with them is a lot smaller than mmu2s. I guess in the end of the day Prusa really knows the numbers. For me the fact that works for a lot of people doesnt make it a success since it should work for the vast majority with problems that can be solved through customer support (which PR support is absolutely great but they cannot do much in this case ). MMU may be a good idea but for the time being the alternatives like dual heads -from a small search in the net- is a more stable approach.I truly hope for it to be improved but im not sure that it can get there. 5hrs print becomes 12h with all the loading and unloading which in turn becomes 24 with all the failures (if it prints at all)
RE: Is MMU2S the best approach for multimuterial?
i wouldnt say the MMU is the best way to handle this but its certainly currently the most cost effective and versatile solution as you retain all the direct drive advantages and have options to recover a print if something fails.
the sensors dont just misfire on their own, it usually means they are not aligned correctly and there are a ton of MMU users out there that you dont hear anything from because it just works for them.
RE: Is MMU2S the best approach for multimuterial?
@pixel
I know that there are probably a lot of users that MMU2S works for them. My point i try to make is that the number of users that have issues with it, or write about them in forums, are a significant number.A number that sometimes cant find a solution even with Prusas excellent support and a number greater than lets say ultimaker's (i'm not trying to sell it i just refer to it bedause of its popularity) dual head users with problems. A product -in my humble opinion- should not have so many users with problems. Sensors may work and be align perfectly but they misfire sometimes out of the blue, because something happened to the filament or some other trivial reason. For example they may be aligned perfectly but when i use a different filament something it may not add up. Or i start a print monitor it for 3 hours and then in the morning i see that there was a failure because for some reason it didn't leave the extruder properly and stuck in the PFTE.They are sensitive to any change in the loading/unloading process. You have to find the perfect equations for filament brand,type, loding unloading speed, temperature IR and FINDA placement etc. A dual head doesnt care for most of them if it loads to the extruder at the start of the print it stays loaded until the end.
I'm not sure about he cost effective or versatile approach. About cost if you keep in mind the number of extra step motors and electronics in mmu2 i don't see how an extra extruder/head would be more expensive. About versatility. Maybe in the context of multiple colours . Not in the context of multiple materials. With multiple heads/extruders is easier to manage differences in temperature or filament nature (ie how tight the gears will be on ..maybe flex with petg, or printing with pla and a little bit older pva). And far quicker also. Also you may have a 0.25 nozzle on one head for printing details and 0.4 on another for normal usage. Differences in temperature can be achieved in less time and as i said above you eliminate all the time of loading unloading. Finally about the recovery. Most problematic scenarios concern loading/unloading. Without those recovering will be far less and the solutions maybe far simpler.
I'm not trying to sell a different printer, i have and enjoy a prusa mk3s and even with my disappointment with mmu2s. I just believe that mmu2s is not on the same level of workmanship/quality and for a multi-material a dual head maybe is a better approach.
And at the end of the day a 10h print becomes 24h (if it simple) just because of all the wiping,unloading, moving the selector, loading,wiping (wipe tower could be smaller or use some kind of bucket since a head only holds one material) and thats, i believe inefficient.
RE: Is MMU2S the best approach for multimuterial?
@pixel
...About cost if you keep in mind the number of extra step motors and electronics in mmu2 i don't see how an extra extruder/head would be more expensive....
This is where you’re wrong. Dual head just costs more. Look at the existing products on the market and compare the price.
You could also think about the technical side what kind of consequences it has:
- Complete redesign of the x-axis due to weight and size (direct drive is big, heavy, bulky).
- New main controller board required with additional stepper driver, mosfet, fan control. You can’t reuse same board
- Higher power usage due two second heater, new power supply required
- ....
The beauty of the MMU design is simplicity in BOM and easy to attach/detach to your existing MK3 printer. That is why it’s so inexpensive and popular.
Again, I understand why you’re upset with the MMU but I’m glad PR released that product. Otherwise I wouldn’t have such a tool at home and wouldn’t be able to do what I’m doing today. Spending $200 or $2000 makes a huge difference.
Often linked posts:
Going small with MMU2
Real Multi Material
My prints on Instagram
RE: Is MMU2S the best approach for multimuterial?
@nikolai-r
Concerning the price. There are printers out there , like davinci or creator pro, with dual heads that are quite cheap. Some of them even cheaper than the mk3s without the mmu2. Comparison with different companies is not usually correct because there is the matter of brand, components ,pattents, targeted market, reprap vs manufactured, opensource etc. About the redisign and maybe the usage of a different psupply. The mmu board drives 3 steppers (that means extra drivers, power etc), so im not comparing the current board with a dual in the case of price but rather the combo of boards and electronics of mmu2 and mk3s with a single board that can handle the dual heads. Thats why i said it maybe a good idea to incorporate this as an option to prusa xl or whatever they design now with a new board (and then maybe offer it as an upgrade kit). I dont say that mmu2 should never be done. Im glad that prusa works on multimaterial options. I just believe that this path is a hit or miss. Most people made it work, a significan number didnt. There more worked and stable approaches out there.Im trying to gain a better perspective from different opinions and thus i appreciate your input. I hope at some point i'll be able to use my mmu2 unit as succeafull as it seems your using yours.
RE: Is MMU2S the best approach for multimuterial?
@nikolai-r
...like davinci or creator pro, with dual heads that are quite cheap...
As far as I can see they have one head with dual extruder. This setup has a huge disadvantage of oozing filament from the nozzle which is not in use.
Often linked posts:
Going small with MMU2
Real Multi Material
My prints on Instagram
RE: Is MMU2S the best approach for multimuterial?
You end up with a whole different set of issues with a dual extruder printer:
- You have to keep adjusting the nozzle height so that the two nozzles are the same height.
- You still have to purge the idle hotend on a purge tower so that there is plastic coming out of the nozzle so there is no holes in the print.
- The idle hot end tends to ooze which either causes the ooze get on the current layer or hardened ooze hits the model and knocks it off the build plate.
- You end up printing a ooze shield to protect the model from the dripping ooze. Otherwise that white sheep has black specks in its wool.
- The increased weight of the dual extruder head causes print artifacts unless you slow down the print speed.
- The width of the x-axis is reduced by the width of the additional extruder.
For me, dealing with getting the MMU2 working was lots better than the time and money I sunk into getting my TAZ5 with dual extruders working.
RE: Is MMU2S the best approach for multimuterial?
@bruce-p4
-About adjusting the nozzle. Im not sure why this has to be done often. I mean every time your change the nozzle maybe. Or after x hours of print (especially if your using metal filaments to align the two heads), but first layer calibration has to be done every now and then for prusa also. Keep in mind that in dual heads you may have different nozzle diameters and avoid changing from 0.4 to 0.25.
-Purging has to be done but not in the extend of mmu2. There is no need to "clear" the nozzle of the old filament. So a smaller wipetower with less wiping may be needed,
-Oozing is one of the problems i read while searching the subject. Some printers drop the temp of one head at 150 and then when needed back to the needed temp to avoid oozing. And if this is timed appropriate with no to little extra time. Some elevate the unused head a bit.
-The weight is a problem but this MAY be reduced by using pancake motors with 3:1 ratio gears like the bondtech. Or maybe not. I mean im sure all the companies out there that use it (and companies like ultimaker and flashforge ), may have found some alternatives to this. In any case if we were to compare speeds mmu2s is already REALLY slow with all these loading unloading etc. And again this is increased with a failed attempt to load/unload in the middle of the print while you are not looking.
-The width is already reduced at some point when using a wipe tower. Either mmu2 or dual head. And i dont think that this is a bad compromise. If you design an XL printer like the one prusa is doing now you may design it in a way that when you do a single head print is full size and the width is reducing only in multi-head/material.
-an added benefit is the need for only one firmware. Now you have one for mmu2 and one for mk3s. Instead of having to support and coordinate two different boards/components we do it only once.
Dual extruder/head is out there and some companies made it work some not quite. MMU2 is slow both in printing time and also in making it work and all the ideas for its improvement come mainly from prusa (however good the company may be, multiple companies working on the same idea usually produces better and quicker approaches to problem solving). I can really see a benefit from dual heads. Even cost wise if it was to implemented in a new printer. Nikolai and you made some valid points and the fact that you are so sure that mmu2 is on the right track gives me incentive to try to make this thing work. But also due to the fact that there is no alternative to prusa for soluble/breakable supports. If with XL comes the option of duality i'll definitely go for it (except if mmu2 becomes stable -and i mean easy for an average user to start printing- and fast).
RE: Is MMU2S the best approach for multimuterial?
The biggest issue with multi extruders vs the MMU/Palette is that you are limited to a number of filaments equal to the number of extruders. I haven't seen a 5 extruder setup and the tool changing systems (where the extra extruders are swapped in and out as needed vs fixed to the X axis) I've seen are all substantially more expensive than a MK3+MMU.
No the MMU is not perfect and I wouldn't even call it ideal, but at the price point it falls in it is almost unbeatable. It's only real competitor is the Palette which has the plus of supporting more printers, but supports one less filament. I've heard pros and cons for their splicing model compared to Prusas retract and insert model, so without hands on experience with both I view that as a wash at this time.
RE: Is MMU2S the best approach for multimuterial?
@gnat
The truth of the matter and maybe where the different of opinions lie, is that i dont actually see the existence of more that two paths something that would be considered the game changer - or even needed. For me more that two paths is for multicolor prints not so much for multimaterial. Yes its interesting to combine flex with petg with snap supports but this cannot be done easily with the mmu2 either no matter how many material it can simultaneously support. Mainly because is going to be quite the tweaking to align the extruder door and mmu2 barrel for both TPU and PETG (not to mention the temp difference). Coloring is something that can be done in post processing and usually giving better results than a four color print. But thats me. Other people may consider it a must. But my point is made on the question of if mmu is the best scenario for multiple materials not multiple colours. I havent tried palette but from what i've read is not the something that improves over mmu.
RE: Is MMU2S the best approach for multimuterial?
...But my point is made on the question of if mmu is the best scenario for multiple materials not multiple colours. I havent tried palette but from what i've read is not the something that improves over mmu...
There are more out there. But this one was easier to find for me ;). MMU2 as multi material unit is real.
Often linked posts:
Going small with MMU2
Real Multi Material
My prints on Instagram
RE: Is MMU2S the best approach for multimuterial?
@nikolai-r
Basically those scripts and later the dribbling and skinnydip methods is what pushed my mmu2 unit from producing utterly useless filament tips to a rather workable state. Even if we have a rather different opinion on the mmu-multiple extruder subject i agree that your approach and the one's from antimix and Erik Bjorgan's/TNDavid gave "a new hope" to some of us that where a step away from using mmu2s as a toiled declogger. I got my first print thanks to those. I think that maybe prusa should do is test and incorporate those ideas in their master fork of prusaslicer.
RE: Is MMU2S the best approach for multimuterial?
I though a lot of time on how to improve the MMU concept, but unfortunately every attempt was adding considerable extra costs.
I had several ideas, but no way, any one, added at least a new motor step (costs++ 😆 ) and then, the board was not enough to drive it and I should add all the additional circuitry to drive the motor.
The PRUSA rolling MMU is genial, since it uses only THREE motor to move five filaments!
I realize that a couple of good of drive gears Bondtech to really grip a filament without shifts, cost about 50€ each; that means 50x5 = 250€++ cost to add, if a new MMU2 design plans to replace the 5 pulley hobbed and bearings with good gears. That would have created basically a 600€ MMU2, but more or less as it is now, just with good filament grip into the MMu2.
The MMU2 is lucky mix that keeps costs reasonable. If I had to spent 600-2000€ just for a multi color option I would have never purchased it.
RE: Is MMU2S the best approach for multimuterial?
The more I read about people's experience with the MMU the less I want it. I'd love to get a multi-material printer, but it seems with the MMU2, not only do you have to upgrade and re-configure it a 100 different ways just to get it to kind of work, but even after that "work" is defined as "you only have to intervene 3-4 times per print and even then every 3rd or 4th print fails." The Octopod Ocotoprint plugin even has a special notification type "MMU needs user intervention." I understand that with a lot of time, patience, and standing by the printer as it prints, MMU can work, but that's just not what I'm looking for.
My MK3S is currnetly working great. It almost never fails a print (when it does it's usually due to my error) and has run for many 100's of hours. I would need a multi-material printer to work just as reliably. None of the solutions on the market that I've seen (multiple extruders, MMU, Pallet) seem to meet that criteria. I'm somewhat hopefully for the E3D tool changes once they work the bugs out, but I'm not sure if it'll turn out to be really expensive. On the other hand, i would gladly pay 2x what MMU costs to have something that actually just works.
RE: Is MMU2S the best approach for multimuterial?
I would need a multi-material printer to work just as reliably. None of the solutions on the market that I've seen (multiple extruders, MMU, Pallet) seem to meet that criteria. I'm somewhat hopefully for the E3D tool changes once they work the bugs out, but I'm not sure if it'll turn out to be really expensive. On the other hand, i would gladly pay 2x what MMU costs to have something that actually just works.
Just of curiosity: Why do you think tool changer will work better than dual head printer?
I think "just works" doesn't apply to any 3D printer. I read on daily base about so many issues with MK* printers (and all the other brands). You made MK3 work for you reliable but there are many people which would disagree with you on this topic like you do on MMU2.
My recommendation would be to get MMU2 only if you have a second printer. There are limitations in regards to flexible filament and the extruder head is more difficult to maintain.
Often linked posts:
Going small with MMU2
Real Multi Material
My prints on Instagram