Mosaic Palette: Single Extruder Multi-Color and Multi-Material 3D Printing

Lots of solutions have been proposed and enacted for multi-color and multi-material 3D printing, from color mixing in the nozzle to scripts requiring manual filament change. A solution proposed fairly early on was to manually splice the filament together, making a custom spool. The printer would print as normal, but the filament would change color. This worked pretty well, but it was tedious and it wasn’t entirely possible to control where the color change happened on the model.

You’ll find some examples of the more successful manual splicing hacks in the pictures below. Scroll down a bit further to find our interview with Mosaic Manufacturing at Bay Area Maker Faire 2016. They have a new product that automates the filament splicing process with precision as the ultimate goal. It unlocks a single extruder printer to behave like a multi-extruder model without stopping and starting.

Mosaic pulled off a very difficult combination of two methods mentioned above. Their flagship product is a machine they’ve dubbed, “Palette”. It’s an automatic filament splicer. Up to four different filaments can feed into Palette, and it will splice them at determined intervals. This would be cool by itself, if only to save the tedium of splicing and winding a custom spool by hand.

The real killer app with Palette, however, is the software that runs alongside it. Palette can take the GCODE output of any properly prepared multi material file from any slicer, and then precisely combine and splice the filament. This can feed into any printer without modifying it, aside from sticking an encoder somewhere in the filament path. The results are indistinguishable from a dual, or quad extruder set-up.

Kill me... it whispers. Free me from this torment.
“Kill me…” it whispers, “free me from this torment.”

Now, it’s not really a reduction in complexity from a multi-extruder set-up. In fact, one could argue that it’s more complex in some ways. Then again, one only needs to look at the abomination that is ORD Solutions 5-extruder 3D printer to see, if not the light, at least a gentle glow.

If we make a quick comparison between the two solutions we can see some advantages and disadvantages for both. The biggest disadvantage for the Mosiac is that the materials have to be fairly close in properties. For example, it probably wouldn’t work to splice polycarbonate and Ninjaflex together. A multi extruder set-up allows for more variation between the filaments, as well as different extruder set-ups entirely. For example, it’s entirely possible to have a .4mm nozzle for printing the body of an object and then have a second extruder with a .2mm nozzle to put down the fine-detailed text on top of a print.

The biggest advantage is the integration with the printer’s existing extruder set-up. Consider a delta-printer. In order to print nicely it typically requires a bowden set-up for the lightest extruder assembly possible. To add five nozzles to the end of a delta would have terrible results. It would vibrate and print terribly with lots of ringing, and eventually, destroy itself in the process. Mosaic simply sits in front of the extruder-cold end, where the filament usually comes in, and does its thing. This holds true for a regular printer too, the lighter the extruder the better.

For most users, the Palette will fit their needs. They can add it on to any printer in their shop, and get color and material mixing. As Mosaic shows in this video, since PLA is filled with just about any fiber and dust manufacturers can get there hands on these days, you can even do conductive paths within a PLA case.

It was a pretty big challenge to get the Palette to work. Building the automatic splicer was one challenge. Anyone who’s 3D printed knows how finicky printers can get about filaments with diameters just a little bit too large.  Mechanically the Palette is fairly complex. It draws in up to four different filaments, measures them, and splices them exactly. It then feeds the newly made filament out to the printer.

They quickly found, that even though the filament was splicing properly, the prints were coming up malaligned. It ends up that most printers are extruding a different amount filament than they claim to. It’s really hard to get a printer perfectly calibrated so it extrudes only the expected amount of filament. Their solution to this is the, “Scroll Wheel.” As mentioned before, this is just a rotary encoder that sits on the exit end of the mosaic, as close to the printer’s extruder as possible. The Mosaic uses this to calibrate its splicing points depending on how the printer is performing.

Pallet getting certified.
Palette getting certified.

As for the previously mentioned software stack, Mosaic has decided to open source it. They have a core demographic they are catering to, but people kept coming up with cool uses for the tool, if only they could hack the firmware for it. So, Mosaic did what every company should do, and let them hack.

Mosaic has already successfully Kickstarted their product and they’ve even shipped the first promised units. If you’d like one, they’re taking pre-orders for post-Kickstarter units. For more information we highly recommend reading through their blog, which documents their entire process from design to manufacturing. It’s a great read and even goes over less covered aspect of consumer manufacturing, like getting your product certified. It’s a cool product and the level of polish shows just how real the industry is getting.

32 thoughts on “Mosaic Palette: Single Extruder Multi-Color and Multi-Material 3D Printing

  1. Nice idea, but I would really prefer true mixing.
    Maybe someone will come up with a 5-color nozzle, but using 0.75mm filament or even thinner.
    So you get substractive coloring with CMYK plus White with way less bulky drivers.

  2. 1) Allow for some waste.
    2) Put a color sensor at one corner of the build platform.
    3) When you expect to change colors, go to that corner,
    and extrude filament until you get the new color.
    4) Resume printing

    1. If there were a way to pull the filament out of the feeder and insert a different one automatically, you wouldn’t have to plan the color change 20 cm before the actual need. You’d only waste what’s in the nozzle.

      1. I thought, what if the extruder had the ability to cut the filament just above the feed-in mechanism, and could grab a new one just by poking the end against the intake while the mechanism spins.

        That way you could have one extruder with a palette of many filament tubes in a holder, and it would simply cut the filament when not needed and grab a new one – purge the stub of old filament in the waste basket, wipe the end against a blade – and continue printing.

    2. Good idea, but you could just allow for enough waste to be sure the blend had time to flow through to the nozzle based on a few experiments. There could be a standardised test print to do these tests and find the parameters for a given environment, printer and blend.

      Why accept extra waste and the need to do a calibration print? Because it is completely a software solution an no extra hardware is required.

  3. I think your analysis is unfair. I’ve actually had the pleasure of meeting both teams in the past and find that they’re completely different tools for a different job. In the case of ORD, they’ve got some pretty ridiculous specs for their machines compared to what else is on the market in that price range. Mosaic is sort of a one size fits all accessory that add’s additional functionality to a printer.

    The Mosaic is a nice piece of tech but to compare it to ORD Solutions and call it an ‘abomination’ is flat out wrong. The whole point of having different nozzles is that you can adjust the flow rates of each material and fine tune for filament parameters. Even the Cartesian build for it makes sense when you have to move that many nozzles and adjust them.

    If you want a cool solution to add to your printer buy the Mosiac, the guys know their stuff. I met them in fall of 2014 and they had a solid idea going then. If you want something more high end super tight tolerances under 5k Go ORD, their machines are built to last and have a lot of specs you can fine tune.

    Sheesh you’re comparing apples to oranges here!

    1. I’m not certain I can be convinced that the ORD thing is, if not an abomination, at least moving away from God’s light a little. I don’t know in what situation you’d need quite so many extruders; and whether therapists are equipped to deal with trauma caused by their management.

      In all fairness though, I agree that the ORD Solutions printers I’ve seen in person have been nicely made. You’re right about the tolerances. I’m just not sold on them as a concept, a little too elaborate.

      1. The fact you don’t know the use case and completely look at it in an unfair light certainly makes me even question your competence to write technical articles about 3D printers. Instead of throwing out a half baked opinion and promoting a product that you’re a fanboy of, why not have some integrity in your writing and get some actual data instead of hand waving arguments. Everyone’s entitled an opinion but technical comparison shouldn’t be about who’s got the biggest and loudest horn and the friends to show of their toys. If you’re going to bring in a product to show off don’t compare it to something that isn’t relevant.

          1. I don’t have to do any research when you’re saying you don’t know about something yet decide to comment on it and include it as a comparison in something else you’re showing off. Compared to the Mosiac, I do agree the ORD printer looks daunting but don’t just dismiss the product, that’s someone else’s hard work. That’s why I think you lack any damn integrity. You should know better as an Engineer than to put a bias aside and focus on the facts not your opinions.

            “I’m not certain I can be convinced that the ORD thing is, if not an abomination, at least moving away from God’s light a little. I don’t know in what situation you’d need quite so many extruders; and whether therapists are equipped to deal with trauma caused by their management.”

            “… I’m just not sold on them as a concept, a little too elaborate.”

            I guess you did a lot of research on me before you made the statement that I have no integrity and operate without “experience or data?”

            I really don’t give a damn what your reputation is. If you’re writing for a technical site and obnoxiously throw around your credentials like that, yet can’t provide an accurate comparison that doesn’t resort to product bashing something you don’t know about, how is that credible?

        1. I really shouldn’t rise to this sort of thing…. I was being nice about. It doesn’t look daunting. It is daunting. It’s a list of all the “cool” features a 3D printer could have implemented without any finesse. It’s a children’s drawing of a race car with jet engines and circular saws. It’s got a carriage, bed, and z that are too heavy for 8mm rods. It’s got the option for 150$ dollar “cool muscle” industrial servos with advanced feedback attached to an already overstressed GT2 Timing belt; which is about as meaningless of an upgrade as I can think of. Moving the bed up is a disaster of a design decision slowly bringing the heaviest moving part and thus the entire assembly’s center of gravity above where it should be for any hope of stability. The cartridge cold-end design with the large metal gear is overkill with no perceptible benefit over better and, surprisingly, cheaper solutions like the E3D Titan or the Bondtech (really? 225USD?). It has a complex filament feed that only supports odd sized half rolls that few reputable sellers commonly stock. You can’t define a plane with five points.There’s no non-stick coating to deal with the sticking problems that happen with multiple extruders. There’s no mechanism to adjust the height of the five nozzles which will eventually get caked on or worn down if they are brass. The wire routing is atrocious. There are threaded rods thrown in for some inexplicable reason. It goes on.

          tl;dr It’s all marketing wank with over-the-top useless features that no one I know uses.

          All that would be forgivable if they pulled a mechanical miracle and the print quality was any good, but for the price it’s laughable. The Prusa i3 MK2 at a fourth of the price blows it completely out of the water. https://www.facebook.com/OrdSolutionsInc/photos/?tab=album&album_id=1220996147927906

          Oh, and I googled for reviews to make sure I wasn’t completely talking out of my ass, here are the only two I found:

          “This review is fabricated. This printer is the biggest pile of **** you’ll ever see. The cooling system is absolute garbage, which overheats and constantly clogs the nozzles. The heated bed is a joke and the printer can’t even heat it to a proper temperature. Over time the heating elements loose power and the temperature discrepancies increase.” http://3d-printers.ireviews.com/ord-solutions-mh3000-five-color-review

          “I too have had a bad experience with this company. More issues than I have time to type out and discuss so here is the short version. Linear shaft was loose, LED lighting in poor location, no Mac instructions, warped print bed, broken Z sensor, setup required drilling and modification of parts, support team doesn’t even work with Mac computers and couldn’t help, you need to use 4 extrudes at once to make use of the entire bed, only a small area (135x217mm) is reachable by all nozzles. I tried to be very patient and even offered ideas as to how they could improve the design and they refused to return my money. I still haven’t received the filament I ordered and even though I asked I still haven’t been refunded for that either. What a waste of my hard earned money.
          STAY AWAY FROM ORD SOLUTIONS. THEY WILL RIP YOU OFF!!!” http://3d-printers.ireviews.com/ord-solutions-rova3d-version-2-review

          Here’s a video of a lady fixing hers, it looks like the bed actually fell off during a print. top notch. https://www.youtube.com/watch?v=hDoMxlIuX5s

          The ones I’ve seen are nice enough in person (completely ridiculous, but okay I guess) , but I’d like to see a complete mechanical redesign before I’m anywhere close to being impressed by or recommending this printer.

          However, it was the only printer on the market that had a vaguely equivalent number of extruders to match the mosaic. So I used it in this article and only poked, what I thought was, light fun at it.

          1. Hi Gerrit, Jenn here from ORD Solutions here. I just wanted to comment to say while we do have a few customers who have not liked the printer who have posted these one sided negative reviews, we have many many more happy customers who LOVE their RoVa3D printers. In fact Katherine (the lady who’s video you posted) is one of those customers, and recommends it often. You can see some of the things our customers have said here: http://www.ordsolutions.com/customer-feedback/ (Including a review from Katherine). Anyway, everyone is entitled to their opinion, and we can’t win everyone over. However I do want to say thank you Theo for your comments :)

            We HAVE listened to our customers feedback and suggestions, and we do have a new printer coming out which we we have completely re-designed, which includes our brand new full color blending head. We are launching it this summer so you will need to stay tuned until then and hopefully we will be able to change your mind about us. For anyone wanting information about our upcoming printer you can sign up to get notifications once is launches: http://www.ordsolutions.com/kickstarter-coming-soon/

            Thank you,

            Jenn Gibson
            Vice President
            ORD Solutions Inc.

          2. Gerrit, You’re comparing an entire 3D printer design, to an accessory that meshes plastic together, all in the wake of throwing your invalid and biased opinion out there.

            “tl;dr It’s all marketing wank with over-the-top useless features that no one I know uses”

            “…Rage…”

            Quit while you’re ahead.

            It’s very amusing to see someone have such a tantrum.

            There’s plenty of other printers out there with different specs/price. There’s Rostock, Ormerod (which I own both), Mendel, Huxley, they’re all different machines in their own light. If you can’t see the point of why a company/person designed something just ask them. It appears the rep from ORD was kind enough to look past your childish reaction and respond.

            Seriously if you’re finally doing research about that product, write an article about it comparing it to other printers. Do your own tests considering you’re the expect writer here, with all the experience. I merely pointed out the fact that you were comparing apples to oranges here, and you decide to go on the internet and find the only bad reviews about a product. You’re probably a really smart guy but honestly the way you’re reaction to criticism is alarming, even worse, you’re pulling down the integrity of the other writers by saying such hot garbage.

            If I really wanted a rise out of you, I’d just say, you’re article stinks. Your not funny or witty with your backhanded comments to criticism, and frankly, drawing comparisons between two noncompeting and different products is wrong. Its damn irresponsible to drag someone else’s work through the mud like that based on a personal opinion.

          3. Jenn,

            Great! I look forward to it! I obviously can’t say that I like your printer after my rant, haha. It’s not that five extruders are necessarily 100% bad, or that cool muscles wouldn’t be great in a proper assembly. I like what you’re attempting.

            Likewise I will clarify that there’s no reason not to buy the ORD Rova3D if five extruders is what you need. There’s nothing about it that will make it not work. It will be comparable in print quality to most printers. It just could work better.

            I look forward to seeing your upgraded version! I would love to revise my opinion. Hopefully I can see and use one in the wild:)

            Thanks,
            Gerrit

          4. Theo,
            I don’t think that my comparison is unfair. They are certainly competing for the most common use case which is multi filler and multi color printing for similar plastics. I think if you want more than two color printing the ORD or the Mosaic are likely your routes. There’s also the Chimera from E3D that fills a similar need. I also covered the things Mosaic couldn’t do in the article and why you might want a multiple extruder set-up.

            In my defense, they weren’t the only bad reviews I could fine, they were the ONLY independent reviews I could find. (Excluding the customer feedback, which as a principle can’t fully be trusted.) I googled for a bit.

            Thanks,
            Gerrit

          1. I did.

            “If we make a quick comparison between the two solutions we can see some advantages and disadvantages for both. The biggest disadvantage for the Mosiac is that the materials have to be fairly close in properties. For example, it probably wouldn’t work to splice polycarbonate and Ninjaflex together. A multi extruder set-up allows for more variation between the filaments, as well as different extruder set-ups entirely. For example, it’s entirely possible to have a .4mm nozzle for printing the body of an object and then have a second extruder with a .2mm nozzle to put down the fine-detailed text on top of a print.

            The biggest advantage is the integration with the printer’s existing extruder set-up. Consider a delta-printer. In order to print nicely it typically requires a bowden set-up for the lightest extruder assembly possible. To add five nozzles to the end of a delta would have terrible results. It would vibrate and print terribly with lots of ringing, and eventually, destroy itself in the process. Mosaic simply sits in front of the extruder-cold end, where the filament usually comes in, and does its thing. This holds true for a regular printer too, the lighter the extruder the better.”

  4. Why Mosaic decided to not do any precompliance tests?
    They even write in the blog “We believe in the importance of preparation, planning, and mitigating risks.” The first thing everyone ever involved in any form of compliance and certification tells you to do some measurements as early in the design as possible. You basically never ever want go to the lab with your ready to sell, finalized product and hope/pray everything will work out just fine.
    Experience shows that in most cases, you will have some EMI/EMC problems no matter how hard you try to evade them. Of course, if you’re lucky you can make some quick fixes with BOM changes and maybe adding some (expensive) clip-on ferrites to the cables solves immunity or emission problems. If you however have to redo your layout again because you need a passive filter or just a larger bulk capacitor on your PCB, and you realize that the required filter is going to be quite large, you now don’t have enough room on your PCB, and not enough space in your case to fit a larger PCB…
    If you already see low frequency radiated emissions almost 10dB above industrial limits, my guess is that the conducted emissions <30MHz are way to high too, and the lower the frequency, the bigger the filters usually get.
    I really don't see how they can even be shure the device violated the standard, if they have not done frequency specific average and quasi-peak measurements of some of the peaks that show to be above the limits in the peak emission graph.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s