Wednesday, February 10, 2010


Full Colour RepRap?

This is an idea that Erik, Batist and I had at FOSDEM.

Both Erik and Ed have now got Ed's Bowden extruder idea working, and Vik has now got a really simple, really slim extruder heater design.

So it ought to be possible to arrange three of Vik's brass tubes in a cluster at 120o to each other with a single heater and thermistor, and to splay out the top ends slightly to interface with three Bowden tubes. If the bottom ends came together in a small manifold block with a single nozzle, and you were to put red, yellow, and blue filament down the three tubes, it should be possible to get the three steppers to mix any proportions to give you any colour you want, I think.

The trick, of course, will be to have the fine nozzle as short as possible, and the mixing happening at the last moment in it.

The flow is very viscous, and so Reynold's number is very low. This means laminar rather than turbulent flow, and therefore maybe streaking rather than mixing. But if the head were to twist back and forth, that might fix that problem.

RGB Colour is for light, it is not applied to soild colour. You have to use the complimentary CYMK (4 colours).
That's why I said RYB, not RGB. I seem to remember that RYB worked for paint when I was a child (which, admittedly, was sometime in the reign of George III...). CYMK might be better, I agree, though 4 would be more fiddly mechanically.
it would be better if you could heat the colors individually and mix then in the extruder head.
RYB was used up to the renaissance before they understood colour theory and realised that it has a hugely limited colour space compared to CMYK and in fact can't reproduce most colours. Typically, it has since been taught in schools as the subtractive colour set and is used nowhere else.
I may be a little off base here... But Won't you need a white base for any of the additive methods to work? It isn't like you have white paper to begin with.
I guess so.. we'll need 5 filaments then- CMYK and white for full colour
My team and I had this idea about a month ago along with several others to implement color printing. We are planning to test several of them out so we will of course post something if we make and headway
As a comparison, "chemset" ( by ramset) make an industrial 2-part adhesive that is thick like toothpaste. It uses two separate supply canisters, and a mixing nozzle. The nozzle is over 10 cm long and has internal flutes to maximise "stir" of the two elements. If you are thinking of making a version of this that uses 3 or 4 elements, and does it at over 150degC, I think you are mad. Go for it! :-)
sweet idea
Yes. I agree we need to have CYMK to get the full color range but it has to start somewhere and the idea itself is really really cool. I believe it may be less complicated too as you would not need to change extruder to change color. I really hope this get's implemented. Good work guys.
Interestingly, you can use the inside of the parts to flush out unwanted mixtures of the part, and print the outside when the colour is consistent.
An other way to do all this is by adding CMYK pigments to the plastic input, preferably late so that it 'responds' quickly, but not to late for adequate mixing to happen. I've noticed that you can go from light colours to darker colors with quite a sharp transition, but the other was around requires more 'flushing' out the old pigments, as can be seen here (multi-colour whistle):

Note, the Bowden extruder also allows you to measure out different lengths of differently coulored plastic and put them in the PTFE tube (resulting in prints like the picture I mention above). You can print a multi-coloured object that way without intervening during the print, they the transitions will be mostly limited on a layer by layer fashion.
When I've painted my room I didn't be able to reproduce a lot of color only with CMYK. The real color paint are made with a lot of different pigment. The CMYK is good for printing but is always an optical illusion of color. I hope you can find a solution! Maybe I'm get wrong ^^
@ lerrigatto: You're right (I learned about it when discussing it with someone who mixes paints on a daily basis). You need even more to be able to create all possible colours, but we don't need to in the first go. ZCorp use CMYK on white powder. This also limits the possible colours. Of course, if you use PLA, which is not white but clear, you're able to create even more colours but you need pigments for those.

Multi-material, whether you use if for supports, super strong composite objects or colour printing, it is simply a very important direction to take (in addition making what we have more reliable, user-friendly and cheap).

With such a setup, being able to create gradients of LDPE and HDPE would also be very interesting. Of course the temperature's need to be managed independently if materials vary more.

Or unfold's ceramic deposition, combined with plastics to create very rigid composites... RepRap is only just getting started!
I found (when switching materials) that surprisingly HDPE and ABS seem combine well to make a range of material with properties in between the two. They are compatible extrusion temperature wise as well.
While using CMYK and White materials will give you a good color range, the mixing problem as well as managing 5 extruders seems pretty challenging. Why not just mount an inkjet CMYK printer head next to your white extruder head and just spray the color on the surface of the plastic.
Side idea. would there be any benifit to using the three tubes with one color, but three different nozzle sizes. You could then bigger fills, etc.?
Regarding the concern over streaking rather than blending, if you angled the nozzles in towards the centre first (at a sharper, near horizontal approach rather than all pointing virtually vertically), injecting into a final short combiner nozzle to pump it out, this might give the colours a chance to push into each other and blend before coming out, so it's less like Aquafresh toothpaste.
I like the idea of feeding thermoset or thermoplastic powder coats into the head. Most will gell and flow at 180*f, they come in 10,000 different colors, and they may add significant improvements on strength (thermoset epoxy's are harder than aluminum.) Powder coats are available with many special qualities as well such as electrostatic disapative, anti microbial, and high heat tollerent to name a few. Just a thought.
First off let me recommend going forward with the RYB extruder rather than the ckmy+white as a first version. RYB will let you produce a range of colors and it is worth remembering that the chance of mechanical failure increases at the square of the number of parts, therefore KISS for the first few generations. You might also consider prototyping this using CAPA\"friendly plastic" both because it is already available in lots of colors and because the low temps involved would be easier to engineer for.

As for the problem of mixing, if you are introducing the fluids tangentially you can run a rod centrally down the length of the tube and spin it, this will create a lot of shear, especially if either the rod or the tube walls (or both) are etched with counter twisting spirals.

Finally it is worth thinking about final version in which the 4 color filaments are made with just enough plastic to carry a load of pigment and the white/natural filament does most of the structural/infill work. Would this mean 2 widths of filament? Also we currently use color to distinguish between multiple materials in the slicing software. Will we need to move to some sort of "materials metadata" format if we start representing color in our models? Are we going to build an RGB>CMYK+W converter into the software or how will the color data be stored and transferred to the extruder? How will the slicing software handle patterns or "skins" which are not usually part of the 3D model per se? I am not trying to be a wet blanket, but the sooner in the process we start thinking about hard problems, the more likely we are to come up with elegant, enduring solutions.
The color "Red" in ink is really "Magenta" - a mixture of red and blue. The "Blue" in ink is really "Cyan" - a mixture of green and blue. "Yellow" really is yellow - a mix of red and green.

With paint (and coloured plastic) you're doing "subtractive" mixing rather than additive as on a computer monitor. Rather than add red, green and blue to black - you subtract red, green and blue from white. But white minus red is cyan, white minus green is magenta and white minus blue is yellow. Hence CMY.

Printers use CMYK (K being "blacK") because the purity of dyes and printer inks means that even when you mix Cyan, Magenta and Yellow you haven't completely blocked all of the light - hence you only get muddy browns and not proper blacks. Hence we use a fourth colour to get us good control of black and grey scales without muddy browns.
Ha, a discussion about colors. being the son of a printer I might be able to add something. As some have already explained the typical 'full color' process uses Cyan, Magenta, Yellow and Black. But in essence there is a fifth color which is the color of your paper, typically white. The colors are mixed in two ways: first the colors are transparent so laying them on top of each other mixes them. but there is no real way to vary the amount of color in each layer or a way to vary the amount of white so only 7 colors are possible. In an actual print the mixing by layering is rarely used. The various tones are created by the same principle the pointillists used called halftoning. The colors places as C,M,Y,K dots that vary in size, small dots are surrounded by white so to make light colors (in your brains that is).
Black is used as steve said because mixing CMY wont give perfect black but you can come very close actually. The other reason is that most text is black and printing that with 3-4 colors is generally not a good idea.
So using CMYK for an FDM process might not be the best model. Zcorp can use this because what they do is actually printing on layers of white material which is similar to 2d printing only you stack them. Since you miss a white base in our process and black can be aproximated I think you need to look at toothpaste and have a system where you have white base material and 3 highly pigmented filaments coming in from the side. CMYW might be better in this case :)
Oh, and I want to vote for support material instead of colors ;)
I found this very interesting article about blending different polymers. The way you stress the mixed polymers will determine the material properties of the immiscible blend:

A must read for RepRappers :)
Post a Comment

<< Home

This page is powered by Blogger. Isn't yours?

Subscribe to
Posts [Atom]