Tuesday, February 21, 2012


Print any colour you like...

...with added stripes.

Some time ago I started to build a mixer nozzle for RepRap that would take four coloured filaments and combine them out of a single nozzle.

This is what I made.  Four 1.75mm filaments are led in by four Bowden tubes.  The filaments can be driven at any relative speeds to achieve any proportions of mixing.

Now Myles Corbett has taken that idea and made it work:

This is his adapted Mendel.  It has two RepRapPro extruder drives at the top leading by the Bowden tubes to the nozzle:

At the moment the drives are being controlled by the same stepper electronics, so they are both running at the same speed.  But we will add an extra electronics so they can be driven in any proportions soon.  He has blocked off two of the four inputs so we are just mixing black and white.

Here is the very first print:

As you can see, it is stripy rather than grey.  We need to increase the mixing in the nozzle.  Of course, molten polymer has a very high viscosity, so the Reynold's number of the flow in the nozzle is very low (and hence highly reversible).  This means poor mixing, hence stripes.  So the next step is to add a labyrinth in the nozzle to force more mixing.

In some ways, more interesting than mixing colours is to mix polymers with different characteristics, such as a hard and an elastic PLA.  Then you could make parts that are rigid in some places and flexible in others with any gradation you like in between.

UAV with warping wings, anyone?

... several hours pass...

Here's the second print.  The horizontal stripes are a bit of serendipity: as I said above, both black and white filaments are being driven synchronously for the moment while we sort out extra stepper drivers so we can rotate them independently; but Myles left the grip-screws on the white filament drive lose, so white was being driven intermittently.  Thus we have various ratios of mix here.

As the filament comes out of the nozzle in two stripes - one black and one white - because of the poor mixing, one side of the object tends to end up blackish, and the other side whiteish.  We might be able to do something clever with this effect.  See Janus particles, for example.

This is a great idea. What's really amazing is that you got it to work so well with so little hardware. I will be watching your progress on this closely.
This is incredible! Being able to print gradients opens up so many possibilities.

Looking forward to more on this. :)
Something else just occurred to me. if you can get a good mix of the colours does this mean that we can print any colour with just three or four primary colours of filament??
Yes - any colour from four (or maybe five - I suspect you'd need both black and white + three primaries) should be possible. We'll need cyan, magenta and yellow, I think.

Get your RepRap to paint your portrait in impasto colours...
Really exciting new use for Faberdashery filament. :)
Exciting times for colourful 3d prints!
Very nice work, full colour here we come!
I wonder if a mixing chamber with a dye injector and natural filament wouldn't also be a good approach.
Possibly, though I am told that mixing concentrated dye into a conventional extrusion process is very difficult - you have to have it already in a plastic carrier anyway.
Amazing!! i think is another big step !!
Could you cause mixing by essentially dithering the colors together? Instead of pushing the white and black together, push the black a quarter step, then the white a quarter step, then repeat.

It might be enough to give an undulating dithered pattern to the mix instead of stripes.

Either way, I'm looking forward to playing with this. Too bad my parts kit is still weeks away.
Science! :) I've wondered for the longest time how mixed colored plastics would turn out - would they actually blend, or just create a rainbow mess? I wondered if you needed to mix dye before injecting, or if that would screw up extrusion. I still wonder if different mixtures have a chance of changing the properties of the resulting plastic such that the melting temperature is affected? :o Glad to see progress marching forward!
Dithering is a neat idea, but I think the reversibility of the flow would make it less effective than one might imagine. One could always scribble with the nozzle a bit to mix on the outside.

Obviously I'd like to avoid having to have some sort of mini electric stirrer in the device...
I guess you would need something like a spiral like those used in the nozzles for mixing two component glue. In those nozzles you can see the glue spiraling through the nozzle and eventually being mixed. I think the joint area will need to be bigger than it is for a glue nozzle, but I guess it will also become a very complex design, and of course, how would you get that spiral machined into a brass nozzle? Create a nozzle through laser sintering maybe, not sure if that would give a good enough quality and whether it could withstand the temperatures required for a longer time. Hmm, interesting challenge.

Anyway, exiting stuff. So now we start to match, and even beat, the quality of commercial machines and perhaps soon we can even do full color…
An idea from Rhys is to use metal (i.e. stainless steel) Luer tips as nozzles in general. This would make swapping easy (to change diameters, for example), and you can get glue-mixer Luer tips. You can also get M5 to Luer adaptors.
Yeah, looks like 5 colors would get you full color. CMYK, plus white to maintain the volume. I'm guessing you could use standard RGB to CMYK equations to get 4 of the filament rates and set W to the top speed minus the sum of the other 4.
Great work, maybe mixing based on vibrating piezo element could help.
I have written code for sprinter that adds two new outputs; extruder steps at 1/2 speed and extruder steps at 1/3 speed. This was intended to drive a triple extruder (see thingiverse 16782) I'm now working on a triple extruder where the active filament(s) are selected by solenoids. The extruder speed needs to be slowed down to 1/2 or 1/3 for color mixing.... I've got the model built, but my latest batch of filament is giving me fits!
I think for 2 colors you need to design a limited slip differential which takes the input power from a SINGLE stepper (say nema17 beefy) motor and delivers it to 2 of the extruders (bowdenesque) and let you control how much % is going from one side of the differential to the other. Basically anywhere from 0/100 to 50/50 to 100/0 or in between. This would be insane if you took it to another level with 4 colors, but my brain hurts too much right now to think about that one.. :D
How are the Bowden tubes attached to the heater block?
This comment has been removed by the author.
This comment has been removed by the author.
Hi, new to this stuff but I am a product designer/inventor @ uni.. Why not use a small chamber(s) to pre-mix the ideal colours from (ideally) primary colours from specially designed CAD data that pre-determines colour? Also a multi-nozzle system could be used, while one nozzle prints one colour the other can be auto-cleaned, then they can be quickly swapped on an axis and so on.. (same with the proposed chamber). Please contact me if your interested in any of these ideas or more! fabianrw@hotmail.co.uk
This is cool. I had thought about some sort of "dopant" injection system with maybe a hypodermic needle going into the center of the plastic stream.
Also, might be able to add a needle valve type arrangement to control how much of each color is added.
This has probably been suggested already but I was wondering how difficult it would be to add the inkjet extruder from an ordinary inkjet printer and the electronics to drive it to the final stage eg hot end of a reprap printer and just use white filament?
It would be cool to mix materials. For example material X and Y could lead to Z. That way it is possible to print a lot of materials with a limited set of base materials (for example alloys).
Post a Comment

<< Home

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

Subscribe to
Posts [Atom]