3D Colour Mendel Printer Idea

I'm not very good at building things but I normally good at thinking of ways to do things. I have an idea for colour in the Reprap project printer. I have uploaded a crude picture of an idea I had that I don't think would take much to implement even on a board with one extruder control. If you placed a number of extruders in a circle eg on a gear and had each one a different colour (like an ink jet printer) each one having a coloured filament going into each. You could then rotate the extruders round to the active slot. As the required colour came into the front position where the single extruder would be, then a connection is made so the extruder would feed its filament to be extruded. Now each extruder could be kept at a certain temp to help keep the speed up. The size of this setup would not be to big and only a wider axis support may be needed but there is room I believe for this already. It wouldn't be to much hassle anyway in order to have colour.

It would be good to have this working on someones printer, for me I am not able to probably build this unless I got step by step instructions. I got all the tools but no idea haha.

Anyone got any further ideas to possible make multi colour for the reprap mendel in one single printer possible.

Michael

a rack and pinion setup may save space on the x axis. have the rack going parrallel with the y axis., this way it wouldn't change the profile in that direction and the carriage could move the same amount. But overall, i believe your idea is a pretty good one. well done

I do think that some kind of tool changer will be the way to multiple [color, material, nozzle diameter].

A machine like Ultimaker has room for 4 nozzles at the print head (could be modded for more), but there are problems. The nozzles would have to have the same Z elevation within a few microns, there would be multi [color, material] ooze, and build volume is lost. You also have to use a separate Bowden drive for each hot end. The BfB 3000 mounts three complete extruders on the print head. Some careful adjustment could get all three nozzles matched in Z, but how do they manage the ooze?

With a tool change, you can wipe the ooze and adjust for nozzle height using the Z axis. One traveling drive could swap hot ends, which would result in fewer components. The drawback is that the print head must carry a mechanism for releasing and engaging hot ends and filament. One such print head could handle any number of hot ends, though.

I also favor a rack over a rotating turret. With a turret, you need to add a mechanism to rotate the selected hot end into place. With a rack, you can use the X (or Y) axis for hot end selection.

I wouldn't want to implement this on a Mendel though. The rack of hot ends would need to travel in the Z axis with the print head. This would all be much easier with the built platform moving in Z, and X and Y axes staying at the top of the machine.

There seem to be four areas of needed development. The first is how the print head should mechanically engage the hot end and filament. The second is electronic support for that mechanism (if active mechanical components are required). The third is software support for the tool change. The firmwares already support multiple extruders, I think. Software to handle the tool change needs to be added. Last, Skeinforge needs to handle multiple colors and materials.

I can see how a rack would be good, I remember a plotter that used to swap pens by going to an empty slot and then picking up a new one. the extruders would still have to be heated when static as to stop the warm up delay. With Just one extruder working at once would there need to be any worry about how many the firmwares support. I don't think the size of a cog of 6 extruders would be all that large. I also reckon that it would be a simpler design and simpler to implement as the extruders would step round at certain intervals and would all be at a fixed height so no risk of slippage.

Any way that is done, I will certainly attempt. I just wish that I could do it myself I will be trying to figure out how to do it (May take multiple life times)

Tool changing is a good solution for multi-material printing, like switching between plastic and support material. But I don't think it works well for colour printing. The main problem with this approach is that even with six different nozzles, you can still only print in six different colours. Even a crude 16-colour printing would require 16 nozzles, and 16 different spools of plastic... The problem is that there's no way to mix or blend colours with this approach, except maybe by depositing tiny beads of one colour and then change tools and deposit tiny beads of the next colour and trying to re-heat them to fuse together on the printed part, which would be painstakingly slow.

So I think that colour printing has to be carried out at the filament level; that is, one nozzle receiving a feed of rainbow-striped filament, where the stripes correspond to where the colour changes have to occur in the model. That way the required colour of plastic can be blended together from primary colours. The colour mixing station would need to be able to either take in granules or filaments of various colours and combine them into a newly extruded rainbow-striped filament, or to take a white or transparent filament and add pigments to it as it goes through.

Edited 1 time(s). Last edit at 08/08/2011 01:14PM by jbayless.

In theory you can archive some color printing ability, meaning you print different areas in one color and be done with it.
Dithering on the other hand cannot be archived, so this solution will not be "true color"

Anoth solution would be if you adapt a inkjet printhead to fit on the printer. This way the top of the outline can be colored each layer. This won't be true color still because the maximal tone can be 50%, so you are limited to pastel colors but you get a continuous color spectrum.

I've been thinking about this, and agree with jbayless; multiple heads only give a limited palette of colours in FDM. Perhaps another option for a range of colours is to mix different-coloured filament as close as possible to the point of final extrusion. I imagine a single hotend fed by small filament (perhaps 1.75mm would be small enough) with cyan, magenta, yellow, black and white filament (we're dealing with reflective/additive colour rather than transmissive colour, ie RG. The filaments are fed by separate extruders, so that they can be fed in at different rates, and mix as close to extrusion as possible, so transitions between colours happen quickly. This would probably mean the extruders and hot end would be mounted on the Z-axis, with the table mounted on the X and Y axis. The gcode could send something like:
G1 X10 EC0.75 EM0.25 EY0.3 EK0 EW1 (you would need a lot of white to make lighter colours)
for the colour mix required to the electronics, which would then extrude the correct proportions of coloured filament; perhaps a short purge would be needed between colours. Also, this would mean slow tool changing would be avoided, and only one hotend needs to be heated. Obviously, the hotend would be kind of special... it would need five channels down it, with a small chamber at the end to mix the colours before extruding. Perhaps this needs patenting!

This is the simplest method I can think of, and yet the 'inkjet' style 3D printers do this so much easier, and with higher resolution. And with built-in support material! Are there any plans around anywhere for homebuilt 3D inkjet printers?!

I thought of the 6 extruders as a starting point that can get us started by printing bits in the 6 different colours. For plastic extrusion printing it will be very hard to get true colour. I accept for the mendel that true colour is not going to be possible for some time and may not even come about but basic number of colours would be very cool. For true colour 3D printing I would love to have a go at an open source Selective Laser Sintering Machine with true colour.

Does anyone know of anyone working on an idea like this ?

I think there is still a valid patent on the color powder printing tech, issued in 2000. Not sure how it applies to sls, but just so you know.

Patent 6165406

Off the cuff I thought that an injection of dye would be the most viable method of coloring. Even something as indirect as adding a sharpie marks on the cold filament at a point before the color needs to change at the hot end outlet. Calculating something like that isn't going to give very high resolution but would allow for colored layers or block area. Taken to the extreme with red, green, and blue markers with variable pressure could give a large range of colors.

I'm not up on colored plastics, but don't they mix pigment with a resin to achieve color?

If so, how about using CMYK pigments and a single natural color resin filament. Mix them in the extruder hot end. Just need the single filament drive. Since pigments would be lower volume, might be able to use some sort of syringe delivery mechanism for them, which might simplify the entire head.

The proverbial granule extruder probably lends itself to this more than the current filament designs -- their ram screw already functions as a mixer.

Just another thought on the idea of changing colours mid run... but instead of swapping hot ends, is it possible to use a single hot end, but have a system for changing filament feeds into that one hot end?

You could have the unit stop, move the extruder head to a waste zone on the board, or alternatively mod your unit to create a bin hole in the bed (5mm hole bored in one corner, with a waste tray to catch the filament underneath the printer), have a rotating carriage above the hot end rotate the new colour filament into place, and feed it into the extruder, and just run the extruder for x amount of seconds until the old colour has been worked out, and the new colour is coming through.

Yes, its wasteful, but it would remove the issues with extruder tip alignment.

You could also have a manual system, that simply stops the unit over the waste zone (but keeps the extruder heater running to keep the filament still liquid in the hot end), reverses the unmelted filament out of the unit, and beeps to let you know to feed in the new colour. Hit a resume button, and it runs a waste purge on the extruder to get the new colour loaded up, and resumes printing again. This way, you would only need one extruder head.

One caveat, I haven't got a printer, so I'm not 100% sure on how the purging would go with just feeding a new colour in after a current one, whether it would completely clear out the old colour, or whether you would just end up with a horrid mix of colour for a large amount of time.

Another downside to this problem might be that there is no defined solid/liquid delineation in the filament, so reversing out one colour might require having the end trimmed by hand to create a new clean tip prior to re-insertion back into the hot end.

... depending on the inner geometry it needs sometimes a really huge amount of filament to flush all the previous material - especially with intensive colours you can have until some meters with fading away the old one

---

Viktor
--------
Aufruf zum Projekt "Müll-freie Meere" - [reprap.org] -- Deutsche Facebook-Gruppe - [www.facebook.com]

Call for the project "garbage-free seas" - [reprap.org]

I have read the posts with interest, but cannot get a grip on the purpose. Using multiple extruders for fill material etc. all makes logical sence because it enables you to build more complex and or accurate parts. The making of a part with different colours as cool as this sounds seems to me to be a lot of effort for little pay back, in real terms.
It certainly has sparked a lot of interest though.

You can do multicolour printing with Filament joining this gives really sharp colour changes



I started another thread on PLA filament testing, this has a link to my coloured printing - [forums.reprap.org]

---

[richrap.blogspot.com]

Hi Rich,

have you tried colour-swapping with ABS too?

Seems to be highly depending on the inner geometry of the melting chamber - i've seen some tests with simply adding black filament after wihite, what produced more than a meter of grey tracks until pure black came out ...

---

Viktor
--------
Aufruf zum Projekt "Müll-freie Meere" - [reprap.org] -- Deutsche Facebook-Gruppe - [www.facebook.com]

Call for the project "garbage-free seas" - [reprap.org]

Hi Viktor,

ABS is next on my list to try when I get a few more colours, I can't imagine it will be much different but I have not tried joining ABS sections together yet.

None of the above was done by colour swapping, it's all smaller joined filament lengths in one continuous print, you get sharp colour changes doing that.

I have only tried the swap method a few times, but still only had a small change in mixed colours. I guess it all depends on how quick the change in hot and cold point is on your extruder, when I pull out a length I'm quite sure that all but 15mm is removed, when I purge with 20mm (feedstock) it's almost always changed over colour.

@ JV4779 - I'm going to try the Sharpie idea, that sounds like it may work to some extent.

Rich.

---

[richrap.blogspot.com]

It takes more like 2m to flush my extruder when removing one filament and swapping to another. Do you think joining them makes the transition faster?

---

[www.hydraraptor.blogspot.com]

I get pretty sharp changes but my barrel goes ~3.2mm down to 3mm down to 2.5mm then 0.5. I think if there was any outward taper (nophead are you talking about the performance of the no-compromise extruder) I could see it taking much longer to flush.

Yes my extruder has a 3.6mm barrel with a slight outwards taper before tapering inwards.

---

[www.hydraraptor.blogspot.com]

nophead Wrote:
-------------------------------------------------------
> It takes more like 2m to flush my extruder when
> removing one filament and swapping to another. Do
> you think joining them makes the transition
> faster?

Is that 2m of input feedstock? I can't imagine where it's all getting stuck. It must be pushing down the middle and slowly pulling the sides in as the material advances down the barrel? I always imagined the melted plastic was hotter at the perimeter of the barrel so it should flow faster, but maybe that's why it's sticking to the outside and the harder material is being forced down the middle.

My nozzles are all the same, 3.2mm all the way down to a 0.5mm opening, and I don't have any PLA or ABS over 3.1mm in size, most of it is 2.85mm

I wonder if it's just an issue with ABS, does your extruder do the same thing with PLA? I'm going to have to try that now with ABS.

---

[richrap.blogspot.com]

The reason my barrel is so big is because "3mm" used to be 3mm +/-x and I got some stuff that was oval up to 3.5mm. Now it seems "3mm" filament is getting smaller and has become 2.8mm +/-x. I wish it was called 2.8mm when centred on 2.8mm. I will probably have to retire the extruder because reprapsource.com have told me their plastic is getting smaller. It was one of the few places I could get 3mm filament that was 3mm +/- 0.1mm. 2.8mm doesn't work very well in a 3.6mm barrel, too much backflow.

I regularly swap between PLA and ABS and it takes about 1m of feedstock with white ABS to clear PLA, but green ABS still was visible in clear PLA up to about 2m. I have never tried coloured PLA.

Viscous liquids flow faster at the centre and almost not at all at the outside of a pipe so I would expect colours to mix.

---

[www.hydraraptor.blogspot.com]

The transition likely depends on how fast you extrude material. Also the diameter of the feed stock and how potent the coloring is.

Faster flow-->Smaller boundary layer
Smaller Diameter-->Faster flow

I'm guessing nophead has such a long transition because of the taper. It creates a small pocket of nearly stagnant flow which takes a long time to mix.

Just tested mine. It takes me ~4in of 1.75mm feed stock to go from black to white, but only .375in from white to black.

i was thinking that the head could hold 4 coloured sharpie pens ..or just the ink ..then use white or clear plastic.