Ideas much appreciated (Very very rough write-up)

Phase 1: Design of hardware+software

• Extruder-Hot End
• Extruder-Cold End
• Modifications to Extruder Carriage
• Motor Selection
• Cable Management (electronics and filament)
• Software

Phase 2: Analysis of color mixing proportions (calibrate model colors to %'s of each color required)
Phase 3: Improvements and updates

Info needed:

• PCB to support 10-11+ stepper(maybe servo) drivers (X, Y, Z, Red,Yellow, Blue, Black, White, "transparent"[if effective], and support filaments) (mixer may or may not be able to just "run" at same speed, simple DC motor maybe) (support possibly on separate extruder)
• Software to use G-code values for different values of amount of each filament to use
• Servo motors vs Stepper motors (for X/Y/Z vs filament feeder vs extruder)
• Mixing efficiency strategies
• Most efficient XYZ movement method (most common, potter's wheel, or some other complicated setup)

Edited 1 time(s). Last edit at 12/09/2012 01:23AM by CPS.

I think one extruder mixing several materials would not be very useful. The current extruder designs store quite an amount of material inside the hot region, and some of it stays there for quite a long time. This is most easily observed when changing from ABS to clear PLA filament, where it usually takes about 50mm filament run until the last foggy shades of the ABS vanishes. However, it may be ok to seperate the colors less cleanly if several full colored materials are mixed.

Pearl EX has a full line of art-safe color powders. Why not start with powder and mix those directly into the hot end?

@dronus, The purpose of this project is to do multi-color 3d printer, mixing several materials will be challenging yes, but not useless

@Simba, how do you control the flow of powder to the hot end? Also, according to Jacqyard Products: "Works great on paper, clay, metal, wood or plastic and cleans up easily with water." Wouldn't it just get washed out if the plastic part gets wet?

The idea is to used the color mix'ed plastic on the outside of the part and the "partially mixed" to be used the infill and in certain situations, put it in a throw away pile of plastic of some sort.

I'll post a 1st draft cad file as soon as i get decent internet connection (to download AutoCad, had to clean out computer awhile ago), probably later this week

Really important info I need help finding... Board to be able to use so many motors, software to accommodate using more than 1 color at a time and transfering to G code, possibly usb/SD connectivity

Will probably need to test filaments individually for color calibration

err why can I not edit my first post, it is kind of important to be able to update it...

Edited 2 time(s). Last edit at 12/11/2012 03:09AM by CPS.

any1 ideas?

Multi-color printing is very much in its infancy in the Reprap world, so there is not much to build on. Adding extra extruders means adding an extra bit to each part of the tool chain. Tools that are designed for monocolor will most likely need an architectural change.

There is a chicken and egg problem, I was encouraging OpenSCAD to support multi-colour, but the general response was "it is not needed, no-one uses multicolor". Of course they won't if the tools don't support it !

Adrian Bowyer is doing work in this area, you might find this link interesting [reprappro.com]. Haven't found a lot of details about it yet. Adrian is using multiple hot-ends rather than a color mixing hot-end. Multiple hot ends is somewhat easier, because the job can be split serially, so not much change is required to the G-Code for example. Color mixing would mean allocating extra axes for each extruder if they operate in parallel.

I think you will need to do a lot of original development, and may have to work with some cobbled together tools until such a time they support multiple colors natively. I've seen a couple of other people doing multi-color, I don't have the links to hand.

I think the place to start is with the mechanics though. Adapting some electronics and firmware to drive it would not be too difficult. I might even have a go at that. The CAD -> slicer route is more of a challenge I think, but perhaps if the basic machine can be shown working it might generate some interest in the slicer side.

Someone did some experiments with coloring the white filament with whiteboard markers, and it seemed promising. Mark the filament before you heat it, but the software needs to estimate accurately how much filament it is using, so it can color it ahead of the time it will need it... what do you think?


Or what about injecting dye? of course you have to heat it so it won't steam and make bubbles...

The idea I had to achieve a full colour print is to combine a conventional ink jet printer and a reprap.

After each layer has been printed you go over it with the ink jet printer to colour in the edges of your part. The extruder and print head share the same carriage.

I think the easiest way of doing it would be to keep all of the normal printer electronics. The wires to the printer motors would go into some "interpretation" electronics which would translate the pwm drive signal into a movement speed of the x and y stepper motors. These electronics would then send back a series of blips every n steps to simulate the encoder wheel normally found in a printer. This would result in a closed control loop and you wouldn't have to create a printer control system.

So the whole process would go:
print a layer in white plastic
Home the printer to an exact spot
Send an image to the printer to print
Enable "printer interpretation" and do what the printer commands
Print another layer and repeat

I believe this is how sls printers produce full colour parts. It's not too hard to put your own inks in a print cartridge which might stick to plastic better, although they might shorten the life of the print head.

Having looked in to thise myself before, I agree with bobc, while there have been a few other discussions, Adrian's work seems to be the most hopefully and the RepRapPro site has been promising multiple-head support for a long while so hopefully they will deliver ;-) The recent video looks hopefully.

Some of the other more commercial efforts do have dual heads e.g. [forums.reprap.org] and some others I can't recall (I think sumpod has a dual head option?) but I haven't actually seen that much info on how well the dual head works (but never looked that well) and they don't seem to have added much to the general RepRap community (I don't think most of them actually released much info). My hope is given Bowden's involvement and the Mendel legacy, the multiple head support should find its way in to the general RepRap community particularly in the Mendel lines once it's actually properly released.

P.S. To be clear, even Bowden's work isn't likely to give full colour printing in the short term. I'm resonably sure mixing won't work well enough to be used to reliably give different colours. And switching between colours sufficiently to produce any resonable degree of dithering except at a very far scale simpy won't be possible. So all we will really have is 3 colour printing (or multi material printing). Perhaps it will be possible to add 1-3 more heads, but that still will be quite far from full colour printing.

Edited 2 time(s). Last edit at 12/18/2012 09:52AM by Nil Einne.

What if you could do something like "tool change" that most CNC mills use, except instead of changing tools, you would be changing spools (hey it rhymes!). Basically print something in one color, move the print head out of the way, get rid of what's left of the current color somehow (either cut it and push it through, or something else), feed in the new spool of a different color, prime the nozzle, and then continue the print.

Take a look at this to get an idea of what it would look like [www.youtube.com]

I can forsee this requiring some extra work on the extruder mechanism, to be able to have it load spools automatically and reliably, as well as managing the various colors, but surely it could be done.

What do you think about this:
[www.reprap.org]

CPS Wrote:
-------------------------------------------------------
> What do you think about this:
> [www.reprap.org]
> ColourMixingReport-jmc.pdf

It's a great read. The conclusion seem to be that an active mixer is required; and that getting multiple feeds plus a stirrer into one heater block requires some careful design.

Perhaps getting good color mixing is a step too far, considering all the dvelopment work required. It might better to start more simply and develop all the parts in stages. Richrap has done some good work with color blending, I think that might make a good base for further development.

In the short term, it would be great just to get some selectable colors out of one nozzle. I think that multiple feeds onto one hot end is looking like the most practical approach. Richrap shows that multiple steppers on the X is doable, but I think for anything more than 3 feeds then Bowden feeds are essential.

I am wondering if it possible to have a stirrer inside the hot-end, it would be driven by the filament coming in. This might not give the best mixing, but would remove the need for external motor and shaft seals.

Maybe a mixing structure like those used for two component syringe dispensers (like [www.sri-supplies.com] ) can be replicated in metal, maybe by milling. Like the cylindrical part of the hotend above the nozzle replaced by two half shells with a milled rectangular labyrinth pattern inside. Basically it forces the fluid to fold onto itself for several times thus give a decent mxing without active stirring. It may be shorter then for two component adhesives as those need to be extremely homogenous.

It seems to me that one approach worth investigating would be embedding the plastic with color-changing microcapsules. (http://books.mcgraw-hill.com/EST10/site/supparticles/Color-changing-inks.pdf)

Not sure if the appropriate chemistry has ever been isolated, but ideally three photochromic capsules (RG would be embedded in the ABS, each responding to a different specific wavelength. If this mixing were accomplished, then the color could be lased on after all of the plastic is deposited. Admittedly this approach would almost certainly require some chemical investigation that goes beyond "simple" engineering, but many of the pieces of this do seem to be in place.

That sounds expensive lol, wouldnt affect the properties of the ABS though?

Similar to the 3-color 3d printer, how about, for the 3 different objects (5 in this case, 1 for each color), but lower the densities of the objects (instead of 1g/s of red for full density, if u want 85% red and 15% blue, for that same area, have the stepper drivers spew out 0.85g/s of red and 0.15g/s of blue)

Allow the mixing objects to be concentrated error free on the visible part of the part(outside), while the infill be the transition between the 2 colors (lets say 85% red +15% blue to a 20% red + 80% blue)

Anyone know of any ways to put so many drivers for electronics (could you just make a pcb to contain them all, and maybe use computer for processing if necessary?)

Let me know if I have to clarify, i just typed this up real quick before i went to bed so i wouldn't forget

To very simply answer you question, almost all colored material, especially plastics is made with powered particles, not dye. Dye is necessary for transparent materials though.

To mix, add into a pellet extruder (ha!)... for filaments, I'm not sure how you'd feed it, just an idea .

All white board markets use powder color filler (not dye) FYI ... this is why marker (dry erase) can never be cleaned from clothing...it has color particles that latch onto clothing. Great for mixing with plastic idea!

Shed Wrote:
-------------------------------------------------------
> It seems to me that one approach worth
> investigating would be embedding the plastic with
> color-changing microcapsules.
> (http://books.mcgraw-hill.com/EST10/site/suppartic
> les/Color-changing-inks.pdf)
>
> Not sure if the appropriate chemistry has ever
> been isolated, but ideally three photochromic
> capsules (RG would be embedded in the ABS, each
> responding to a different specific wavelength. If
> this mixing were accomplished, then the color
> could be lased on after all of the plastic is
> deposited. Admittedly this approach would almost
> certainly require some chemical investigation that
> goes beyond "simple" engineering, but many of the
> pieces of this do seem to be in place.

I think the idea of a post-activiation of color makes the most sense, followed by covering plastics with an inkjet to transform one color into color. You'd loose the least material properties this way, have the most color options, low additional material amounts, etc.

The color changing dye would be tricky. Obviously, if it responds to light (laser) then it needs a certain period after which it wouldn't change color. Such a material is likely to photobleach over time, but its a start.

... maybe some years to go, but resembles one of the ideas out from Eric Drexlers books:
Hollow micro-spheres, embedded with a 'nano-robot', consisting of a bead with different colours on the sides and the ability to rotate by presence of intense (laser-) light, so it will show the colour represented by the laser-wavelength.

So any surface coated with such microspheres will change the colour, when illuminated with a RBG-laser ... maybe several times, for recolouring by wish

---

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

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

what happens if some1 pokes one of the objects u made with a laser, and it all turns 1 color

yea, but nano robots are out of the scope of reprap especially now

Edited 1 time(s). Last edit at 02/28/2013 11:01PM by CPS.

CPS Wrote:
-------------------------------------------------------
> what happens if some1 pokes one of the objects u
> made with a laser, and it all turns 1 color
>
> yea, but nano robots are out of the scope of
> reprap especially now


We call that a 3D etch and sketch

... it's not so far away - first 'recolourable' (but passive) microspheres can be found in e-paper: [en.wikipedia.org]

Other types with colours and activating or (re-) colouring by magnets or ultrasonic were tested too ...

---

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

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

how about a 3d 100%(only seen 85% but 100% is possible i guess) transparent oled printer with a resolution of 8kx8kx8k, a 3d printer that prints true 3d tvs

Any ideas for going about the pcb board that can handle so many motors and how to integrate it into software?

Great idea. but any idea how to route the city of invisible wiring?
Also, most plastics are ~92% transparency. So if we print acrylic, you can make the center pixels just a little bright than outside to compensate. now that is cool

who is printing oleds that actually work as an array (beyond just blue/green light)?

most of the major tv manufacturers print oleds (that is how they make them), they use a 2d printer, just look up on youtube

Sounds to me like the easiest way would be with a single nozzle fed by several coloured filaments. this would save the complexity of head changing,

If you used a square or hexagon nozzle block and fed the filament from the side you could tap the 4 or 6 sides and feed in. The heater element placed on the top to directly heat the area just above the nozzle. You could vary the extrusion speed of each of the bowden extruders to change the colour. It would be best to minimize the cavity size just above the nozzle so colour change would be rapid, but have sufficient volume to allow a good melt.

First test would be a two coloured nozzle feeding black and white to see if greyscale could be achieved.

Already ruled out nozzle switching (Except maybe for the support)

The 2 color mixing has already been achieved by someone else. The biggest problems are the software and the drivers for all the motors.

VRML is slated to take over for these reasons, though I dont understand it.

the biggest problem is time/filament to achieve mixing levels needed. For a real color print, i.e., not one color per layer, you are going to have to turn on a dime, i.e. mixing in only a 1 mm^3 area, not 100 mm^3 or wherever that is at now.

We have to think longer term too. A FFF model with 6 filament colors is truly awesome, but the holy grail is a system more like reprap+ a 3 inkjet nozzle of sorts. Inkjets simply work, and they can hit the accuracies and low volumes we need. The question becomes surface integrity and bond strength to the plastic which seems suspect that it would rub off...

For most prints it is not needed to "turn colors on a dime" because there is a lot of infill that you simply don't see. For rare cases where there simply isn't enough room, a "waste" structure can be built. The waste structure doesn't have to be waste because it can be anything else, such as infill for another part that you are printing.

Yes inkjet would be ideal i suppose but if you are using white filament and you are making a multicolored part, you will be spending a ton of money on the ink.

Another use of the mixing nozzle is the addition of multi-material and color mixing of magnetic/conductive/etc filaments