No more filament? - Quest for a Universal Pellet Extruder for 3D Printing

Pellets - Let's shake things up!

A few months back I asked the question on G+, Would anyone like to discuss Pellet / Granular extruders? I ask because I have been working on various designs for quite some time with success and also plenty of failures , it's not easy, it needs collaboration and I would like to see it stay open-source and accessible to all at a very low (no) cost. 

Directly using pellets or granules of many different types will be another very big game-changer in 3D printing.

Using all sorts of different materials for 3D printing is good news, it helps people find more uses for the technology and further's the progress of design, manufacturing and innovation.

Since designing various Universal Paste Extruders, it's been a real pleasure to see so many people use, re-design and further extend paste and food printing in so many different ways.

Using pellets and other granular materials for 3D printing are a next evolutionary step in both simplicity and flexibility.

It has plenty of challenges...

I have been working on and off for a while to design and test other types of extruders for granular materials for 3D printing, my focus was initially using Sugar with great success but this year I have also moved onto plastic pellets (resin's) that are normally used for injection molding or to make the 'standard' filament's we already use in home 3D printers.

This whole area needs more development and a focus from the open-source RepRap community, makers and developers to refine a system to use all sorts of different granules and materials for 3D printing.

This Blog post is an introduction / development update of where I'm at, and at least a way to form some prior art statements about ideas and directions for this sort of extruder development in 3D printing.

The point of this blog post it launch this as an open-source community project. The files have been uploaded to Youmagine here, if enough people take an interest and can assist with further development it could be a good idea to have a section on the RepRap Forum or Google+ Group for Pellet extrusion development.

I started out with a focus on sugar - the main idea being to use it as a support material as it can be washed away. PLA sticks really well to sugar (and sugar like materials)

Isomalt is a really nice sugar often used for sugar-craft and as a sugar substitute, it has a nice advantage of being less likely to burn or go brown than normal cane sugar.

It worked surprisingly well when you get a good temperature (around ~155 Degrees C) you can get a flow of Isomalt sugar from the nozzle.

- First, lets get back to building one for yourself. - 

Lots of info below, or if you like, you can watch my video about the project here first - 

It's also in HD over on YouTube.

If you have been with me for a while, you know I like to use as many off-the-shelf parts as possible, so it's as easy for people all around the world to build and enjoy 3D printing. Without the need for custom machining, or expensive equipment.

That does not mean you should not make refined custom / machined versions of my work, I know may people do exactly that, and make a successful business out of them. That's great ! - just remember me when you become rich and famous :) and please do try to contribute back into Open-Source developments and Open-Innovation in whatever way you can. * end of philosophical rant  *

Because I decided to use a 'standard' J-Head Hot-end for experiments with this project, I had to base the dimensions around the 6.35mm hole normally used for a PTFE liner

J-heads now come from a LOT of different sources, for reference all the ones here are from the original designer (Reifsnyder Precision Works) over on hotends.com

I tried a number of different 200mm long x 6mm wood Auger bits from DeWalt and Black & Decker, most were almost identical and operated the same.

A 6mm wood Auger bit fitted perfectly and with only a slight modification I had the basic outline of a granular extruder without using any custom parts other than being 3D printed.

A little bit of cutting and polish and you have a granular feed system.

I found it better to cut off the end of the drill, but position in the J-head does need to be quite specific, you will need to experiment for the best melt and flow.

I drilled out the end nozzle to be 0.65mm. No other modifications to the J-head are required.

J-head MK3 - with Aluminium tape to help with cooling.

Different J-head's can work, I experimented with both MK4 and MK3 - The MK3 worked better.

While experimenting with different materials and plastic pellets I designed two different Motor drive systems - One (Above) using a normal NEMA17 motor and Herringbone gear system - just like a 'normal' extruder.

The second design used a NEMA17 with a Planetary gear box fitted. This can generate a lot more force but requires a very large number of motor steps (20,000/mm) so bear that in mind.

Rotational speed and torque are both important, so I would recommend starting and experimenting with the Non-Planetary version first.

3D Print the parts - I printed all of them in ABS. (and also experimented with metal coatings - more on that later)

Most of the fittings are M4, they should be easy to understand from the models and images here, but you need to assemble in the correct sequence for everything to be tightly clamped together.

Bearings are standard 606 style for the Gear and Drill bit.

M3 fixings for the motor and gears.

The mounting method uses my standard universal quick-fit carriage mount.

Quick-Fit carriage - Many versions available for different machines - allows the use of many different extruder types - all my designs use this method.

Auger should be fixed firmly and be straight down the middle of the j-head - make sure your bearings are fitted tightly - they should be a tight fit and may require some pressure and slight heating of the plastic to get a firm fit.

You will need to experiment with the placement of the Auger depending on the style / type of your J-head.

Make sure to secure the J-head well and remember you will need to cool the J-head body with a fan just in the same way you would when using PLA filament.

Important bit - The design of this pellet/granular trap (shown above) is really important for good operation. I experimented with many different versions / designs and I'm sure it can still be greatly improved.

You should end up with something like the above that you can experiment with. Note all the different colour parts, I had quite a few versions of the design before this one.

Then you need some pellets or granular materials to melt. is a typical selection above and below. Some are round, some sharp shards and most are in cut round pellet form.

I finally managed to source a lot of different materials, but at the start I cut up normal filament into small pellets - crazy I know, but at least it allowed some testing as I could not find many places last year that would sell me small batches of different plastic pellets.

Top shows Woodfill pellets and right shows PET

A friend kindly send me a large selection of recycled material for testing, but as can be seen above (left) they are quite large and irregular sized shards, so as much as I want to use recycled plastic, for this extruder it's not going to be possible because of the screw size. I will be using them for other interesting things - more on that soon :)

By far the easiest way is to now buy pellets from ColorFabb here. I purchased a selection.

And if you look closely at the above image you will see 1.75mm and 2.85mm pellets, so I'm reasonably sure these are actually from filament cut up - maybe a good way to recycle any out of tolerance spools, but regardless it's good to know they are small pellets - ideal for this type of extruder.

The ColorFabb PLA pellets are not the same, they are 'pearls', and unfortunately don't quite work in this size of extruder. I bought a lot of these not knowing the size, so check before buying pellets from wherever you manage to source them. - I will find use for them :)

You can even get Woodfill - but, have not tested it yet.

Fill hopper, make sure you have a fan on constantly cooling the thermal break on the J-head and lower half of the extruder body. Experiment with temperature and speed of stepper motor drive.

You can use the standard extruder retraction to stop or limit the flow ! but you may need to slow it down to under 15mm/sec.

Experiment and play, and please let me know how you get on.

One really important thing - Make sure you run the Auger motor in reverse! - that's anti-clockwise as you want to push down the pellets into the hot-end and not drill them out.


I ran out of time with this project in the summer (2014) due to other projects and shows, so the files being released still need work and refinement to get to a working solution that can process many different types of material in pellet form. I have still yet to try the Colorfabb Woodfill pellets as I didn't want to block the second extruder after already blocking the first one with 'A-PET' material.

This is a very exciting project, do let me know what you think about it, all and any feedback is good. And if you decide to make one or want to evolve the design, please feel free and spread the word.

Thanks for reading, have fun, experiment and play. Also please do let me know how you get on.

Until next time, and have a Happy Christmas / New-Year Holiday / Seasons greetings.

Rich.

- A little notice - 

My 3D Printing work (and video's) are released under Creative Commons CC-BY-SA License.

But - All my images, photographs and words published anywhere are covered under International Copyright law, so please do ask me for permission and remember to state where they came from - me, Richard Horne. - (AKA RichRap)

Sorry to have to put this up, but I see a lot of my images and words being used without my permission for commercial activity, by people who want to make money, that's fine but do ask my permission first and be open about what you want to do with them. I'm cool about most things.

3-way Quick-fit Extruder and Colour Blending Nozzle

Three extruder's and One Nozzle - RepRap material blending

Hello everyone, I'm really excited about this one, I hope you like the results - 

Very first print.

It’s taken a while to get this post written up, but a few recent things have really spurred me on to get it finished.

It’s been over a year now since my first layer selective colour experiments and filament joining, not a month of RepRap has passed that I have not thought about multi-colour/material 3D printing.

I started work on colour mixing earlier this year and had some promising results with a basic, manual setup. Things moved onto other 3D printing developments and I got sidetracked for a few months.

Dual extrusion has been the focus for many machines, but this has some significant problems of alignment and oozing of the inactive hot-end and the fact you cant mix/blend materials with two separate hot-ends.

Background - 

I always wanted a single hot-end with multiple driven feeds, so that’s the focus of this particular development.

Before I forget, the files for this project are up on Thingiverse here.

In recent months we have seen some more movement with filament mixing, Adrian Bowyer and Myles Corbett have done ground-breaking work on both multiple fed hot-ends and mixing of plastic filament. The final Report Myles submitted recently is an essential read, great work indeed.

Another wonderful advanced example of layer selective of colour printing has been this set of stunning 3D printed iPhone cases by Matthew Bennett (norcalbarney).

Photo and 3D printed iPhone cases by Matthew Bennett - Thingiverse

Seeing the work by Myles on the RepRap Blog and the prints by Matthew spurred me to re-start and finish off my own rather elaborate colour blending extruder development.

Stop-Start -

I had previously stopped my development because I was having some issues getting two colours to mix well in the nozzle; I introduced mixing slits and a spiral in the PTFE tubes (in and after the PEEK block) and added steps in the nozzle barrel, it had reasonable mixing results but was very tricky to make up by hand.

The same month I saw that Myles had exactly the same issues with the black and white mixing I decided that not mixing the colours well would produce more interesting effects and give more artistic looking prints. So during #30DOC in June I restarted the development and added three feeder tubes. I had initially intended these to be separately driven Cyan, Magenta and Yellow feed, but after some testing realised that all sorts of blends could be made by running single or multiple extruder's at intervals separately or together.

That’s why I’m not calling this extruder a colour mixing setup. I’m opting for blending together 1 2 or 3 materials to produce some really interesting effects. 

Unfortunately my Antique Lathe died in February, which is one of the main reasons this multi-head hot-end didn’t get further development, but with a bit of work I managed to do everything I needed (all of the following below) with just a Pillar drill and hand-tools.

Introduction Video Part 1 -

Building the Extruder's (X3) - 

I’m using a custom variant of the Huxley sized quick-fit extruder from June’s 30 Days of Creativity, along with Herringbone gears and push-fit Pneumatic connections on the bottom.

Many thanks to RepRap Pro for supplying me with the M6 Hobbed bars, these really are made so well and work perfectly with this extruder as it's a modified eMaker Huxley variant.

Quick-Fit 3-way adapter - 

I was going to design a more complex bracket, but I decided to keep it simple for the first test, and as I had plenty of room on my MendelMax X axis, this was the simplest design to test out the idea.

I know! what a monster!

It’s designed to fit on the Quick-Fit Xcarriage.

3-Way Hot-End - (Version 1 - ... it dies a bit later on, but bear with me, it gets better with V2)

Without a Lathe, I decided that I needed a simple hot-end that clamped together; this would allow easy manufacture and experimentation.

First steps were to make a simple over-sized heating block and using a step-drill I bored a stepped hole in the middle, 4mm at the bottom and 14mm at the top.

The nozzle is a M6 brass bolt with a 0.5mm hole, this is cut down to 8mm long and held in place with an M6 brass half nut.

The filament feed tubes would be held in with a block of PEEK clamped down onto the heating block. 

And clamped to the hot-end with a Stainless steel plate.

 Nice and simple construction.

The PEEK block has three holes (you could have more), drilled on a slight angle so they spiral into the stepped heating block. It uses 4mm PTFE tubes to allow the extruders to be angled around the Quick-Fit X-Axis plate. 

I was aware this was not going to last a long time, but I wanted something simple to test the idea with actual driven extruders first.

I was expecting rope seal adhesive to be a good way to bond the PTFE tubes into the PEEK, but this turned out to be a very bad plan. Maybe my Rope seal is just no good, I still have yet to get any reasonable results using it. Nophead however uses a different brand for bonding his heating resistors without any issues at all, so I may revisit this compound in the future.

Blending nozzle V1, assembled – the Stainless plate is clamped to the heating block using long M3 Stainless screws, the PEEK has some Copper tape wrapped around it to aid cooling.

This only has a single 6r8 Resistor, so takes a few minutes longer to get to temperature than a ‘standard’ sized heating block, but gives a good stable performance when insulated up.

First manual test, very easy to push the separate filaments by hand, this might just work!

I finished the manual tests and mounted it all on the printer, ready for test and calibration.
Testing each stepper driver, extruder, hot-end heating and thermistor feedback took about 1 hour.

I got it all working and 'calibrated' well enough for a first print, when... the hot-end assembly died. :(
It did print the striped heart box shown in the very first picture in this blog, then slowly the tubes started to pull out of the PEEK block.

The Glass rope adhesive was soft and had allowed the tubes to be pushed out, quite a mess.
I really think my glass rope adhesive is bad or out of date, it only ever seems to get semi-set no matter how long I leave it, even at elevated temperature.

Never mind, lesson learnt and I know exactly how I'm going to make the next one -

3-Way Hot-End - (Version 2 - ... the one that didn't die! and worked gloriously well)

The heating block and nozzle worked perfectly, so I'm re-using that. I just need to make a much better PEEK and tube holder -

Start with some raw materials, 4mm PTFE tubes, PEEK and Aluminium bars. I'm saving up for a new Lathe, but it's going to take a while to get a new one. So at this point I still only have a pillar drill and hand tools, here goes -

PEEK was cut, drilled, then filed and sanded in the Pillar drill to get a domed plug.

Aluminium bar was also cut, drilled and grooved (with a hack saw!) all rotating in the Pillar drill. (who needs a Lathe!)... - I still do actually :(

Holes are tapped M4 threads.

Use a pencil sharpener on one end of each tube, just cut a little.

Picture shows, normal tube at the bottom, lightly sanded (500 grit) middle, top is threaded (see below)

Before you try to cut the thread on a PTFE tube it VERY important that you lightly sand the outer surface of the tube or you will never be able to grip and turn it in the die!

It's quite easy to cut the thread, you can use a piece of sand paper to grip on the sanded tube while you rotate it in the die, don't cut too fast and go both directions.

You do not need to use the liquid PTFE, but I had some and wanted to try it out.

The thread is more than enough to hold in the tube, but I wanted to see if the Liquid PTFE made it easier or harder to screw the tube into the PEEK.

It did add some lubrication, but does not seem to make much difference, most of the Liquid is pushed out, but I feel a little happier knowing it was added.

All done, let it dry and trim the tubes at a 45 degree angle to each other and to the walls of the heating block. 


Once cut you must drill/ream out the PTFE tubes back to be 2mm internal bore, it will have narrowed due to the threading and fitting into the PEEK block. Do this very carefully and make sure you remove all the PTFE swarf.

Then you can assemble it.

This time the stainless plate is clamping both the aluminium heat-sink and PEEK, this works exceptionally well, the Aluminium block stays warm and can be fan cooled if required, Stainless plate is cold.

New V2 hot-end and extruder assembly.

I know what you are thinking, how heavy is that lot! well it's not as bad as you would think, the extruders are lighter because of the Nema14 motors, but still, it's a significant weight - Note:- I'm running this on my BIG MendelMax using M8 smooth rods and it's still working fine.

I'm getting an X/Y travel speed of 170mm sec without any issues at all on my bigger MendelMax, I changed the acceleration down in firmware to 4000mm/sec, and set X/Y jerk to 15mm/sec.

I have tested print speeds of between 20mm/sec and 65mm/sec at the moment. It seems to be most happy at around 45mm/sec and a 1.4mm extruder retraction. (using 3 x 1.75mm Faberdashery filaments)

Electronics Expansion -

Ideally you would need to drive all three (or more) extruder's as separate outputs from the firmware and send sets of Gcode commands to enable and mix the colours for the object being printed. That’s all future stuff and quite doable as the next stage.

For now I'm using my standard RAMPS 1.3 setup with a modified Sanguinololu PCB to act as a stepper expansion board.

I decided as a first development to expand a single Extruder channel to drive 4 separate stepper drivers and control these stepper enable lines separately so filament blending could be tested with just standard Gcode and normal firmware. Slightly crude, but simple and very effective.

You still need to populate some components on the Sanguinololu, and fit Pololu or stepstick drivers. Components around the steppers are required, including the Microstepping select jumpers, these should be set to whatever your normal Extruder drive is set (mine are x8)

You do need to make sure each extruder drive has a separate enable line. Some are shared on the Sanguinolou, so cut the tracks to make them separate.

Also use a 4k7 resistor for the enable pull-ups as the main controller has a very weak pull-up (100k) already fitted on the enable line, if you don't fit smaller ones on this expansion board it won't switch the extruder's off quickly when you disable them.

You only need 5 wires coming from the original Extruder driver, Step, Dir, Enable, and +5V.

The Sanguinololu board must have +12V and GND connected to it's normal power rail input.

The 3 separate extruder drivers on the Sangunnolo are controlled by the single enable line from the main electronics. For ease of testing the enable line can therefore be switched on or off for each extruder. Spare output pins from the micro could easily be used for enabling each extruder, but for a proof of concept, switches allow the easiest way to experiment and evaluate performance and operation.

Make sure you set the trim-pot reference voltages to about 0.42v for NEMA14 motors if you are using Pololu setpper drivers. If you are using Stepsticks, that’s turned about 95% of the way around (about 1A)

All ready to go...

Setting the Firmware-

Now that 1, 2 or 3 extruder's will be driven together, you need to set the Esteps in firmware

For me the Esteps are 360 (x8 micro-stepping) for one extruder. So if you want to run a single extruder at a time and switch individually between them, 360 is the correct number.

More interestingly you can run any two extruder's at the same time, so the two filaments are combined / blended together, you still have three output combinations and it's actually possible to swap out the 3rd inactive filament for a different colour while it's printing using the other two, then switch to another set of two extruder's using the new colour/material. to do that the Esteps need to be half the value as you have two feeds both giving half the volume to the nozzl, 1/2 of the number above = 180

And finally running all three at a time can give some great looking prints and different colours from each direction of the object! to do that you will need 1/3 of the Estep value = 120

It's easier to see the different effects in the video than in the pictures below, but take a look at both.

Video Part 2 - Construction and Prints -

Prints!-

You must be wanting to see some prints by now :) - Here you go -

Single extruder feed at a time, Gold, Red and Blue filaments used.

Single extruder feed at a time, Gold, Red and Purple filaments used.

More pots, Single extruder at a time.

The one at the back also had a colour swap of gold for the last section.

Two extruder's at a time, Pink, Blue and Pearl White filaments used.

You can see in this picture, that Pearl White is the common colour and Pink and Blue are alternated as the print progresses. (Ignore the yellow outline, that was the purge from the previous used colours)

This picture shows the two extruder colours and how they work on a printed object, this is the same part just rotated 180 degrees, using Orange and Silver filaments.

Two extruder's at a time, Punk star Pink (Magenta), Electric Blue (Cyan) and Mellow Yellow filaments.

Here you can see Greens, Oranges, darker Blues and Reds in the mix. 

Lots of stretchlets.

Orange and Glitter*.

This shows the same pot, rotated 180 degrees - Using Cyan, Magenta and Yellow.

Lower part of the frog printed with Magenta and Yellow, top has Magenta one side and Cyan on the other.

Frog - Red, Purple and Gold filaments.

Left frog - Pink and Silver.

Orange, Purple and Glitter*.

A Universal Paste Extruder body printed with Red, Yellow and Pearl White filaments (2 extruders enabled at a time). - Same part, just rotated.

Mega Stretchlet bangle, Red, Yellow and Blue, all on together.

Close-up of Stretchlet - see the 'toothpaste' effect.

I used up all my 1.75mm filament stock in these test prints, but I'm keen to build up a whole printer in glorious multi-colour blended wonderfulness!

And Imagine using different materials, soft PLA for modifying hardness or doing a flexible 'hinge' in the part. or Glow in the Dark with a hint of whatever colour you like for daytime and night viewing.
Mixing both ABS and PLA together would be an interesting experiment, maybe one to try at some point.

For now more work getting the electronics a little more integrated and maybe a multi-fed Bowden version?

Other news - 

Raspberries - 
I had the pleasure of presenting at the recent Raspberry Jam in Bristol on Monday this week, (not 3D printing this time around, but maybe next time) it was a splendid evening and we had some really great presentations and demo's. Thanks to everyone that came for the Jam.

And a big thank you to Patrick for giving me one of his first really nice Laser-cut (home-made) Acrylic Raspberry Pi cases. If you don't want to print one, you can get hold of his very nice version here -

Clay -
I do get lots of messages about the Universal paste extruder, thanks and keep them coming, and I am planning on doing much more, both on clay and other materials and also the Peristaltic pump extruder as soon as I can.
I do have Metal clay to try and also Paperclay (very good for low warp - perfect 3D printing material)

Rostock - 
Finally I need to mention the Rostock delta printer. If you don't already know?, go and read all about it , Right now, it's totally Awesome with a really big A.

Thanks as always for reading, and please post a comment or contact me to say what you think.

Until next time.

Rich.