My machinery 2021: Direct Granules Extruder

Waste plastic turned into raw material for 3D printing.

Similar projects worth following
For this year's Hackaday Prize I dive deeper into 3D printing. The aim is to reuse waste prints with a direct granules extruder. I got this working nicely during the past weeks and I also discovered a way to use industrial pellets as raw material, replacing the need for filament (on plastic spools). Another benefit of using granules as raw material is the capability of printing a wider range of materials that are otherwise too soft or to brittle to be processed as filament.

The extruder is meant to be a drop-in replacement for existing desktop filament extruders with no additional electronics needed, thus it is working with a relatively weak stepper motor and uses one normal heating cartridge. No extruder you can buy anywhere, yet!

With my machines I aim to give students a better understanding of how technology works and of course I'd like to drive innovation as well. "Keep things simple" is the mantra that I follow. Simple to build, out of parts that are simple to source and bring it to live with simply some lines of software code.

I am fighting common thoughts that complex machines like 3D printers must be composed of complex parts manufactured with complex machinery. I am replacing expensive tool chains with my brain ;-) That often means breaking with conventional ideas.


- Build a reliably working extruder prototype (V3).

- Create build instructions (including video), even so it is a prototype made following my intuition.

- Explain in detail how the extruder works and so why normal wood screws are better than anything else: In contrast to conventional extruders, the forwarding happens in the cold zone only.

- More details about how to print with household sugar and so using materials with essentially different properties than common FDM materials.

Done, but not explained in detail, yet:

- What to consider when preparing raw material

- How to feed the extruder with conventional, industrial pellets

- Find ways to convert waste prints into new raw material for the extruder

To do:

- More print tests with a larger variety of raw materials

- Build version 4 with series production in mind instead of the handmade version 3

openSCAD and stl files of the Rocket I have printed as a test run of the extruder.

Zip Archive - 311.58 kB - 11/07/2021 at 07:27



Build instruction of Direct Granules Extruder V3. Keep in mind, that this handmade prototype was largely build using the rule of thumb.

Adobe Portable Document Format - 391.30 kB - 10/25/2021 at 08:40


3D files of the printed parts needed for Extruder V3 (STL and OpenSCAD)

Zip Archive - 3.95 MB - 10/25/2021 at 08:33


Python scripts that generate G-code files to minimize warping of 3D prints. The scripts generate G-Code for a gear rack and a box.. Also included are stl files of such a gear rack and box.

Zip Archive - 709.24 kB - 09/03/2021 at 10:43


The Python script (tested on Debian Linux) and Arduino sketch I have written for my robotic arm v1.0 You can control the robotic arm with your keyboard and store/replay movements.

Zip Archive - 5.97 kB - 08/19/2021 at 06:17


  • 1 × Aluminum bar 16x16, 300mm length
  • 1 × Aluminum plate 3x65x130mm
  • 1 × 3mm threaded rod 1m
  • 1 × Glass block 25x25x8mm
  • 1 × Teflon tube 8mm inner diameter

View all 12 components

  • My Prusa MK4 no longer needs filament to print

    Norbert Heinz06/07/2024 at 13:57 0 comments

    I am in the process of converting a Prusa MK4 to granule printing. Part 1 of the process shows how to build the core component, extruder V5.0 and first test prints:

  • A closer look at the auger screw(s) of my direct granules 3D printer

    Norbert Heinz05/10/2024 at 11:40 0 comments

    The simplest component of my granule extruder must meet certain criteria to get a working 3D printer:

    Next step: Converting a Prusa MK4 to direct granule printing.

  • Extruder V4.1: Rough build instructions and thermal optimization

    Norbert Heinz04/27/2024 at 12:54 0 comments

    I have been shrinking the tube of my extruder in order to reduce friction and so to get a better retract:

  • V4 of my granules Extruder has a full Metal Core

    Norbert Heinz04/06/2024 at 10:05 0 comments

    Version 4 is made of a brass / stainless steel tube and has of course a simple wood screw an auger:

    A 20 minutes test print of a tiny 27x25x12mm track link shows the capabilities of the new design:

  • Let's have a look at the energy needed to grind PLA pellets

    Norbert Heinz03/17/2024 at 13:53 0 comments

    To get a better idea how to improve my pellet grinder, I had a closer look at the energy needed to turn pellets into powder:

    With the simple experiments carried out in the video, you get a better idea of what material throughput a low power router turned into a plastic mill can deliver.

  • Get raw material for my direct granules extruder

    Norbert Heinz03/01/2024 at 11:41 9 comments

    My "old school" way of getting raw material was using a blender and a sieve to get grain sizes of 2mm and less:

    The resulting grains under the microscope:

    To get fresh material from industrial pellets, I have turned a router into a grinder:

    With that, the resulting grain size is less than 1mm:

    Future tests will show, if the smaller grains can be used to design a smaller extruder.

  • Cookies 2.0 with a paste extruder

    Norbert Heinz12/22/2021 at 16:05 0 comments

    While waiting for parts to arrive for my granules extruder upgrade, I have created a syringe based paste extruder:

  • 3D printing Sugar in more detail

    Norbert Heinz11/10/2021 at 08:02 0 comments

    My lazy Sunday afternoon project of 3D printing sugar is followed by a more in depth view on what can be done with that exotic material and what can't be done:

    Sugar isn't plastics

    Due to a different chemical composition, the physical properties of sugar are totally different from that of common plastics used in FDM printing. A glass temperature exists, but only academically speaking. The transition from solid to liquid appears almost instantly. The viscosity is clearly lower than that of common plastics and the surface tension is higher. With that, sucrose, as used in my experiments, drips out of the extruder without forming strings.

    It works in general

    The very first try was a copy of the pinion on the stepper motor of my extruder. You can recognize it as being a gear, even the 3mm teeth can be seen, also the layering. The top surface is very rough. That's due to the low viscosity and the surface tension of molten sugar. At low print speeds, there is always a drop of sugar sticking to the tip of the nozzle and droplets form on the top layer.

    Print it fast to get rid of droplets on the surface:

    You need to print sugar with a higher print speed (30mm/s) to give the molten sugar not enough time to form droplets while cooling down.

    Print it slow to avoid cracking

    even when printed, sugar is still a very brittle material. With high print speeds, tensions build up in the material which can lead to cracks in thin walls. With a low print speed of just 10mm/s I could print 1.8mm walls without any problems.

    The more you know about a material, the better the results

    Part cooling and heated build chamber will make things better

    The diameter of the cone at the base is 100mm., the angle is 45degrees. The closer the print came to the tip, the less time there was for the sugar to solidify. That caused the forming of droplets and finally to an aborted print. Part cooling for small structures could be an improvement. Also a heated build chamber can help to avoid tensions in the material. There is more experimentation to be carried out, but not yet...

  • Hackaday Prize finals video, discussing the "Rocket"

    Norbert Heinz11/07/2021 at 07:26 0 comments

    Hackaday Prize 2021 comes to an end today. My project was voted into the final round and as part of the competition, an up to 5 minutes video for the jury has to be recorded. Here is mine:

    For this video I have created a "Rocket" in openSCAD ((see files section):

    The printer settings were:

    Hotend temperature: 181°C

    Bed temperature: 65°C

    Layer height: 0.2mm

    Layer width: 0.5mm

    Infill: 0%

    Rocket dimensions: 100x100x150mm

    Lets have a closer look at the result, my extruder delivered:

    First thing you will notice is the tip of the rocket:

    This is definitely where a part cooling fan would be helpful. Even though there is one implemented in my Extruder design, I did not screw it in place. The reason is, that is blocks the view on the nozzle and this is why I rarely used it during my test runs. These tests were always about the extruding capabilities, rather than about the overall print quality of the whole printer. So forgive me not to have added the fan before starting the print...

    While talking about the extruding capabilities, a smooth surface is a result of constant extrusion and that is what a good extruder MUST deliver.

    The detailed look at the surface demonstrates, that the surface is a it should be. What seems to be a crack in the surface at a first look is in fact nothing else than a single grain of non red color that entered the extruder (I didn't clean the blender nor the hopper that carefully).

    Stringing is not addressed, yet:

    In the video about the construction details (see previous blog entry), I have demonstrated how it  works. The issue is that the screw squeezes out an extra portion of plastics when going intro reverse. That must be addressed in an additional parameter in the slicing software, but doesn't exist, yet. More on that in a coming video/blog entry.

    Ghosting can be seen on the right fin of the rocket. This is due to the weak printer mechanics, not due to inconsistent extrusion.

    The first layer looks good, which is also an indication of constant extrusion:

    All in all, the quality, this handmade prototype delivers, is better than anything I have seen from a direct pellet/granules extruder. It only can get better with a better printer mechanics, a part cooling fan and a more precise machined extruder following this design (that is under construction).

    More on the sugar print will follow, soon:

  • RepRap Ltd starts reproducing my extruder design

    Norbert Heinz11/02/2021 at 15:43 0 comments

    I got in contact with Adrian Bowyer and now he tries to make a "copy" of my design, publishing the process on twitter via the RepRap account. As experienced 3D expert, the copy is of course no 1:1 rebuild of my drawings. It is also ruled by the materials he has in stock (e.g. the Teflon tube has an inner diameter of 7mm, not 8mm as I recommend).

    RepRap via Twitter:

    Can't wait to see it working!

View all 22 project logs

  • 1
    Why there is no step by step build instruction for Direct Extruder V3, but of course an instruction exists

    Direct Granules Extruder V3 is a prototype you can build from cheap, commonly available materials with low tech tools. It is not for beginners in 3D printing, but advanced tinkerers should be able to make a copy. Since not being intended for beginners, I don't provide a step by step build instruction. Instead you get drawings and pictures in the "files" section that should enable you to create your own extruder.

    Another issue is, that this design is based on other physical principles than existing extruders. It is essential to get a deeper understanding of those principles that I have described in the logs of this projects that also include videos.

    Leave a comment if you still have questions about the build process of the direct extruder.

View all instructions

Enjoy this project?



benjjyman wrote 11/07/2021 at 11:51 point

1 can we vote for you?

2 one problem with design is ptfe tube for use with recycled ABS, or Nylon or anything above 260 degrees celcius , other than that this brilliant

3 could you please explain warping python script ?

  Are you sure? yes | no

Norbert Heinz wrote 11/07/2021 at 14:43 point

(1) You can like my project (on top of this page), but votes are up to the Hackaday Price jury.
(2) Yes, in a future design, the PTFE tube will no longer reach into the hot zone.
(3) What explanation is needed? It was just a proof of concept, not yet meant to be an implementation in a slicing software.

  Are you sure? yes | no

heinz wrote 10/23/2021 at 04:05 point

Nice, I'm a fan of your design style. 

How is the screw located in the whole assembly? Does the screw end where the teflon tube ends and the 1.5mm rod continues to go into the heater block?

  Are you sure? yes | no

Norbert Heinz wrote 10/23/2021 at 11:17 point

Thanks! Right now I have added a log entry with a video, that will hopefully answer your question:

  Are you sure? yes | no

heinz wrote 10/23/2021 at 14:07 point

Ah okay, screw ends on the cold side or a little bit beneath. Didn't realize the 1.5mm is rod is for sealing the nozzle on retraction, thought it had some stirring function.

  Are you sure? yes | no

nodemcu12ecanada wrote 08/23/2021 at 14:24 point

Very inexpensive 6 degrees of freedom robot arms are readily available.

You have to get a power supply and do the controls yourself. I used a $4 NodeMCU.

  Are you sure? yes | no

Norbert Heinz wrote 08/23/2021 at 15:03 point

I never said, my robot arm was the first and only one. ;-)

Countless arms using RC hobby servos exist, I also have built one before:
These are all interesting devices to learn basics about robotics. Add some new design details to the army of DIY robots is why I created mine from wood.

  Are you sure? yes | no

Jason wrote 08/18/2021 at 21:35 point

Super cool!  I'd love to see some details and files that I could use to put together kits for local schools!

  Are you sure? yes | no

Norbert Heinz wrote 08/19/2021 at 06:24 point

Thanks! Sorry, I had forgotten to upload the file. The package is now available on Hackaday.
I have also added a link to my pages in the project log

  Are you sure? yes | no

duda.sertorio wrote 08/18/2021 at 17:09 point


  Are you sure? yes | no

Norbert Heinz wrote 08/18/2021 at 18:45 point


  Are you sure? yes | no

Similar Projects

Does this project spark your interest?

Become a member to follow this project and never miss any updates