Filament Recycling Process
recycling
filament extruder
3d printer
Here is a description of my current process for recycling 3d printer filament waste back into (usable) 3d printer filament. I took these pictures over multiple runs, while still refining the process, so there will be a few different colours.
For an evaluation and a look at the results, see here.
Material #
First, there are some prerequisites for successfully recycling filament for 3d printers. The main measure of success I use is how well the recycled filament can print on a normal printer with normal settings (just maybe temperature adjusted by +- 15C), especially compared to commercially offered filament.
The input materials have to be the exact and only polymer that is also targeted as the output. So for PLA it needs to be 100% PLA, respectively 100% PETG for PETG (those are the two main materials I currently recycle). A purity of 100% is only theoretical, so I try to stick to >99.5% and sort the input materials I receive manually.
I check and maximally reduce the following contaminants:
- dust (using a high-pressure air blaster)
- grains/granules of any type: because it’s unknown material (high pressure air)
- filament spaghetti, because it attracts dust really well and only gives small amounts of material
- badly burnt failed prints, when a clogged nozzle started “caramelising” the material
- the other material when I can tell them reliably apart: PLA/PETG
- sometimes other polymers like ASA or ABS are also mixed in and need removal (acetone solvent or other methods to identify)
Once the pieces are clean, they need to be shredded into granules.
Preprocessing before shredding #
This mainly includes big pieces that do not fit in one go into the “shredder” (kitchen mixer) and therefore need some reducing in size.
For that, I put them in an oven and set a temperature of ~210C. The pieces then flatten down and can be cut down with steel-sheet scissors. These pancake shapes are well shreddable by the mixer, at a size of about ~30gr per piece.
I’ve recently also started using a jigsaw for cutting bulky pieces, but that is somewhat dangerous and produces non-uniform shapes that are not always reliably shredding in the mixer, because of their still jagged shape.
“Pancakes” before shredding into granules in the mixer. The curved shape is so that they move chaotically in the mixer chamber. Completely flat pancakes can often get stuck diagonally in the chamber, because of the air currents made by the mixer.
An example of some multicolour PLA, here preprocessed in ~2cm sized chunks for shredding in the mixer. These chunks are massive plastic and somewhat unsafe to shred, so I prefer the pancake method now.
Shredding #
The shredding process just takes the input materials and reduces them to a small and consistent powder. For the extruder to work properly, each granule has to be either <4mm in all dimensions, or <4mm in two dimensions and <6mm in ONE dimension. Think of a rice grain: that would be oversize in one dimension, so still allowed.
A view of the mixer’s chamber, before shredding a pancake. Any piece needs to reach all the way to the bottom of the chamber to be shredded, and not wedge itself diagonally somewhere higher above the blades.
The shredder is just a kitchen mixer, though rated for 2kW 24000rpm. In the interest of safety, I added a few reinforcements around the chamber, which was not necessarily designed for plastics.
Sometimes my reinforcements would crack. Then I just print new ones with even more perimeters for part strength.
Granules #
Here is an example of some granulated material. To achieve the size constraints for extrusion, I pass everything through a sieve, and shred the granules that are too big again.
Hermetic bags are used to keep the granules clean during storage, and also attempt to keep the humidity away. But because of the very high surface area of the granules, they still absorb humidity from the air relatively fast.
Almost any humidity present in the granules during extrusion will manifest as steam bubbles during the melting, and will leave a rough, sandy surface on the filament. With higher humidity, the expanding steam also changes the target diameter of the filament for the worse, sometimes making it unprintable.
Dry granules will yield a smooth and shiny surface on the filament, which also enables a consistent volumetric flow rate and makes the result precise in diameter. The first picture shows the transition from dry to smooth gray granules, a notable improvement.
Drying #
After storage, the granules are heated in the oven (at ~50C) immediately before extrusion. This dries them from any humidity they already absorbed, and to prevent them from re-absorbing any again, they are shuffled from the oven into the running extruder immediately. This minimises exposure to air, where the polymer can soak up humidity quite fast; especially when warm.
Extrusion #
Here is the exit of the extruder, with the injection screw and melt zone above, insulated in the cylinder shape. The granules are heated up and put under higher pressure. The molten plastic then exits at the bottom through the nozzle, which gives it the round shape of filament.
The white piece with text “SO-D” is part of the filament height sensor, which also has a little metal wire resing on the newly made filament. This information is used by the extruder to adjust pulling and spooling-up speeds for a consistent filament diameter.
I added some modifications to this system to help it with longer timescale disturbances, like air temperature rising or granule density changing over the whole batch. The mechanism I designed is printed in black and rides on the 3030 aluminium frame as a stabilising rail. It can move the extruder’s height sensor assembly up or down quite granularly, thanks to the M4 feedscrew. One turn of the M4 rod corresponds to a height change of just one thread pitch: 0.7mm. This mechanism has slightly improved the diameter consistency, reaching up to 1.75+-0.03mm in precision over an entire 1kg spool.
Extruded filament cooling path, with lots of fans to cool down and stretch it into proper dimensions
Extrusion path with yellow filament #
The hopper with dry granules
Winding up the new filament at the start of the spool
Results #
A detailed writeup is here.
This recycling project is supported by
