HomoFaciens refines a granules-fed FDM extruder, easing plastic waste recycling

Pellets are the promise, but repeatability is the fight

HomoFaciens’ version 7.1 granules-fed FDM extruder gets at the part of desktop recycling people actually care about: can you feed shredded plastic into a printer without turning the whole setup into a maintenance hobby? The appeal is obvious. If you can skip tightly controlled filament spools and run granules instead, you open the door to cheaper raw material, easier storage, and a real path for turning waste plastic into usable feedstock.

That is the real value of this project. It is not just a novelty extruder. It is a direct answer to the filament bottleneck, where cost, diameter consistency, and spool handling all add friction before a print even starts.

Why granules matter more than filament on paper

The earliest HomoFaciens granule extruder notes make the premise plain: waste plastics have to be shredded into granules, and granule size is a critical property of the feedstock. That sounds simple until you remember that pellet systems are sensitive to everything filament hides for you, from feed uniformity to air entrapment to how the melt behaves inside the screw and tube.

That is why this kind of machine attracts tinkerers who like the idea of recycling at the bench and production-minded builders who want more flexible material handling. Granules are less fussy to source than filament, and in the right setup they can support recycled plastics or specialty blends that would be awkward or expensive to wind onto spools.

The hard part was never the idea, it was the hardware

HomoFaciens has been iterating this concept for years, starting with experiments that included shredding PLA in a blender and trying PET from old plastic bottles. The big lesson from the older versions was not that pellet printing fails. It is that consistency is hard, and the failures tend to be mechanical rather than conceptual.

Version 7.0 already showed the core problem: V6.2 printed well, but not consistently. The main technical headache was that molten plastic could rise toward the cold end and solidify on the extruder wall. Add a not-very-powerful Prusa MK4 stepper motor into the mix, and the design has to stay within a narrow band where auger diameter, torque, and bore geometry all cooperate instead of fighting each other.

What version 7.1 changes

Version 7.1 is not a rewrite. It is a cleanup pass aimed at reducing the fiddly bits that make a homebuilt extruder harder to reproduce. The tube is based on a 40 mm long stainless-steel M6 threaded sleeve, and the previous brass insert has been removed. That matters because the V7.0 approach required more manual finishing with a drill bit and hand reamer, which is exactly the sort of extra work that turns a promising build into a one-off.

The new bore is drilled with a 6.8 mm bit and then opened to 7 mm with a hand reamer. The upper section is widened to 8 mm for 6 to 8 mm of depth, which HomoFaciens says can increase throughput by up to 20%. Three notches and a chamfer at the top further improve feed. Those are the kinds of small mechanical changes that make the difference between a clever prototype and something you can actually keep using.

What the test prints say

The V7.1 test work is where this stops feeling theoretical. HomoFaciens used a hollow cylinder print, 100 x 100 x 100 mm, with 0.7 mm wall thickness, 0.4 mm layer height, 0.7 mm extrusion width, and 30 mm/s print speed. That is not a decorative bench artifact. It is a straightforward stress test for whether the feed system can keep material moving evenly under real print conditions.

The result that matters most is also the simplest one: microscope inspection of the printed wall found no air bubbles. For a granule-fed extruder, that is a big deal. Trapped air and inconsistent melt are the sort of defects that quietly wreck wall quality long before a print visibly fails.

The project also says the extruder was tested on a Prusa Mk4 and kept up through bridging tests. That is another useful signal, because bridging pushes a system harder than a slow, neat perimeter does. If the feed can stay stable there, it starts looking less like a bench experiment and more like a machine worth refining.

How it compares with standard filament

Filament still wins on convenience, no question. You buy a spool, dry it if needed, load it, and print. The dimensional consistency is built in, and most desktop machines are tuned around that workflow. Granules-fed printing, by contrast, asks you to do the upstream work yourself: shred, size, dry, feed, and tune.

But filament also locks you into the economics and logistics of spools. That is where pellet systems get interesting. NASA’s Spinoff coverage of AI SpaceFactory’s Starforge printer says filament can cost up to $150 per pound, while pellets can come in at about a couple of dollars a pound. Pellets can also be strengthened with glass or carbon fibers, which makes them more attractive for specialized material recipes. NASA also notes that AI SpaceFactory won the 2019 3D-Printed Habitat Challenge, which gives this feedstock approach some serious pedigree beyond a garage workshop.

So the trade-off is clear:

• Filament is easier, more predictable, and much more mature on desktop printers.
• Granules can be cheaper and far more flexible, especially for recycled or custom materials.
• Pellet systems demand better mechanical tuning, more material prep, and a higher tolerance for tinkering.

Who should care about V7.1 right now

This is not ready for someone who wants to swap filament for pellets and never think about it again. It is for the builder who is already comfortable drilling, reaming, testing, and adjusting. It is also for anyone who wants to explore recycled-material workflows without waiting for a polished commercial platform to do all the hard parts.

Prusa’s MK4S product page says the MK4 family is used by hobbyists and big companies alike, which makes the MK4 a fitting test bed here. It sits right at the edge of hobby and production, exactly where a granule-fed feed system starts to make sense. If the goal is to move from one-off experiment toward repeatable small-scale production, this is the kind of platform that can expose what still needs fixing.

The honest answer is that pellet extrusion is no longer a pure science project, but it is still a tinkerer’s machine. Version 7.1 gets closer by stripping out friction, improving feed geometry, and showing clean test results. The promise is real now. The next hurdle is whether that promise survives long runs, varied recycled feedstock, and the kind of ordinary abuse that filament printers take for granted.