Skip to main content

Open materials, recycling and distributed manufacturing

📍 Where we are: Part V · Beyond the Desktop — Chapter 13 of 16. The last chapter covered open file formats and open data. Now we look at the material that flows into the printer, and whether a community can make it from waste.

The simple version

Think of a home kitchen that bakes its own bread from flour it mills itself, then composts the scraps back into the garden. Distributed recycling is the same idea for plastic: shred an old bucket, melt it into filament, print something useful, and when that breaks, shred it again. The open-source twist is that the "mill" (the extruder) is itself a published design anyone can build. The catch is the same as in the kitchen: homemade flour is cheap, but the loaf is only as consistent as your mill.

What this chapter covers​

We follow plastic through a small, open recycling loop: the filament extruder known as the recyclebot, the Precious Plastic community and its machines, and the research field called distributed recycling via additive manufacturing (DRAM). We then look at open material data, at economics, and at the quality limits that decide whether recycled feedstock is a hobby, a research tool or a real supply chain. Everything named here has a public source; where I could not verify a claim, I say so.

The recyclebot: an open filament extruder​

Most desktop printers (see anatomy of an open printer) eat filament, a thin plastic line usually 1.75 mm wide. A filament extruder melts plastic and pushes it through a die, then cools and winds it. The open-source version from Joshua Pearce's group at Michigan Tech is called the recyclebot, and the group's name for its first version is the RecycleBot. A later design, the RepRapable Recyclebot, is published in HardwareX with build files; it can print many of its own parts on a RepRap, much like the original RepRap idea from Chapter 1. The paper reports a material cost under US$700 and about a day to build, a throughput of about 0.4 kg per hour at about 0.24 kWh per kg, and filament diameter within ±4.6% [1]. Treat these as the authors' reported figures for that design, not guarantees for your build.

An early peer-reviewed demonstration (2013) used high-density polyethylene (HDPE, the plastic of milk jugs) on the 3 mm filament standard then common. The filament averaged 2.805 mm, and 87% of samples fell between 2.540 and 3.081 mm; the authors call this less consistent than commercial filament and attribute it to a variable extrusion rate and manual drawing of the filament [2]. That spread is the heart of the matter: a printer assumes one nominal diameter, and a spool that wanders from it causes under- and over-extrusion.

You can check that yourself. The snippet below uses made-up diameter readings (illustrative, not measured) and counts how many fall inside a tolerance. I ran it with Python 3 and the standard library only.

import statistics
readings = [1.74, 1.76, 1.75, 1.81, 1.73, 1.69, 1.77, 1.75, 1.78, 1.72] # mm, illustrative
nominal, tol = 1.75, 0.05
ok = [r for r in readings if abs(r - nominal) <= tol]
print(f"in tolerance: {len(ok)}/{len(readings)} = {len(ok)/len(readings):.0%}")
print(f"mean {statistics.mean(readings):.3f} mm, stdev {statistics.stdev(readings):.3f} mm")

It prints in tolerance: 8/10 = 80% and a standard deviation of 0.033 mm. A mean of exactly 1.75 can hide a poor spool, so look at the spread, not just the average. Closed-loop diameter sensing on the extruder is the usual remedy, and it is an active open-hardware topic.

DRAM: distributed recycling via additive manufacturing​

Distributed recycling means doing the work near where the waste appears (a school, a village, a makerspace) instead of hauling it to a central plant. A systematic review of 92 papers from 2009 to 2019 split the DRAM chain into six stages: recovery, preparation, compounding, feedstock production, printing and quality assessment [3]. That split is useful because most failures happen at the boundaries between stages (for example, dirty input producing weak filament).

StageWhat happensTypical open tool
RecoveryCollect and sort by polymer typeCommunity collection, labels
PreparationWash, dry, shredOpen shredder designs
CompoundingMix with additives or other polymersRecyclebot or similar extruder
FeedstockMake filament or pelletsRecyclebot, pelletizer
PrintingFused filament or granular printingRepRap-class printer
QualityTest diameter, strength, contaminationCalipers, tensile tests

Research from the same group argues that coupling recycling and printing at home scale can lower the embodied energy and cost of making simple products, compared with industrial manufacturing and shipping. In a case study of three ABS products made from computer waste, the authors report that the coupled route roughly halves embodied energy [4]. A life-cycle analysis of HDPE filament likewise found distributed recycling used less embodied energy than the best-case centralized route [5]. Pearce also reviewed the broader economics of free and open scientific hardware and reports an average saving of 87% over proprietary equivalents [6]. Note that these savings are often calculated against commercial lab instruments, and they count the builder's labour at little or no cost.

Loop diagram of distributed recycling: waste plastic is sorted, shredded, extruded by a recyclebot into filament or pellets, printed, used and returned, with quality checks at the extruder and printer stages. The DRAM loop: waste is recovered, prepared, extruded, printed and used, then returned; diameter and contamination checks guard the two stages where quality is lost. Original diagram by the authors, created with AI assistance.

Precious Plastic: the community route​

Precious Plastic was started by Dave Hakkens in 2013 and is now run by a non-profit foundation in the Netherlands; its machine designs, blueprints and tutorials are free, published under Creative Commons Attribution-ShareAlike 4.0 [7]. Its core machines are a shredder, an extruder, an injection moulding machine and a compression oven, with a sheet press added later. It differs from the recyclebot in spirit. Recyclebot research aims at feedstock for 3D printers. Precious Plastic mostly makes finished objects (tiles, sheets, moulded products) by moulding and pressing, and it comes with a community platform with a map, library and marketplace for machines and parts [7]. Many chapters of this book are about printers; this is the sibling that skips the printer entirely.

Open material data and honest limits​

An open material database would let you look up a polymer's tested properties and licence. I did not find a single widely used, openly licensed database for recycled or hobby-grade AM materials, so I do not name one. What does exist in the open are per-study datasets in papers and repositories, and the NIST benchmark data in Chapter 15, which concerns industrial metals and one polymer resin rather than recycled filament.

The limits are real:

  • Consistency. Diameter and moisture vary from batch to batch; see the snippet above. In one test, printing with PLA that had been recycled once lost 10.9% of its tensile strength, showed more scatter, and clogged the nozzle, while stiffness was statistically unchanged [8]. Reviews of FDM-printed polymers likewise find that mechanical properties depend on the material and the printing parameters, so virgin-versus-recycled results from one setup do not transfer automatically [9].
  • Contamination and sorting. Mixed or dirty plastic makes weak parts. Sorting is manual and error-prone.
  • Maturity. The 2020 systematic review found few works on the recovery and preparation stages [3]. A 2020 logistics model states that the environmental and economic advantages of distributed recycling "have yet to be demonstrated" and that its operational complexity limits application [10]. A 2023 review (recent, secondary evidence) lays out the remaining challenges and opportunities for coupling distributed recycling with printing [11].
  • Certification. Parts from recycled feedstock are rarely qualified for safety-critical use.

Honest verdict: strong as education, local repair and research; unproven as a replacement for industrial recycling. Evidence note: the recyclebot design paper, the economics review and the logistics model are the only sources I found for what they report, so I cite them despite modest citation counts.

Why it matters​

A printer is only as good as what you feed it. Open extruders and open recycling designs let a school or a community in a place with weak supply chains make its own feedstock, and they make the economics of local making inspectable. They also teach a useful habit: measure your input.

In the real world​

Pearce's group publishes build files and papers openly, which is why a lab in another country can copy a recyclebot. Precious Plastic workspaces appear on its public map. Most users are labs, educators and small workshops, not factories.

Key terms​

  • Recyclebot — an open-source extruder that turns waste plastic into 3D-printing filament.
  • Filament extruder — a machine that melts plastic and forms it into a constant-diameter line.
  • DRAM — distributed recycling via additive manufacturing: local recycling coupled to local printing.
  • Distributed recycling — recycling done close to where the waste is produced.
  • Precious Plastic — an open-source community and set of machines for small-scale plastic recycling.
  • HDPE — high-density polyethylene, a common recyclable plastic, used in the first recyclebot study.
  • Feedstock — the raw material fed to a printer, such as filament or pellets.
  • Circular economy — a model in which products are reused or recycled instead of discarded.

Where this leads​

Plastic is the easy case. Next we look at the hard ones: printing metal, and printing very large things, in Open-source metal and large-format printing.