3D printing stopped being a prototyping exercise. Parts go into products now, which means the materials have to survive real applications, and engineering polymers are expensive.
This change is what makes in-house filament production worth it. And that's also why we launched the Filament Maker X10: a compact, industrial-grade extrusion system that runs powders, regrinds and pellets into finished spools at up to 1.5 kg/h (ref. PLA), unattended, on a workbench.
Filament Maker X10 - Become your own filament supplier.
It holds ±30 µm across the entire spool, verified by a 3-axis optical sensor with 10 µm resolution taking more than 25 measurements per second.
Virgin PLA costs around €10/kg. A lab scrapping a couple hundred kilos a year recovers a few thousand euros — against a system costing many times that. Recycling PLA was always an ethical argument, not an economic one.
PA12 powder from MJF and SLS printing costs €50–100/kg, and 50–60% of the powder in a build goes unused. A busy facility scraps well over a thousand kilos a year. Same machine, same process, completely different math.
FMX10-produced rPA12 + 10% rCF filament.
FMX10 parts made out of FMX10-produced rPA12 + 10% rCF filament.
In-house production is always the more sustainable option. It becomes the more economical one once the material is valuable enough — and that's now true for a widening set of polymers.
Enter your material, your annual volume and your waste stream, and the calculator returns cost per kilogram, payback period and CO₂ avoided for your setup.
An industrial extrusion line produces properly, but asks for capital, floor space, three-phase power, chilled-water plumbing, long changeovers, and an operator who understands extrusion watching the run. Reasonable if filament making is your business. Not reasonable if you just want your own supply.
Compact filament makers, including our own Filament Maker TWO, answer a different question well: can this polymer become printable filament at all? Both machines stay in the line-up, and where bench space is the binding constraint, Filament Maker TWO is still the machine we recommend.
What didn't exist was anything in between. The Filament Maker X10 is built for that gap: production output and unattended running, single-phase power, no plumbing, no installation project.
FMX10's Cooling Path.
Extruded material has to cool below its glass transition temperature before it can be spooled at printable quality. On a compact machine that distance is short (17 cm on the Filament Maker TWO) and it's the cooling, not the melting, that caps output. Push more material through a short cooling distance and the filament arrives at the winder still soft, where it deforms under spooling tension. Diameter drifts, the spool goes out of tolerance, and the run is wasted.
On the FMX10, filament travels 2.80 meters from nozzle to spool: 1 m along a silicone conveyor belt, then 1.8 m back to the spool inside the machine body. Folding the path back is what keeps a production-length Cooling Path inside a 1785 mm footprint.
Gradual cooling means filament is placed on the belt rather than stretched through the air. That's what unlocks up to 1.5 kg/h (roughly five times the Filament Maker TWO) without giving up dimensional consistency. Four independently controlled heating zones run to 450 °C at ±0.1 °C, which opens up high-temperature polymers, low-viscosity polymers, and filled and composite materials in the same machine.
The Cooling Path is air-cooled, which matters beyond the plumbing it saves. Water-bath cooling drives moisture into hygroscopic polymers — nylons especially — and moisture in the filament shows up later as bubbling, poor layer bonding and weak parts. Air cooling avoids introducing it in the first place.
At full output one FMX10 covers roughly the filament consumption of 30 printers, so a larger farm simply runs more than one.
3D Print farm running FMX10 filament.
Supplier-grade diameter, minimal operating expense
Output alone doesn't make filament usable. The FMX10 holds ±30 µm across the entire spool. The reference value is recycled PA12 with 10% recycled Carbon Fiber: a filled, recycled feedstock rather than an easy virgin one.
It comes from where the measurement happens. A 3-axis optical sensor sits at the end of the Cooling Path, after thermal shrinkage has settled. It resolves to 10 µm and takes more than 25 measurements per second, feeding a closed-loop control system that corrects continuously without anyone tuning the process mid-run.
The 3-axis filament sensor in action.
Filament data lands on the dashboard, and you can monitor and operate the machine remotely. The only hands-on steps are starting it, changing spools, and shutting it down.
Commodity polymers — PET, PLA, PP, HDPE, PETG, ABS, PS, LDPE.
Engineering polymers — PA6, PA11, PA12, PA66, PC, POM, PMMA, PPS, TPU, TPE, TPV, PEBA, PVA, PHA, PHB.
High-performance polymers — PEEK, PEI, PEKK, PPSU.
Plus composite and additive inputs: carbon fiber, glass fiber, ceramic powder, metal powder, wood, nanoparticles, chemical additives, polymer blends and biodegradable polymers.
Virgin or recycled, powder, regrind or pellets: all feed the same machine, which is what makes MJF and SLS waste powder a viable input rather than a special case.
We tested filament produced on the FMX10 from our own recycled MJF PA12 powder to ISO 527, against the original printed MJF part and against commercially available FDM filaments.
Peak tensile strength along the print plane. Recycled PA12 produced on the Filament Maker X10, against a printed MJF part and commercial FDM filaments.
3devo values: recycled MJF PA12 powder, extruded on the Filament Maker X10, ISO 527. rCF = recycled carbon fiber. Reference values: HP 3D High Reusability PA 12 datasheet, printed MJF part, ASTM D638 · Stratasys FDM Nylon 12 datasheet · Polymaker Fiberon PA12-CF10 datasheet, dry, ISO 527. Commercial PA12-CF grades vary from roughly 52 to 77 MPa in-plane depending on fiber loading.
Recycled PA12 matches the original MJF part in-plane (47 MPa against 48 MPa) and matches commercial virgin PA12 FDM filament at the same 48 MPa. Blended with 10% recycled carbon fiber it reaches 72 MPa, exceeding the original part by half and sitting inside the range of commercial PA12-CF filament, which runs from roughly 52 to 77 MPa in-plane depending on fiber loading.
Peak tensile strength across the layer boundary — the weaker axis for any FDM part, and where the carbon-fiber blend makes the difference.
3devo values: recycled MJF PA12 powder, extruded on the Filament Maker X10, ISO 527. rCF = recycled carbon fiber. Reference values: HP 3D High Reusability PA 12 datasheet, printed MJF part, ASTM D638 · Stratasys FDM Nylon 12 datasheet · Polymaker Fiberon PA12-CF10 datasheet, dry, ISO 527.
Unfilled recycled PA12 reaches 23 MPa across layers. That gap is ordinary FDM anisotropy measured against an isotropic powder-bed part — a consequence of how FDM builds, not of recycling. Adding 10% recycled carbon fiber closes it: 51 MPa, past the 48 MPa MJF reference and level with commercial PA12-CF10 at 52 MPa.
Over the past year we've processed more than 2,000 kg of MJF PA12 waste powder in-house, extruded it on FMX10 machines and printed it into brackets, covers and casings for our own production lines — parts that used to be CNC-milled.
One of them is a filament stabilizer bracket fitted to the FMX10 itself. The powder becomes filament on an FMX10, and the printed bracket goes into the next machine off the line.
FMX10 filament stabilizer bracket part, made with rPA12 + 10% rCF filament.
The model works commercially beyond our own floor. In the UK, filament supplier Filamentive began selling rPA12 earlier this year, a recycled nylon filament made entirely from MJF powder waste, produced on the Filament Maker X10. Filamentive describes it as the first commercially available filament of its kind. Read their announcement.
As an illustration rather than a promise: one setup recycling around 2,000 kg a year, running on office hours, can avoid roughly €80,000 of material purchasing and about 9.6 tons of CO₂ compared with buying virgin material.
Actual figures move a long way with the material, the volumes and the conditions on each site. That's exactly why we built a calculator instead of a claim — the same inputs that change the answer are the ones only you have.
Run your own cost and CO₂ numbers: enter your material, your annual demand and your waste stream, and see the payback and carbon avoided for your setup.
Before launch, fifteen machines ran for more than six months across six pilot customers and our own production line: automotive, medical, retail lighting manufacturing, print farming and commercial filament supply. Not demonstrations. Real production work, on their own materials, in their own conditions.
Throughout that phase we collected their comments, their process data and their printed results, and fed all of it back into the machine. Different sectors, different feedstocks, different ambient conditions, and what came back shaped what the FMX10 had to be: throughput high enough to supply a production floor, tolerance tight enough that the filament behaves like supplier-grade material once it reaches a printer, and neither of those depending on someone standing next to the machine.
One pilot, a retail lighting manufacturer, now runs an FMX10 in the same room as its printing farm, worked by the same operator. Filament production is happening alongside the printing rather than as a separate operation with its own specialist. Across all five sectors, the machine is in real applications.
Individual case studies will follow over the coming months.
The Filament Maker X10 is available now. Join our live demo to watch it run, or calculate how much cost and CO₂ you can avoid by implementing FMX10 in your workflow.