3D Printer (Desktop FDM)

identity

• name: 3D printer — this entry covers the desktop FDM (fused deposition modeling) type specifically, the kind found in and around woodworking shops. FDM is one of several 3D-printing processes; SLA (stereolithography), which cures liquid photopolymer resin with a UV laser, is a mechanically unrelated sibling process and is treated here as a contemporary, not a variant (see lineage and lineage-misuse) [3P1][3P2].
• family / category: Digital fabrication → the additive branch — a computer-controlled machine that builds a physical object by depositing material layer by layer from a moving nozzle, rather than removing material (a router or CNC mill) or ablating it with a non-contact energy beam (a laser cutter). Sibling to CNC Router and Laser Cutter (Laser Engraver); all three are "computer-controlled and programmed from CAD/CAM" but differ in fundamental mechanism [CNC-cross][LC-cross].
• one-line: A computer-controlled machine that melts thermoplastic filament and deposits it layer by layer onto a build plate to construct a solid object from a digital model — the only member of the digital-fabrication sibling group that adds material rather than removing or ablating it.

✦ how NOT to use it

• not-for:
  • Producing genuine structural, load-bearing wood parts as a substitute for actual joinery or milled stock — see matching → cannot replace.
  • Cutting, ablating, or removing existing material — an FDM printer only adds material; it cannot perform a router's or laser's job in reverse.
  • Printing outdoors-rated or heat-exposed parts in plain PLA — PLA's low heat resistance (glass transition ≈60°C) makes it a poor choice near a spindle, in a hot garage, or in direct sun [3P19].
• dangerous:
  • A named, fire-department-confirmed fatality anchors this class. Calvin Yu died in San Francisco after a Tronxy X5SA printer (purchased via AliExpress/Alibaba) overheated and ignited a couch on June 11, 2020; he died the next day. The San Francisco Fire Department's investigative team concluded the fire was caused by the printer; his parents subsequently sued Alibaba and Tronxy 3D for product liability, reported across multiple outlets in June 2022 [3P23].
  • RepRap's own wiki documents the specific technical failure mechanism, from the community that builds these machines, not a third party. A dislodged heating cartridge falling onto a printed part started a fire in one documented case because the firmware's thermal safety check only monitored the setpoint, not the rate of temperature change — a real, named mechanical/firmware failure mode, not a generic "electronics can fail" caution. Fan failures causing hot-end overheating and molten-filament drips are separately documented on the same page as a related ignition risk [3P24].
  • IF a print will run unattended for an extended period (overnight, or while the shop is empty), THEN this is a real, community-documented risk category — but this entry holds the "unattended overnight print causes a house fire" claim at a lower confidence tier than the Yu case above. A widely discussed 2024 house fire in St. Johann, Austria, was linked by the printer's own owner (Reddit user "Stablebullet") to his Creality Ender-6, which had reportedly been running for hours — but Austrian police attributed the fire to "a technical defect," and the owner's own follow-up comment expressed uncertainty over whether the printer or a charging laptop was the actual cause. This entry treats that case as disputed, not confirmed, and separates it clearly from the Yu case above [3P25].
  • A recurring hardware-defect pattern, not a one-off: Bambu Lab's A1 was recalled by the US Consumer Product Safety Commission in 2024 for electric shock and fire hazards; a second, distinct fire-hazard issue surfaced in 2025 tied to an overheating thermistor on the AC power board (reaching roughly 140–160°C and melting nearby plastic), which Bambu Lab acknowledged and addressed with a redesigned board and unit replacement/repair program in Q3 2025 [3P26].
  • Burns, moving parts and shock are the routine physical hazards alongside fire. NIOSH reports extrusion nozzles typically running at 190°C to 260°C and heated beds at about 55°C to 120°C, both hot enough to burn skin, with burns common when users pull melted plastic off a still-hot nozzle [3P29, section 3.7]; stepper motors, pulleys, carriages and fans can catch fingers, long hair, loose clothing or head coverings [3P29, section 3.8]; and shock or injury is possible during maintenance or a malfunction if the unit is not de-energized [3P29, section 3.10].
  • Setup and stop controls for this class: (1) fit and keep the guards, covers and enclosure interlocks the maker supplies (NIOSH cites nozzle socks or covers and interlocks that pause printing when an enclosure opens) [3P29, section 4.4]; (2) before clearing a jam, removing a part or servicing the machine, let the hot end and bed cool, then switch off and unplug, unless the maker's manual expressly prescribes that step with power on, in which case follow that procedure exactly [3P29, sections 3.7, 3.8, 3.10]; (3) keep the printer clear of combustibles and of fire-suppression nozzles, in a space with fire detection and suppression, and not on or beside soft furnishings (the Yu case above) [3P29, section 4.4][3P23]; (4) inspect wiring, connectors, heater cartridge and thermistor seating and fans, and stop and unplug at any damage, a failed fan, smoke, melting or a burning smell (the failure modes on RepRap's safety page) [3P24]; (5) check the specific make and model against the CPSC recall list and the maker's own notices before use and again from time to time (the A1 case above) [3P26]. Firmware thermal protection and an enclosure reduce some of these risks; neither one guarantees the printer is safe to leave running.
  • Emissions are a dangerous-class hazard too, not only fire. Ultrafine particles and VOCs from heated filament, with inhalation the main route, are documented by NIOSH, which also reports respiratory symptoms in 27 of 46 surveyed workers in one study [3P29, section 3.5]. Running a printer indoors without the emission gate in correct use (ventilation or a filtered or exhausted enclosure, chosen for the printer and filament, per the SDS and maker's instructions) belongs in this class, whatever the filament.
• damaging:
  • Printing structural materials (ABS, ASA, nylon) in an open-frame printer without accounting for ambient temperature control invites warping and delamination — the exact failure mode Stratasys's own 2000/2004 heated-build-chamber patent (expired 2021) was built to solve, and still the dominant print-failure mode on unenclosed machines with these materials [3P12][3P18].
  • Printing nylon without pre-drying it degrades both the print and the part's strength — nylon absorbs ambient moisture readily and should be printed from a sealed or desiccant-controlled filament box [3P19].
• beginner-trap:
  • First-layer adhesion failure is a documented beginner problem with its own manufacturer troubleshooting article. Prusa's Knowledge Base article for its FDM printers, "First layer issues," warns against compensating for poor adhesion by setting the nozzle lower (Live Adjust Z) and says lowering it too far can damage the print sheet, damage it says the warranty does not cover; it suggests firmware bed level correction for a persistently uneven surface, and gives sheet-specific cleaning (acetone on the smooth PEI sheet only, about once a month; none on textured or satin sheets) [3P27]. Its manual First Layer Calibration and Live Adjust Z procedure is written for the MK2/S, MK2.5/S, MK3/S/+ and MK3.5/S; other models, Prusa's own newer ones and other makers', may calibrate automatically or differently, so follow the procedure for the model in hand. (Two details this line carried before the 2026-10-17 fix pass, bare-finger PEI contamination and a smooth/textured re-calibration trap, were sourced to a resin-printer article and were not confirmed on the FDM page above, so they are removed.)
  • Warping on long, flat prints (a straightedge, a long fence, a router-table insert) is a real, cross-corroborated risk, worse with ABS/ASA/nylon than PLA/PETG on a wide footprint — a clean plate, correct Z-offset, stable bed temperature, and a slow, wide first layer all reduce it [3P28].
  • "Just increase the infill" is a common but flawed fix for a weak load-bearing jig. A peer-reviewed study found that a part with horizontal flat surfaces other than its top and bottom develops an "interrupted shell," which can cause a drastic drop in load-bearing capacity — walls, ribs, and gussets (not dense infill) are what actually carry bending and tension loads on a printed part [3P21]. A beginner who cranks infill percentage on a weak jig design without addressing wall count or orientation is fixing the wrong variable. Better walls and orientation improve a printed jig; they still do not by themselves rate it for powered-machine workholding (see correct use).
• lineage-misuse:
  • Confusing FDM with SLA/resin printing — both are called "3D printers," but they are chemically and mechanically unrelated processes (thermoplastic extrusion vs. photopolymer curing) with entirely different materials, failure modes, and safety profiles (resin printing's principal hazards — uncured-resin skin contact, VOC/odor from the resin itself — are not the same set of hazards documented here for FDM; this does not mean FDM is free of vapor hazards, since heated FDM filament releases its own ultrafine particles and VOCs, covered under correct use and dangerous [3P29]). Treating advice for one as applicable to the other is a category error [3P1][3P2].
  • Confusing a 3D printer with CNC Router or Laser Cutter (Laser Engraver) — the three share a CAD/CAM-driven, computer-controlled workflow, which invites treating them as interchangeable "the shop's digital tool," but the underlying process is opposite in each case: additive (printer), subtractive-contact (router), non-contact ablative (laser). This is the exact gap CNC Router's own entry flagged as an unsourced inference — now sourced via the anatomy account above [CNC-cross][3P17].
  • Expecting a printed part to substitute for solid wood's structural properties — held as Haft's own reasoned inference, not a found misuse account (honest gap). Despite direct, repeated searching, no single named forum or practitioner account of someone actually making this specific reasoning mistake was located. The claim is instead grounded in a well-known maker's own tested strength/stiffness comparison (see matching → cannot replace) [3P22] — a sourced mechanical basis for a plausible mistake, presented honestly as inference rather than as a documented case.

correct use

• what it's genuinely good at: producing custom, low-volume, geometrically complex parts directly from a digital model without tooling, molds, or machining setup — jigs, fixtures, brackets, organizers, prototypes, and small hardware that would be slow or impractical to make by hand or on a subtractive machine [3P15][3P19].
• IF the part is a low-load, indoor, room-temperature fixture (a drill guide, an alignment jig, an organizer), THEN PLA on an unheated or lightly-heated bed is normally sufficient and is the easiest material to print reliably [3P19][3P20].
• IF the part needs to survive shop heat, mechanical stress, or repeated flexing (a hold-down clamp, a dust-shoe adapter, a spool holder), THEN PETG is the more common shop default — tougher than PLA, more heat resistant, and it flexes before it snaps rather than shattering [3P19]. Prusa's own PETG page lists "mechanical parts, holders, clamps" among PETG's common uses [3P30]; a material name is still not a load rating (see the load-bearing line below).
• IF the part will sit near a spindle, motor, or in a hot/enclosed space, or needs UV stability outdoors, THEN ABS or ASA is the better choice — but an enclosed printer is normally needed to control warping, and ASA specifically for UV exposure [3P19]. That enclosure is for warping; it does not settle emissions, so ABS and ASA go through the emission gate below before they are printed indoors.
• IF the jig or fixture is genuinely load-bearing, THEN print orientation and wall count matter more than infill percentage — more perimeter walls generally carry bending/tension loads better than higher infill, and parts should be oriented so the expected force runs along the layer lines rather than across the layer boundaries, a load-bearing weak point (see beginner-trap) [3P19][3P21]. Material, orientation and wall count alone do not make a printed part a rated fixture. For workholding on a powered machine (a router-table hold-down, a clamp or featherboard holding stock against a cutter), treat a printed design as illustrative unless that specific design has its own test or load evidence, and keep the machine's own guards and rated workholding in use. Printed clamps are not ruled out; they are unrated until shown otherwise. (This is an engineering inference from the reviewer's CX-418 review, not a documented failure of a named printed part.)
• IF the printer will run indoors (any filament, PLA included), THEN set up emission control before the first print. NIOSH reports that heated filament can release ultrafine particles and volatile organic compounds, that emissions vary with filament type, color, brand and extruder temperature (higher temperatures gave larger emissions), and that inhalation appears to be the main exposure route [3P29, section 3.5]. Its engineering-controls section cites one study recommending a minimum of six air changes per hour or a portable HEPA filter, another finding three air changes per hour insufficient for regular use, and filtered or exhausted enclosures that cut particle emissions by 97% to 99% in testing [3P29, section 4.4]. Choose the control for the actual printer and filament, read that filament's Safety Data Sheet, and follow both the SDS and the printer maker's instructions. A filament being PLA, or a print having no noticeable smell, is not evidence that its emissions are low.
• IF the plan is "the printer is enclosed, so emissions are handled," THEN check what the enclosure actually does. An enclosure that only holds heat in (for warping, see variations) does not by itself remove particles or vapor. NIOSH describes the enclosures that reduced exposure as ones that exhaust outdoors or through filters while keeping temperatures within the maker's operating specification, cites a UK HSE finding that waiting about 20 minutes after printing before opening the enclosure reduces exposure, and says a qualified person should assess any retrofit enclosure or exhaust so it does not increase fire risk or void the printer's approval or warranty [3P29, section 4.4].
• concrete, shop-scale numbers: most shop jigs and fixtures use well under 50 g of filament (roughly a dollar of PLA/PETG) [3P19].

✦ matching — for the app's read

• does: builds a solid part, additively, layer by layer, from a digital (CAD/CAM) model — jigs, fixtures, brackets, organizers, custom hardware, prototypes, and templates [3P15][3P19].
• suits: shop-tooling fabrication — hold-down clamps, cam clamps, dust-shoe adapters, wasteboard-leveling jigs with a dial-indicator mount, bit/tool organizers, touch-off blocks for consistent Z-zero, router-table inserts with a custom bushing recess (the dumbau/LumberJocks case above), soft vise jaws, center-finders, and hose-clamp fittings for non-standard hose profiles [3P19][3P16][3P15] (for clamps and hold-downs on a powered machine, see the load-rating caveat in correct use). General prototyping — testing a part's fit or geometry before committing good stock to it.
• kit-fit:
  • pairs with: CNC Router and Laser Cutter (Laser Engraver) as the shop's third digital-fabrication tool — commonly used together (e.g., a laser-cut or CNC-routed jig body paired with a 3D-printed bushing or bracket) rather than as substitutes for one another [CNC-cross][LC-cross].
  • substitutes for: hand-fabricated jigs and fixtures made from scrap wood, hardware-store brackets, and off-the-shelf organizational hardware, when a custom geometry is faster to design and print than to fabricate by hand.
  • cannot replace: CNC Router or a table saw for actual furniture-grade structural wood parts. This is the entry's central lineage-misuse point, grounded in tested data rather than a found misuse anecdote (see honest gap below): a well-known maker's own tested comparison found that 100%-infill PLA (91 MPa) and PETG (72 MPa) can match or exceed some wood products in raw bending strength (spruce 72 MPa, MDF 23 MPa, particleboard 10 MPa) — but the ranking inverts on strength-per-weight, where spruce/balsa/plywood win decisively, and solid wood is more than twice as stiff as PLA/PETG at an equal cross-section [3P22]. A printed part that tests strong in a bend test is not a like-for-like substitute for a structural wood part.
• not-for (project-keyed): don't reach for a 3D printer when the job needs a genuine structural wood member — see cannot replace above. Don't reach for PLA when the part will sit in shop heat or under sustained mechanical load — see correct use's IF/THEN material guidance. Don't expect an FDM printer to do what a laser or router does — it cannot ablate or mechanically cut existing stock; it only adds material (see lineage-misuse).
• access: starting a simple print from an existing model (Thingiverse, Printables) is close to push-button on modern consumer machines. Designing a functional jig from scratch is a real, separate skill (CAD literacy) with its own learning curve — several of this entry's own named sources (Hornberger, Ted Wakefield) describe learning CAD specifically to make shop tooling, not as something that came free with owning the printer [3P15]. Setup and operating gate, before the first print: put the printer where its emissions are controlled for this printer and filament (ventilation or a filtered or exhausted enclosure, per the filament's SDS and the maker's instructions), clear of combustibles, with every maker-supplied guard and cover fitted; let the nozzle and bed cool, then switch off and unplug before reaching in, except for a maintenance step the maker's manual expressly prescribes with power on; stop and unplug at any damage, failed fan, smoke, melting or burning smell; and check the make and model against current recalls before use (see correct use and dangerous) [3P29][3P24][3P26].
• revival: not applicable in the "old tool rediscovered" sense this corpus uses for hand tools — the desktop FDM printer is a genuinely new (post-2009) tool class, not a revived older one. Its own "revival" story is closer to a democratization story: a ~€30,000-class machine becoming a few-hundred-dollar desktop tool within a decade of its founding patent's expiry [3P8][3P6].

anatomy

• extruder — cold end + hot end. The "cold end" is a stepper motor with a hobbed drive gear that grips and feeds filament; the "hot end" (heater cartridge, thermistor, nozzle) is where the filament actually melts [3P17].
• direct-drive vs. Bowden configuration. Direct-drive mounts the cold end and hot end together on the moving print head. Bowden separates them — a stationary cold end pushes filament through a long PTFE tube to a lighter, remote hot end, reducing the moving head's mass (see lineage for the Bowden fix's own 2009 origin) [3P17].
• heat sink / fan on the hot end — prevents "heat creep" (filament melting prematurely above the nozzle, a named jam cause, especially with PLA) [3P17].
• nozzle — interchangeable, commonly 0.4 mm [3P17].
• build plate / bed, often heated — a heated bed counters warping from thermal contraction as printed plastic cools; an unheated bed largely restricts practical material choice to PLA. Bed surfaces (PEI, BuildTak) aid first-layer adhesion and part release [3P17][3P18].
• motion system — stepper motors (commonly NEMA 17, 200 steps/revolution) driving belts and leadscrews along Cartesian X/Y/Z gantry axes, or three converging arms on delta-style machines [3P17].
• This anatomy directly sources the mechanical account CNC Router's own entry flagged as an unsourced inference: FDM deposits material layer by layer from a moving nozzle onto a bed, exerting no mechanical cutting force on any prior layer — the structural reason an FDM printer is additive and "cannot gouge material the way a crashed router bit can" [CNC-cross][3P17].

variations (living)

Subtype-first, per this corpus's own matching discipline — the printer's build volume, drive configuration, and enclosure type each change what it's actually suited for; treat them separately rather than as one generic "3D printer."

• Cartesian FDM (the dominant desktop form) — X/Y/Z gantry axes; the RepRap/Prusa-descended mainstream configuration [3P11][3P17].
• Delta FDM — three converging arms instead of a Cartesian gantry; favors tall, narrow build volumes and speed over the widest range of part geometries [3P17].
• Direct-drive vs. Bowden — see anatomy; direct-drive suits flexible filaments (TPU) better, Bowden favors lighter, faster head movement for rigid filaments [3P17].
• Open-frame vs. enclosed — an enclosure holds ambient chamber temperature up, which matters for warping-prone materials (ABS, ASA, nylon); an open frame is adequate for PLA/PETG but not for enclosure-dependent materials (see matching) [3P18][3P12]. This is heat retention for warping; an enclosure that only holds heat is not by itself emission control (see correct use) [3P29].
• SLA / resin printers — a genuinely different process, not a "type of" FDM printer (see lineage and lineage-misuse): cures liquid photopolymer with UV light rather than extruding thermoplastic. Mentioned here for contrast, not built out as a subtype of this entry — a full SLA/resin entry, if built, would need its own anatomy, materials, and misuse classes rather than inheriting this one's [3P1][3P2].

trades

• Cabinetmaking / fine furniture / hand-tool and machine woodworking — shop-tooling fabrication (jigs, fixtures, bushings, organizers) is the dominant documented woodworking use; the tool augments the shop rather than replacing any of its structural-fabrication tools [3P15][3P16][3P19].
• Product design / rapid prototyping — the tool's original and still-primary commercial and hobbyist use case, from which the woodworking-shop application (jig/fixture printing) is a downstream adoption [3P6][3P9].
• Digital fabrication / makerspace — alongside CNC Router and Laser Cutter (Laser Engraver) as the third leg of a shop's computer-controlled toolset [CNC-cross][LC-cross].

lineage

Ancestors

• No direct mechanical ancestor in the woodworking tool tradition — FDM is a 20th-century invention with no hand-tool or hand-powered-machine lineage the way a CNC router traces back through the router to the hand plane, or a laser traces back (loosely) through optical/energy tools. Its nearest true ancestor is the broader idea of computer-controlled, CAD/CAM-driven fabrication that it shares with CNC Router [CNC-cross].
• Stereolithography (SLA) predates FDM by several years and is a genuinely different process, not a predecessor mechanism. Chuck Hull filed US Patent 4,575,330 ("Apparatus for production of three-dimensional objects by stereolithography") on August 8, 1984, granted March 11, 1986; he co-founded 3D Systems in 1986, which shipped the first commercial 3D printer, the SLA-1, in 1987 [3P1]. SLA cures a liquid photopolymer vat layer-by-layer with a UV laser — a fundamentally different physical process from FDM's molten-thermoplastic extrusion, not an earlier version of the same mechanism. Honest secondary-sourced note: French researchers Alain Le Méhauté, Olivier de Witte, and Jean-Claude André reportedly filed a competing stereolithography patent in France in January 1984, ahead of Hull's filing, but their employer (Cilas Alcatel) abandoned it and never commercialized it [3P3] — not independently verified against a French patent register this run, carried at lower confidence.

The founding invention

• Fused Deposition Modeling — S. Scott Crump / Stratasys. US Patent 5,121,329, "Apparatus and Method for Creating Three-Dimensional Objects," inventor S. Scott Crump of Minnetonka, MN, assignee Stratasys, Inc.; filed October 30, 1989; granted June 9, 1992 — read directly in full text [3P4]. Crump and his wife Lisa Crump founded Stratasys in 1989, the same year the patent was filed [3P5]. The patent's own References Cited section names Hull's SLA patent as prior art, a direct primary-source link between the two technologies.

Descendants — an open-hardware-driven Petroski case, not a single inventor's fix. (Distinct in structure from most of this corpus's Petroski cases: the "failure" the descendants fix isn't a technical flaw in the original patent — it's the patent itself, as a barrier to affordability, expiring.)

• The RepRap project (2004–2008) — the direct cause of the desktop-printer explosion, independently museum-documented. Adrian Bowyer, a mechanical engineering lecturer at the University of Bath, introduced the RepRap concept online in February 2004 and launched the funded project in summer 2005 (~£20,000 EPSRC funding), explicitly targeting a self-replicating machine costing roughly €350 in materials against a then-typical commercial 3D printer price of ~€30,000, released under the GNU GPL [3P6]. RepRap 0.2 printed its first self-part on September 13, 2006; RepRap 1.0 "Darwin" made over half its own parts by February 9, 2008 [3P6]. Bowyer and Vik Olliver displayed the first complete "child" machine made by its "parent" at 14:00 UTC on May 29, 2008, at Bath University [3P6]. The original Darwin RepRap (built by Bowyer and Ed Sells, Bath, 2007) is held by the Science Museum Group, London — accession object 2013-59, read directly [3P7].
• The Stratasys FDM patent expired in 2009 — twenty years after its 1989 filing — and this is directly, causally linked in sourced reporting to the affordable-desktop-printer explosion that followed [3P8]. The same year, MakerBot was founded on January 20, 2009 by Bre Pettis, Adam Mayer, and Zach "Hoeken" Smith, who met at the NYC Resistor hackerspace, explicitly "to build on the early progress of the RepRap Project" [3P9]. Their first machine, the Cupcake CNC, was finished for SXSW in March 2009, with the first batch of 20 units shipped by mid-April 2009 [3P9]. MakerBot's second product, the Thing-O-Matic (2010), is held by the V&A (London) — accession CD.27-2018, read directly — the museum's own catalog text calls it "the first desktop 3D-printer to be widely manufactured at a relatively affordable price for consumers," priced at $1,225 at its September 2010 Maker Faire NYC launch, discontinued spring 2012 when the Replicator succeeded it [3P10].
• Josef Průša / Prusa Research — the design that made the RepRap lineage mainstream. Průša joined the open-source RepRap project as a student in 2009; he designed the Prusa i3 in 2012 as a simplified RepRap Mendel derivative (print time for its own parts cut from roughly 20 to 10 hours; dual Z-motors added; 3D-printed bushings), founded Prusa Research the same year, and the i3 was reported as the world's most-used 3D printer by 2016 [3P11].
• The heated build chamber — a named, dated, expired patent directly targeting warping and delamination, FDM's dominant print-failure mode. US Patent 6,722,872 B1, "High Temperature Modeling Apparatus," filed June 2000, granted April 2004; inventors William Swanson, Patrick Turley, Paul Leavitt, Peter Karwoski, Joseph LaBossiere, and Robert Skubic, assignee Stratasys. It patented heating the entire build chamber — not just the bed — while isolating that heat from motors and electronics. The patent expired February 27, 2021, reported in industry press the same week [3P12].
• The Bowden extruder — a real, dated, named open-hardware fix, not a patent. RepRap Blog user "eD" posted "Bowden extruder concept" on April 18, 2009, stating the problem directly: the moving hot end and its drive motor added too much mass and bulk to the print-head gantry. The fix, credited in the same post to "follow Forrest[ Higgs's] previous attempts," decouples the drive motor from the hot end via a PTFE-lined Bowden-cable sleeve, so only the lightweight hot end moves; Forrest Higgs himself commented on the same post confirming his own prior attempts [3P13]. Honest gap: no dedicated patent for either the Bowden or direct-drive extruder configuration was located despite direct searching — this is an open-hardware-community innovation, not a patented one, and this entry states that plainly rather than inventing a patent citation.

Contemporaries

• CNC Router — the subtractive sibling: removes material by spinning-bit contact rather than depositing it. Shares CAD/CAM-driven, computer-controlled operation with the 3D printer but is mechanically opposite [CNC-cross].
• Laser Cutter (Laser Engraver) — the non-contact sibling: ablates or cuts via a focused energy beam rather than mechanical contact (subtractive) or material deposition (additive). Together the three form the shop's "digital fabrication" trio [LC-cross].
• SLA / resin printing — the other major 3D-printing process (see ancestors above); shares "3D printer" as a category name and a build-by-computer-model workflow, but is chemically and mechanically unrelated to FDM. A genuine naming-confusion risk for a shop buyer — see lineage-misuse.

history

• The FDM/SLA founding decade, both patents directly read. Crump's FDM patent (US 5,121,329, filed 1989, granted 1992) and Hull's SLA patent (US 4,575,330, filed 1984, granted 1986) were both fetched and read in this or the companion research pass; Crump's own patent cites Hull's as prior art, the direct documentary link between the two technologies [3P1][3P4].
• The RepRap decade (2004–2012) is the entry's real historical spine — an open-hardware movement with primary-source dating throughout, not a single company's product history. From Bowyer's 2004 concept post through the 2008 self-replication milestone, the 2009 patent expiry, MakerBot's founding the same year, and Průša's 2012 mainstreaming design, the chain is documented at every step with dates and named people, not inferred [3P6][3P8][3P9][3P11].
• A peer-reviewed academic record exists for the project itself, not just its own wiki: Jones, Haufe, Sells, Iravani, Olliver, Palmer, and Bowyer, "RepRap – The Replicating Rapid Prototyper," Robotica vol. 29 (2011), pp. 177–191, Cambridge University Press [3P14].
• Honest gap: Bowyer's 2019 New Year Honours MBE "for services to 3D printing" is reported across multiple outlets but was not independently verified against a primary UK honours-list document this run — carried at lower confidence.

✦ human side

• Marius Hornberger — a named, dated woodworker's own conversion account, with public design files. Hackaday (Dan Maloney), Dec 20, 2018, profiled Hornberger, a woodworker running a shop-focused YouTube channel, describing himself as a 3D-printing skeptic three years earlier who adopted it for soft vise jaws, a center-finder for round/square stock, a custom dust-collection hose clamp for spiral-reinforced hose (his own design solving a fit problem standard band clamps don't handle), and switch panels — with his design files posted publicly on Thingiverse [3P15]. This is a named, dated, first-person skeptic-to-user conversion account with a public, checkable body of work — the kind of source this corpus values most.
• A named forum design closing the loop back to this entry's own citing sibling, CNC Router. On the LumberJocks thread "3D Printing for the Woodworking Shop" (thread 303421), a member printed a Bosch router-table insert with a recessed bushing mount for a Leigh dovetail jig, calling it "the perfect application of 3D printing for the woodworking shop." The underlying model, "Bosch Router Table Insert with Bushing Recess," is by Thingiverse user dumbau, parametric and Customizer-adjustable [3P16]. Sourcing note: the forum's own JS-gated rendering meant this was confirmed at search-snippet depth, not independently re-fetched as full thread text — held at slightly lower confidence than a directly-fetched forum page elsewhere in this corpus.
• Named Hackaday commenters, dated, with specific real shop-use accounts (lower authority than a full article, but genuinely named and dated): commenter "Weirdwhit" (Dec 20, 2018) describes printing a hole-saw bushing to guide a hole saw where a pilot bit couldn't be used; commenter "Ted Wakefield" (Dec 23, 2020) describes buying a hobby CNC router, learning FreeCAD to design jigs, and later preferring a 3D printer over a bigger CNC for shop-part fabrication [3P15-comments].
• RepRap's own community, self-documented rather than told secondhand. Unlike several entries in this corpus that rely on a manufacturer's retrospective account of its own founding, RepRap's history is unusually well self-documented in near-real time on its own wiki and blog — including the Bowden-extruder fix (see lineage) posted by its own inventor within days of the idea, not recalled years later [3P13].

✦ proof of existence

• Two independently verified, directly-read museum accession records — a genuinely strong proof-of-existence pair for a tool class barely two decades old. The Science Museum Group (London) holds the original Bowyer/Sells "Darwin" RepRap self-replicating 3D printer (2007) — accession object 2013-59 [3P7]. The V&A (London) holds a MakerBot Thing-O-Matic (2010) — accession CD.27-2018, on display in the Design 1900 to Now gallery, Room 76 — described in the museum's own catalog as "the first desktop 3D-printer to be widely manufactured at a relatively affordable price for consumers" [3P10]. Together these bracket the entire RepRap-to-affordable-desktop-printer arc with two national-museum-grade records, independent of any manufacturer's own account of its history.
• Two foundational US patents, both directly read. S. Scott Crump's US 5,121,329 (1989/1992, the founding FDM patent) and the Swanson/Turley/Leavitt/Karwoski/LaBossiere/Skubic heated-build-chamber patent, US 6,722,872 B1 (2000/2004, expired 2021) [3P4][3P12].
• A dated, named, first-hand open-hardware design record, distinct from a patent but no less a documentary record: the RepRap Blog's own April 18, 2009 Bowden-extruder post, with its inventor's own contemporaneous account [3P13].
• A peer-reviewed academic record of the project itself: Jones, Haufe, Sells, Iravani, Olliver, Palmer, and Bowyer, "RepRap – The Replicating Rapid Prototyper," Robotica vol. 29 (2011), Cambridge University Press [3P14].
• Honest gaps, named directly: no Smithsonian collection record specifically for a 3D printer as a technology-history object was confirmed this run (the Smithsonian's Digitization Program Office 3D-scans historical artifacts for printing, a different thing entirely, and this was not treated as a substitute); the Le Méhauté/de Witte/André competing 1984 French SLA patent claim was not independently verified against a French patent register; Bowyer's reported 2019 MBE was not independently verified against a primary UK honours-list document.

references

The reviewer's corrections (3)

• BLOCKER — F1. Omitted exposure controls: indoor printing and ABS/ASA recommended without a ventilation or emissions gate.
• BLOCKER — F2. The incident narrative gave no usable physical-safety controls.
• MAJOR — F3. Wrong-process source (a resin SL1/SL1S article) and an overbroad first-layer instruction.

Limits the reviewer named (they travel with this entry): The reviewer did not independently certify every historical statement, the patent-expiry causal claim, museum accessions, fire reports or material-test numbers. Two sources could not be read in full: the PMC paper (a browser challenge) and the CPSC recall URL. The fixer read the NIOSH guide at summary depth. This is a bounded review of named findings, not a whole-entry certification.