-
Notifications
You must be signed in to change notification settings - Fork 1
FDM Part XII Cross cutting workflows
FDM Polymers — A Technical Reference ›
Material selection, calibration, bed adhesion, multi-material printing, post-processing, cost/procurement, tribological filaments, fiber-reinforcement fundamentals, moisture management, testing and qualification, and design for FDM — eleven synthesis chapters that consolidate the per-polymer guidance scattered through Parts II–XI into single-purpose references. Read these once after the polymer chapters; return to them as workflow questions arise.
Chapters: 22 Material selection · 23 Calibration · 24 Bed adhesion · 25 Multi-material · 26 Post-processing · 27 Cost and procurement · 28 Tribological filaments · 29 Fiber fundamentals · 30 Moisture management · 31 Testing and qualification · 32 Design for FDM
Polymer selection is the engineering decision that constrains every downstream choice on a print — process parameters, hardware, post-processing, cost, and ultimately service performance. The framework below is constraint-first: identify the hard requirements, intersect the candidate polymer sets, then refine by cost, printability, and available hardware. When two polymers tie on the binding constraint, pick the cheaper or more printable one — the volume's chapters cover when the second-tier choice is the right one to spend on.
Service temperature is the first filter. The polymer's glass transition (for amorphous polymers) or melting/crystallization envelope (for semi-crystalline) sets a hard ceiling on continuous-load service: amorphous parts loaded above Tg creep; for semi-crystalline polymers the ceiling is set by HDT and creep, typically far below Tm — use the continuous-service column in Appendix A.1. Mechanical character is the second filter: stiffness vs toughness vs flexibility. Most FDM polymers are stiff (PLA, PETG, PC, nylons, PPA-CF); a minority are tough (PCTG, PC blend, PA12, PEBA); a few are flexible (TPU, TPEE, PEBA). Environment is the third filter: UV exposure, chemical contact, moisture, fuels, food contact, biocompatibility. Cost and printability are the fourth filter — usually a tie-breaker rather than a primary constraint, but decisive when the application has no special requirement on the other three axes.
The sequence below answers "which polymer" by elimination. Evaluate all constraints that apply, mark every yes answer, then intersect the candidate sets. If only one hard constraint applies, the listed set is the starting answer; if none apply, use the default-tier choices at the end.
| Step | Question | If yes -> | If no -> |
|---|---|---|---|
| 1 | Does the part need to be flexible (Shore D < 70)? | TPU, TPEE, PEBA, foaming elastomer (Ch 16) | Continue to step 2 |
| 2 | Does the part see continuous service above 150 °C? | PPS-CF up to ~180 °C; above that, PAEK or outsource (Ch 18–19) | Continue to step 3 |
| 3 | Does the part see continuous service above 100 °C? | PPA-CF, PC blend (high-PC), PPS-CF (Ch 14, 15, 18) | Continue to step 4 |
| 4 | Outdoor UV exposure over months or years? | ASA, PMMA, PVDF (Ch 10, 17) | Continue to step 5 |
| 5 | Aggressive chemical exposure (acids, bases, hydrocarbons, fuels)? | PP, PVDF, PPS, POM by chemistry (Ch 11, 17, 18) | Continue to step 6 |
| 6 | Food or water contact? | copolyester (PCTG, or Tritan-based filament — chemically a distinct TMCD terpolymer, see §8.1; preferred — subject to the §8.9 caveat), PP, PETG (Ch 8, 11) | Continue to step 7 |
| 7 | Repeated impact loading or living-hinge cycling? | PCTG, PA6/PA12, PP unfilled, PEBA (Ch 8, 13, 11, 16) | Continue to step 8 |
| 8 | High stiffness-to-weight for structural parts? | PA-CF, PPA-CF, PC-CF, PPS-CF (Ch 13, 14, 15, 18) | Continue to step 9 |
| 9 | Wear, friction, or sliding-contact applications? | POM, PC/PTFE, iglidur tribopolymers (Ch 17, 15, 28) | Continue to step 10 |
| 10 | ESD-dissipative surface required? | ESD-PC (Ch 15) | Continue to step 11 |
| 11 | Flame retardance (UL94 V-0) required? | FR-PC, PPS, PEI (Ch 15, 18) | Continue to step 12 |
| 12 | Optical clarity required? | PCTG (Ch 8), PETG clear, PMMA, PC clear (Ch 7, 17, 15) | Continue to step 13 |
| 13 | None of the above special requirements apply? | PETG (cost), PLA (printability), PCTG (toughness step-up) | — |
Table 22.1 — Material selection decision walkthrough. The order of the steps reflects constraint dominance — the hardest-to-satisfy requirements come first — but real parts often answer yes to more than one row. Mark every applicable row, intersect the candidate sets, and then choose by cost, printability, and available hardware. Step 13 applies only when no special requirement dominates; it handles the bulk of hobbyist work.
| Application class | Default polymer | Step-up if budget allows |
|---|---|---|
| Display models, cosmetic prints | PLA | PCTG for toughness, PETG for cost |
| Functional prototyping (room-temp) | PETG | PCTG for impact, PC blend for heat |
| Outdoor parts (UV, weather) | ASA | PMMA for clarity, PVDF for chemistry |
| Electronics enclosures (passive) | ASA or PETG | PC blend for heat tolerance |
| Engine-bay / under-hood | PC blend | PPA-CF for stiffness + heat |
| Drone or RC airframe | PCTG or PP-CF | PA6-CF or PPA-CF for performance |
| Lab equipment, chemical contact | PP | PVDF for aggressive chemistry |
| Living hinges, snap-fits (high cycle) | PP unfilled, PA12 | PEBA for dynamic flex |
| Gaskets, vibration dampers | TPU 95A | TPEE for heat, PEBA for dynamic flex |
| Wear surfaces, low-friction bushings | POM | PC/PTFE for load-bearing stiffness and heat, PEEK for heat + chemistry |
| Structural brackets, fixtures | PETG or PC blend | PA6-CF or PPA-CF for stiffness |
| Food contact (resin-level compliance) | Copolyester (PCTG, or Tritan-based filament) | Verify per-filament certification |
| Athletic footwear, cushioning | PEBA or foaming TPU | — |
| Aerospace, FR-rated electronics | FR-PC | PEI if hardware tier supports it |
Table 22.2 — Application-to-polymer quick reference. Use as a starting point for procurement decisions; the per-polymer chapters in Parts II–XI carry the engineering detail needed to confirm fit and the brand surveys needed for purchasing.
Every selection this chapter's framework produces ultimately rests on datasheet values, so the ways datasheets mislead belong here. Three failure patterns recur across every filled and engineering polymer family; the family chapters carry only the deltas — §13.7 for the nylons (moisture compounds the gap toward a factor of two), §14.11 for the PPA family (the gap widens on the highest claims, and fiber loading plateaus above ~20%) — and the same cautions apply unmodified to the PC family (Chapter 15) and PPS (Chapter 18).
Published stiffness is a ceiling, not an expectation. TDS modulus is derived from optimally printed or molded, fully dense specimens; a real part carries layer-line porosity and imperfect fiber alignment, and printed-part stiffness lands below the datasheet figure across brands and families. Plan around a 20–30% reduction for design, more for moisture-sensitive matrices, then confirm against a specimen printed and conditioned the way the real part will be (Chapter 31). And compare like with like before comparing at all: vendors publish both a tensile (Young's) modulus and a flexural (bending) modulus and headline whichever is larger — an honest comparison holds method (ISO 527 vs ISO 178), orientation, and dry/conditioned state constant. Mixing them manufactures agreements and scandals that are really units mismatches.
Heat figures diverge by test method, and not consistently in one direction. HDT at 0.45 MPa, HDT at 1.8 MPa, Vicat, and the deformation tests used by independent reviewers load the specimen differently and are not expected to agree (§1.4); a product can measure above its TDS heat figure on one test and below on another without either number being wrong. A lone heat number means little without the method behind it, and none of them is a continuous-service rating — the continuous-service column in Appendix A.1, built from RTI-class data and creep behavior, is the design basis.
Brand moves the result more than the category label does. Two filaments sold under the same nominal class can rank one way on their datasheets and the opposite way once printed and measured: fiber loading, matrix grade, and sizing chemistry (§29.4) shift the printed envelope by 10–25%. Datasheet figures therefore cannot reliably rank products from different makers; where a vendor reports a property honestly as an orientation-dependent range rather than a single number, that range is the most accurate thing on the sheet. Independent test datasets (Appendix D.1) are most useful as a cross-check on relative rankings; your own machine (§23, Chapter 31) is the final authority on absolutes.
Every new filament — even a re-order of a previously calibrated brand and color — requires per-spool calibration on the actual machine before engineering-grade work. Resin batches drift, additive packages change between revisions, and printer state shifts with nozzle wear, extruder gear wear, and ambient conditions. The workflow below is the consolidated sequence the rest of this volume references; the sequence is order-sensitive — each step's output is the input to the next.
A word on scope before the workflow. Full calibration is worth the time when the print is functional — a part that bears load, mates with other parts to a tolerance, seals, runs hot, or will be tested or certified. For those prints the dimensional accuracy, mechanical strength, and repeatability the workflow buys are the whole point. For purely decorative or cosmetic prints — display models, figurines, visual prototypes — most of this sequence is overkill. A model that only has to look right does not need a measured extrusion multiplier or a tuned pressure-advance value; the vendor's generic profile, perhaps with a quick temperature check for surface finish, is sufficient, and the hours spent on flow and shrinkage calibration return nothing visible. The judgment is simply whether the print has a job to do beyond being looked at. The rest of this chapter assumes the answer is yes; if it is not, drying (Step 1) and a temperature check (Step 2) are the only steps that meaningfully affect a cosmetic result, and the remainder can be skipped without consequence. Calibration proves the process; Chapter 31 covers proving the part.
Drying is the first step because moisture confounds every measurement that follows. A wet filament shows artificially low max volumetric flow (steam disrupts melt cohesion), artificially low effective extrusion multiplier (voids in the bead reduce mass per unit length), and wildly inconsistent pressure advance values. The Part I §3.5 drying-protocol table is the reference; dry to the upper end of the recommended range and time, with 30 minutes margin beyond the spec to be safe on first calibration.
A temperature tower prints a single tall geometry with the nozzle temperature stepped down by 5 °C per 30 mm band, spanning the vendor's recommended range plus 5 °C above and below. Score each band on three axes: surface finish (smooth and consistent), bridging (no sag), and stringing (no fine filaments between features). For cosmetic parts, the optimal temperature is the lowest band that scores well on all three; for functional parts, use the highest band that still bridges without stringing — interlayer adhesion rises with melt temperature, and the tower does not test it. Stock temperature-tower model files exist on community model repositories; the specific tower geometry is less important than scoring the bands consistently.
Max volumetric flow (mm3/s) is the rate at which the hotend can melt and extrude filament without under-extrusion. The test prints a single-wall geometry with the flow rate stepped upward — typical bands span 5 mm3/s to 20 mm3/s in 1 mm3/s steps. The failure point is visible as a sudden transition from solid wall to thin, gappy, or visibly under-extruded bead. The measured maximum (MVFmax) is the highest band before that failure. Slice at no more than 0.8 × MVFmax for general work; engineering-grade work uses 0.6–0.7 × MVFmax — both fractions of the measured maximum, not of an already-derated value — to build process margin against drift.
Extrusion multiplier (EM, also called flow or flow ratio depending on the slicer) is the scaling factor applied to the slicer's calculated extrusion volume. The default value of 1.0 assumes perfect filament diameter, perfect extrusion stepper calibration, and zero melt-shrinkage during cooling — rarely all true simultaneously. The 12-sample wall measurement method: print a single-wall hollow cube at a known wall-width slicer setting (typical 0.45 mm for a 0.4 mm nozzle). After cooling, measure the actual wall thickness with calipers at 12 points distributed around the cube. Compute the average; divide the slicer's target wall width by the measured average to get the EM correction. Apply, re-print, re-measure to verify within ±0.5%. Typical converged EM values: 0.93 for high-fiber-loaded grades, 0.97–1.00 for unfilled engineering polymers (some unfilled copolyesters converge lower, near 0.94), 1.03–1.05 for softer elastomers and PC blends.
The YOLO method — a faster alternative for many users. The wall-measurement method has one real weakness: a caliper reading on a single ~0.45 mm wall is at the edge of what hand calipers resolve reliably, and the vase-mode print itself can vary in width with cooling and seam placement. The YOLO flow-rate test, built into OrcaSlicer (and available as community test models for other slicers), sidesteps the caliper entirely. It prints a single plate of small blocks, each sliced at a slightly different flow modifier — typically a range of -0.05 to +0.05 in steps of 0.01 — and the user picks the block with the cleanest top surface: the smoothest fill, no gaps between the surface-pattern arcs, and no raised or sunken seam between the inner and outer regions. The chosen modifier is applied as new = old ± modifier in a single pass. Because surface quality is judged rather than measured, YOLO is often the better choice for unfilled polymers on a well-behaved machine: it is faster, needs no calipers, and judging "is this surface smooth" is a more forgiving task than resolving hundredths of a millimetre on a thin wall. The wall-measurement method still earns its place where an absolute, traceable number is wanted — qualifying a new material, documenting a profile for publication, or calibrating fiber-filled and elastomeric grades whose top surfaces are textured enough that the visual judgment becomes ambiguous. A reasonable default: YOLO for routine per-spool tuning, the 12-sample measurement when the value has to be defensible.
Pressure advance (PA, sometimes Linear Advance) compensates for the elastic lag between extruder gear motion and nozzle bead deposition. Without PA, the bead width drifts at the start and end of every line — thin entries, thick exits. The test prints a single-layer pattern with the PA value stepped upward across the bed; the optimal value is the band where line ends and beginnings appear visually consistent with the rest of the line. Typical converged PA values: 0.020–0.040 for PLA, 0.030–0.060 for PETG/PCTG, 0.025–0.050 for PC blends, 0.04–0.08 for fiber-reinforced polymers.
Amorphous polymers shrink 0.3–0.5% in the print plane on cooling; semi-crystalline polymers can shrink 1.5–3% depending on crystallization behavior. The XY compensation factor in the slicer scales the model outward to compensate, producing dimensionally accurate parts on cool-down. The Califlower Mk2 model — a multi-feature shrinkage test with both external and internal dimensional checks — is the practical community-standard reference. Print, measure key dimensions, compute the average shrinkage as a percentage, set the slicer compensation. Typical converged values: 0.20% for PCTG, 0.25% for nylons (CoPA and CF-filled grades), 0.35% for ABS, 0.45% for ASA, 0.5% for filled PP; unfilled PA6 and unfilled PP can run ~1–2%, consistent with the semi-crystalline range above.
Z-direction shrinkage is typically smaller than XY because the layer-by-layer deposition allows partial relaxation between layers. The standard test is a 100 mm tall hollow cylinder; measure the actual height with calipers, compute the shrinkage percentage, set the slicer compensation. Many users skip this step on first calibration — the magnitude is usually under 0.3% for amorphous polymers, where engineering tolerance permits it. Skip with intention rather than by accident.
Calibrated values belong in the filament profile, not in the printer's persistent storage. Most slicers support per-filament storage of nozzle temperature, bed temperature, max volumetric flow, EM, PA, XY shrinkage, and chamber temperature. The pressure advance value can also be embedded in the filament-specific start G-code if the firmware supports it (the command differs by firmware: M900 K[value] on Marlin, M572 D0 S[value] on RepRapFirmware and Prusa Buddy firmware, and SET_PRESSURE_ADVANCE ADVANCE=[value] on Klipper). Store the profile, label it with the calibration date and the spool batch code, and re-verify EM and PA on first use of a new spool from the same brand — batch-to-batch drift of 5–10% is normal.
Bed adhesion is the interfacial chemistry problem framed in §3.3: wetting and intermolecular attraction between the molten first-layer polymer and the build surface. Polar polymers grip polar surfaces (PEI, glass, powder-coated steel); non-polar polymers (PP, PE) require non-polar compatible surfaces. The per-family chapters in Parts II–XI cover the specifics; this chapter consolidates the choices into one table.
| Polymer family | Best surface | Adhesive/release | Bed (°C) | Removal notes |
|---|---|---|---|---|
| PLA | smooth PEI | none | 50–60 | Cool fully; pops free |
| PETG | textured PEI | glue stick if over-gripping | 80–90 | Over-grips on smooth PEI; release layer essential |
| PCTG | smooth PEI | glue stick or PVP | 70–90 | Similar to PETG; release layer reduces sheet damage |
| ABS / ASA | smooth PEI | glue stick on first prints | 95–110 | Brim required for parts >100 mm; enclosure mandatory |
| HIPS | smooth PEI | glue stick | 100–110 | Limonene-soluble; brim for large parts |
| PP (unfilled, CF, GF) | PP-coated sheet or PP packing tape | Magigoo PP for difficult parts | 85–105 (PP sheet) / 80–100 (tape) / 20 (cold-bed) | PP-on-PP self-adhesion; cool fully for release |
| PE / HDPE | PE-coated sheet or PP packing tape | Magigoo PP | 80–100 | Same principle as PP; sparse commercial PE-sheet availability |
| PA6 / PA66 | G10 garolite | PEI + PVP as alternative | 90–110 | The garolite plate grips strongly; cool fully |
| CoPA (PA6/66 copolymer) | smooth PEI or CryoGrip Glacier | glue stick if over-gripping | 50–70 | Low-warp copolymer; the Appendix B bench profile runs the bed at 50 °C; garolite unnecessary |
| PA12 / PA612 / PA11 | smooth PEI | glue stick on demanding parts | 70–90 | Lower-shrinkage than PA6; PEI grips reliably |
| PA-CF / PA-GF | G10 garolite | Magigoo PA on PEI as alternative | 90–110 | High warp tendency tames best on garolite |
| PPA / PPA-CF / PPA-GF | smooth PEI + glue stick / PVP | Magigoo PC also works | 90–120 | G10 garolite acceptable; chamber temperature drives success more than surface |
| PC / PC blend | G10 garolite (long-term) | glue stick / PVP on PEI | 100–115 | Over-grips PEI catastrophically; release layer non-negotiable |
| PC-CF / PC-GF / ESD-PC | G10 garolite | Magigoo PC | 100–120 | Hardened nozzle mandatory; bed temperature near upper bound |
| PC/PTFE | smooth PEI | Magigoo PC | 90–120 | All-metal hotend required; chamber recommended |
| TPU / TPE | textured PEI or CryoGrip Glacier (40–50 °C) | glue stick as release if smooth PEI used | 40–70 (64D+ hard grades at the top end) | Over-grips smooth PEI; release layer if smooth surface needed |
| TPEE | textured PEI or CryoGrip Glacier | glue stick as release if smooth PEI used | 50–70 | Similar to TPU; higher bed than soft TPU |
| PEBA | smooth PEI | none | 50–60 | Releases cleanly without adhesive; easier than TPU |
| PMMA | smooth PEI | glue stick | 100–110 | Brittle; thermal stress dominates; cool fully |
| POM | glass + glue stick or POM-coated sheet | Magigoo PA | 100–115 | Low surface energy; brim mandatory; ventilation required |
| PVDF | smooth PEI | none for small parts; high-temp adhesive/release for large or warp-prone parts | 90–110 | All-metal hotend; chamber recommended for warp control |
| PPS-CF | G10 garolite | Magigoo PA | 80–120 | Chamber product-dependent: Polymaker and Flashforge print without a heated chamber, Bambu specifies 60–90 °C; hardened nozzle mandatory |
| PEI / PEEK / PEKK | industrial adhesive | beyond prosumer scope | 140–155 | Industrial chamber required; outside this volume's scope |
| PVA / BVOH | smooth PEI | none | 50–65 | Print directly from a dryer; ambient air degrades quickly |
| PVB | smooth PEI | glue stick | 70–80 | Hygroscopic; dry before printing; IPA-smoothable after |
| PHA / PLA-PHA | smooth PEI | glue stick if cold-bed | 0–60 | PHA prints cooler than PLA; some products require unheated bed |
| PCL | glass + glue stick | tape | 20–30 | Very low Tm; minimal bed heating needed |
Table 24.1 — Bed adhesion strategy by polymer family (consolidated reference). G10 garolite is the engineering default for any high-warp engineering polymer where the print would damage a PEI spring-steel sheet on removal; Magigoo's family of polymer-specific adhesives handles most remaining edge cases. Cold-bed approaches (PP, PCL, some PHA grades) use room-temperature bed during printing and elevated temperature only at end-of-print for release.
Three principles cut across the table. First, smooth PEI is the default surface for most polar polymers; textured PEI reduces grip and is preferred when over-grip damages the sheet on removal. Second, G10 garolite is the right answer for high-warp engineering polymers (PA6, PC, PPA, PPS) because it grips reliably during printing, releases cleanly on cool-down, and tolerates repeated thermal cycling without surface degradation. Third, polymer-specific adhesives (Magigoo PP, PA, PC) substantially outperform generic glue stick on the hardest filaments and are the engineering choice when print success rate is the metric.
Where Table 24.1 names an adhesive or release layer, the choice is rarely between brands that behave identically. Build-plate adhesives fall into four mechanical classes, and the class matters more than the label. Glue sticks are water-soluble solid adhesives (typically PVP-based, some PVA): cheap, available anywhere, and adequate for PLA and basic PETG, but they wear after a handful of prints and apply unevenly. Sprays coat a large bed quickly and uniformly — useful for big first layers — at the cost of overspray onto the machine and a need for ventilation. Liquid pen and brush adhesives are the purpose-built tier: formulated to grip while the bed is hot and release as it cools, lasting many prints per application. Temperature-activated adhesives are a liquid sub-type whose grip rises with bed temperature, giving strong hold on hot beds and easy release once cool. Table 24.2 surveys the products a prosumer user is likely to encounter.
| Product | Type | Material range | Notes |
|---|---|---|---|
| Magigoo Original | Liquid pen | PLA, ABS, PETG, HIPS, ASA, TPU | The default all-in-one. Pen applicator, holds hot and releases on cool-down, ~100+ applications per pen, cleans with water. The safest general-purpose choice for common filaments. |
| Magigoo PP / PA / PC | Liquid pen | Polymer-specific: PP, the nylon family, the PC family | Chemistry tuned per family. Magigoo PP is one of the few practical options for polypropylene; Magigoo PA and PC target the high-warp engineering polymers where generic glue stick fails. |
| Vision Miner Nano Polymer | Liquid brush | High-temp: PEEK, PEI, PPSU, PC, nylon; also PLA, PETG, ABS, HIPS, PVDF | The engineering-tier choice. 120 mL brush bottle, formulated for high-temperature materials, works on glass, PEI, and carbon surfaces; a single coat lasts many prints on lower-temp filaments. |
| Layerneer Bed Weld Original | Liquid | PLA, PETG, ABS, ASA, PVA, CPE — not nylon or PP | An aggressive adhesive aimed at stubborn corner-lift on large flat parts. The vendor explicitly does not recommend it for nylon or polypropylene. |
| Bambu Lab Liquid Glue | Liquid | PETG, TPU, and other common filaments | A clean liquid alternative to the glue stick; beginner-friendly, lower residue. Often used as a release layer on over-gripping plates. |
| TH3D Bed Cement | Liquid | Broad; works on PEI, flex plates, glass, garolite | 100 mL bottle with an applicator tip, priced near half the leading brands. Vendor guidance: reapply roughly every five to ten prints. |
| Dimafix | Temp-activated pen | ABS, ASA, PC, PP, nylon and other high-warp materials | Grip increases with bed temperature and falls away as the bed cools. Pen format; favored for warp-prone engineering filaments. |
| 3DLAC (and similar sprays) | Spray | PLA, PETG, ABS, ASA; nylon-specific variant available | Fast, uniform coverage on large beds. Overspray and ventilation are the trade-off; on very grippy stock plates it functions more as a release layer than an adhesive. |
| PVA glue stick (generic) | Solid stick | PLA, basic PETG; release layer for many materials | The budget baseline — washable, universally available, reapplied every few prints. A purple-tint stick shows coverage. Adequate for undemanding work and as a sacrificial release layer on PEI. |
Table 24.2 — Build-plate adhesives accessible to prosumer users. Material ranges are as stated by each manufacturer; treat them as starting guidance, since adhesion also depends on bed surface, temperature, and first-layer settings. Sealed-ecosystem and industrial-only products are out of scope. Prices and formulations drift — Appendix D frames the brand landscape as point-in-time.
Two selection rules cover most cases. For common filaments — PLA, PETG, ABS, ASA, TPU — an all-in-one liquid pen such as Magigoo Original is the lowest-friction choice, and a PVA glue stick is the budget fallback; both grip adequately and release on cool-down. For engineering filaments — the nylon, PC, PPA, and PPS families, and anything in a heated chamber — a polymer-specific or high-temperature adhesive earns its cost: Magigoo's family-specific pens, Dimafix for warp-prone materials, or Vision Miner Nano Polymer for the highest heat tiers. A point worth keeping in mind from Table 24.1: with the highest-warp engineering polymers, the adhesive is not a substitute for the right build surface and chamber temperature — it is the last few percent of reliability on top of a correct surface choice, not a rescue for a wrong one. Finally, every product here is also a release agent: on a build surface that grips a given polymer too hard (PC or PETG on smooth PEI is the classic case), a thin adhesive layer protects the sheet by giving the part something sacrificial to bond to.
Table 24.1 names a surface for each polymer; this section is the surface itself. Modern prosumer machines have largely converged on the flexible spring-steel system: a thin steel sheet, coated on one or both sides, held to the heated bed by an embedded magnet. The sheet flexes off the magnet after a print so parts pop free, and a machine's plate is really defined by its coating, not the steel. The coatings divide into six functional classes:
-
Smooth surfaces — smooth PEI (a polymer film) and smooth-PEI-like coatings — give glassy-flat, glossy first layers and the strongest grip on polar polymers. That grip is the catch: PC, PETG, and PCTG over-adhere to hot smooth PEI and can tear the coating or pull fragments on removal, which is why Table 24.1 pairs them with a release layer.
-
Textured surfaces — textured PEI, most commonly a powder-coated PEI steel sheet — trade gloss for a matte, lightly stippled first-layer finish and noticeably reduced grip, making them the safer default for over-gripping materials and for cosmetic parts where a matte underside is wanted.
-
Satin surfaces sit between the two: a fine, even micro-texture that yields a soft semi-matte finish with grip closer to smooth than to coarse textured — a middle option for users who find smooth too grippy and textured too coarse.
-
Patterned plates carry a decorative relief (woodgrain, geometric, marble-like) that transfers to the part's bottom face; they are a finish choice, with adhesion behavior tracking whatever base coating carries the pattern.
-
Engineering plates are the high-temperature, high-warp tier: rigid G10/garolite and equivalent purpose-made sheets, used bare for PA, PC, PPA, and PPS because they grip those polymers hot and release cleanly on cool-down without a consumable adhesive.
-
Cold plates are the inverse case — surfaces (and a workflow) for materials printed on an unheated or barely heated bed, such as polypropylene on a PP-faced sheet or PCL: adhesion is managed by surface chemistry and tape rather than bed heat, with the bed sometimes warmed only at end-of-print to aid release.
| Surface class | Typical finish | Grip | Best-fit materials |
|---|---|---|---|
| Smooth PEI | Glossy, glass-flat | High on polar polymers | PLA, PVB, PEBA; PETG/PCTG/PC only with a release layer |
| Textured PEI (powder-coated) | Matte, lightly stippled | Moderate | PETG, PCTG, TPU, ABS/ASA; over-grip-prone materials generally |
| Satin | Soft semi-matte | Moderate-high | PLA, PETG, ABS — a middle option between smooth and textured |
| Patterned / decorative | Relief pattern in part underside | Tracks base coating | Cosmetic prints; adhesion follows whatever coating carries the pattern |
| Engineering (G10 / garolite) | Fine matte | High on engineering polymers; clean cool-down release | PA, PA-CF/GF, PC family, PPA, PPS-CF |
| Cold plate (PP-faced, tape, glass) | Varies by facing | Chemistry- and tape-driven, not heat-driven | PP and PP-CF/GF, PCL, some PHA grades |
Table 24.3 — Build-plate surface classes. The class describes the coating's behavior; specific branded plates (below) are implementations of one or more of these classes. Grip is also a function of bed temperature and first-layer settings, so treat the column as a relative ranking rather than an absolute.
Within these classes, a few product families are common enough on prosumer hardware to name specifically. G10 / garolite sheets are a glass-epoxy laminate sold as flat rigid plates (Holden Enterprises and others) rather than flexible steel; they are the engineering-plate workhorse for the warp-prone polymers and tolerate repeated thermal cycling without coating wear. BIQU CryoGrip plates are flexible spring-steel sheets engineered around a cool-release effect — grip is strong while printing and falls away sharply as the plate cools, with the Glacier variant documented for TPU and for nylon at moderate bed temperatures; in slicer terms a CryoGrip sheet is treated as a high-temperature smooth-PEI-class plate and is kept within its rated temperature ceiling. Prusa ships three spring-steel sheets that map directly onto the classes above: a smooth PEI sheet, a textured powder-coated PEI sheet, and a satin sheet, plus a separate polypropylene-faced sheet for the cold-plate PP workflow. Bambu Lab plates follow the same pattern under different names — a textured PEI plate as the general-purpose default, a smooth/high-temperature plate for glossy first layers and higher-temperature engineering work, a cool-plate option for low-temperature materials, and an engineering plate intended for the warp-prone families. The naming differs by vendor, but the underlying surface classes in Table 24.3 are what actually determine behavior; matching the class to the polymer in Table 24.1 matters more than the brand on the box.
Three practical points. First, the coating is the consumable, not the steel: smooth PEI and powder-coated textured PEI both wear, and a release layer on over-gripping pairs (PC or PETG on hot smooth PEI) protects that coating directly — the §15.10 cost case for garolite on PC is exactly this. Second, surface class and bed temperature are coupled, not independent: the same plate grips harder hot, so a too-grippy result is sometimes a bed-temperature problem rather than a wrong-plate problem. Third, match the plate to the dominant material — a textured PEI sheet covers most common filaments, a bare engineering sheet earns its place the moment PA, PC, PPA, or PPS is in regular rotation, and a cold-plate or PP-faced sheet is effectively mandatory for polypropylene rather than optional.
Multi-material printing extends FDM beyond single-color, single-polymer parts to functional combinations: rigid bodies with flexible seals, structural bodies with soluble supports, color-coded mechanical assemblies, and cosmetic models with mixed transparency. The hardware ecosystems implementing multi-material capability split into four architectures, each with distinct constraints on what can be combined.
Single-extruder multi-material (MMU) systems share one nozzle across multiple filament feeds via an upstream switching mechanism. Filaments swap by retracting the active filament, advancing the new filament through the cutter or splicer, and purging the melt zone before resuming the print. The architecture is mechanically simple and cost-effective but introduces a substantial purge tax: typical purge volumes per swap are 50–200 mm3, accumulating quickly across a multi-color print to volumes that may exceed the model itself. Chamber compatibility between filaments is also a hard constraint — every loaded filament must tolerate the chamber temperature dictated by the highest-temperature material in the print.
Dual-hotend systems mount two independent hotends on a shared toolhead. One hotend lifts out of the way while the other prints. Filament swaps require only switching which hotend is active — no purge, no melt-zone contamination. The cost is mechanical complexity and additional calibration: hotend offset (XY and Z) must be characterized for each machine. Dual-hotend systems support combinations that MMU cannot, because the two filaments never share a melt zone — different temperature windows, different filler systems, even incompatible materials can be combined.
IDEX (Independent Dual EXtruder) systems mount two complete toolheads on independent gantries, enabling true parallel printing — two parts at once, or one part with simultaneous deposition of two materials. The cost-of-ownership and footprint are highest in this tier; the throughput advantage is meaningful for batch production.
Tool-changer systems mount several complete, independently docked toolheads that a single motion system picks up and parks as needed (Prusa XL-class machines). Like dual-hotend systems they are purge-free — each filament keeps its own dedicated melt zone — and they scale to four or more materials; the cost is per-tool XY and Z offset calibration and the price of each additional complete toolhead.
Filaments in any multi-material print must share a compatible chamber temperature envelope. Three practical tiers:
| Tier | Chamber temp | Compatible filaments |
|---|---|---|
| Low | ambient – 35 °C | PLA, PVA, BVOH, PVB, PHA, PCL — and TPU/PEBA with care |
| Mid | 35 – 50 °C | PETG, PCTG, ABS, ASA, HIPS, PA12, PA612, PP, PE — and PLA at the lower end |
| High | 50 – 65 °C (active) | PC, PC-CF, PA6, PA6-CF, PPA-CF, PPS-CF (PEI-CF only marginally; vendor spec is 85 °C+) — and ABS/ASA at the lower end |
Table 25.1 — Material compatibility groupings by chamber temperature. Mixing across tiers risks deforming the lower-temperature material (TPU softens above 50 °C; PLA softens above 55 °C). Filaments at tier boundaries may be mixed downward (run high-tier material at low-tier chamber if the geometry permits) but not upward. Note that vendor slicer compatibility schemes draw their tier lines differently — §11.9 reports schemes that class PETG and PCTG as "low-temperature" where this table's chamber bands place them in Mid; each taxonomy is internally consistent, and this table is the volume's own reference.
Support material is the most common multi-material application. Three strategies dominate:
Soluble supports (Chapter 20) — PVA paired with PLA, BVOH paired with PLA/PETG/ABS — dissolve away in a water bath after printing. Highest geometric capability (internal cavities, undercuts, trapped supports) at the highest cost (purge tax, dissolution time, filament cost).
Same-material supports use the same filament for both model and support, distinguishing them via geometry (sparse infill, thin walls, spaced from the model surface). No filament cost premium; removal requires mechanical separation. The default approach for single-extruder FDM without multi-material capability.
Breakaway interface supports — vendor-specific products like Polymaker PolySupport or specialty PPA-compatible breakaway filaments — print as the support body but with a deliberately weak bond to the model material. Snap apart on cool-down. The middle ground: lower cost than soluble supports, better surface finish than same-material supports.
Hybrid strategies are common in production work. A PC-blend model with HIPS body supports and HIPS interface layers — HIPS-on-PC is a verified weak pair, so the interface releases cleanly from the PC body without limonene dissolution, and the support volume cost is cheap HIPS. (A PCTG interface does not work here: PC and copolyesters bond strongly, so the interface welds rather than releases. For low-temperature work, the PETG/PLA pairing — either as model or as interface — is the analogous verified weak pair.)
Single-extruder multi-material printing on a four-color part with 100 filament swaps and a 100 mm3 purge per swap consumes 10,000 mm3 of purged material — roughly 12 g for a 1.24 g/cm3 polymer. For a small model the purge weight may exceed the model weight. Mitigations: reduce purge volume per swap (calibrate against actual cross-contamination; many systems use higher default purge than necessary); design the model to minimize swaps (color blocking rather than fine detail); use dual-hotend hardware which eliminates purge entirely for filament transitions; or accept the purge cost as the price of color capability.
Post-processing converts a raw FDM print into a finished part. The techniques cluster into five categories — mechanical, chemical smoothing, coating, heat treatment, and assembly — each with polymer-specific compatibility constraints. The single most important reality from the per-polymer chapters: the same chemistry that drives a polymer's engineering value typically forecloses solvent-based finishing on that polymer.
Sanding, polishing, and machining work on every FDM polymer with the right abrasive and the right technique. Three principles: wet over dry minimizes airborne particles (essential for CF/GF-reinforced grades where fiber fragments are documented respiratory hazards); progressive grit from coarse (220) through medium (400, 600) to fine (1000, 2000) produces the smoothest finish without skipping steps; polymer-specific temperature awareness matters — PLA, PETG, and PCTG smear under friction heat from power tools while POM, PC, and PPS tolerate aggressive sanding without deformation. For optical clarity on PCTG and PMMA: sand wet through 2000 grit, then polish with plastic-polish compound. For matte finish: stop at 800–1000 grit.
Vapor smoothing and bath smoothing in compatible solvents produce glass-smooth surfaces by surface-only re-melting the printed bead structure. The compatibility matrix is narrow:
| Polymer | Compatible solvent | Method | Hazard tier |
|---|---|---|---|
| ABS, ASA | Acetone | Vapor smoothing in closed container with brief warm-vapor exposure | Flammable; ventilation required |
| HIPS | Limonene; acetone with caution | Limonene bath dissolution/finishing; acetone can surface-finish but over-etches easily | Skin sensitizer (limonene); flammable (acetone) |
| PVB | Isopropyl alcohol | Vapor or bath; vendor-recommended workflow | Low; household-tier solvent |
| PLA | No safe household-tier solvent; DCM and ethyl acetate work | Heat-gun gloss possible; mechanical only for matte | DCM: see regulatory note below; flammable (ethyl acetate) |
| PETG, PCTG | No safe household-tier solvent; DCM works (same hazard tier as the PC row) | Mechanical preferred; XTC-3D / 2K epoxy for gloss | DCM: see regulatory note below; fume hood mandatory |
| PC, PC blend | Dichloromethane (DCM) | Bath or vapor; produces strong solvent-weld bond | DCM: see regulatory note below; fume hood mandatory |
| Nylons (PA, PPA) | No common workshop solvent works | Mechanical only; surface coatings for gloss | — |
| PP, PE | No workshop solvent at room temp | Mechanical only; flame treatment for paint adhesion | — |
| TPU, TPE, PEBA | No effective vapor solvent | Mechanical only; coatings limited by flexibility | — |
Table 26.1 — Solvent-smoothing compatibility by polymer family. Most engineering polymers in this volume have no practical workshop-solvent smoothing route — the chemical resistance that drives their engineering value forecloses chemical finishing. Regulatory note (DCM): under EPA's TSCA final rule on methylene chloride (89 FR 39254, May 2024), all consumer uses are prohibited — DCM is no longer lawfully sold to US consumers — and most industrial/commercial uses are prohibited or phasing out; a limited set of continuing uses operates under EPA's Workplace Chemical Protection Program, and EPA has adjusted some compliance dates since publication (e.g., extensions for certain non-federal laboratories, November 2025). Status checked July 2026 against EPA's methylene-chloride risk-management page (epa.gov/assessing-and-managing-chemicals-under-tsca/risk-management-methylene-chloride). OSHA regulates DCM as a potential occupational carcinogen (29 CFR 1910.1052); IARC classifies it 2A.
Mechanical finishing and coatings (2K epoxy, XTC-3D) are the practical routes for PETG, PCTG, nylons, polypropylene, and the high-performance specialty polymers.
Surface coatings cover three application classes: gloss and fill (XTC-3D, 2K epoxy clearcoats, plastic-bonding primers fill layer lines and produce smooth painted surfaces); color and aesthetics (automotive primers and topcoats over polymer-compatible bonding primers); function (conductive coatings for ESD, vapor barriers, anti-static treatments, UV protective clears for outdoor parts). Low-surface-energy polymers (PP, PE, POM) require flame or plasma treatment plus a plastics adhesion promoter (SEM- or Bulldog-class) before any coating will adhere reliably; standard automotive primers fail to wet these surfaces, and the 3M 4298UV / Loctite SF 770 primers are adhesive-bonding promoters (§26.5), not paint primers. Higher-surface-energy polymers (PLA, PETG, PCTG, ABS, ASA, PC) accept standard primer-topcoat workflows directly.
Annealing serves different purposes for amorphous and semi-crystalline polymers. For amorphous polymers (PETG, PCTG, ABS, ASA, PC), annealing relieves residual stress from rapid layer cooling — modest improvement to dimensional stability and creep resistance, no change to HDT. Temperature must stay below Tg by 10–15 °C to avoid distortion; typical schedules run 60–80 °C for the copolyesters and styrenics and 85–100 °C for PC blends (vendor PC schedules sit well below the theoretical Tg−10–15 ceiling of ~130 °C — see §15.11), for 2–4 hours. For semi-crystalline polymers (PLA, PA, PP, PPA, PEEK), annealing develops crystallinity — substantial gains in HDT and stiffness, dimensional change of 1–3% as the polymer chains pack into ordered regions. Schedule must stay above Tg but below Tm. Per-polymer detail in §3.6 and the polymer chapters; unfilled PPA anneals per its vendor protocol (§14.9), but without fiber reinforcement the warp risk during the heat soak is higher — support the part and follow the vendor schedule.
The volume's annealing schedules, collected from the polymer chapters and the vendor protocols they cite:
| Polymer | Purpose | Temperature (°C) | Time (h) | Support medium | Dimensional change | Expected gain |
|---|---|---|---|---|---|---|
| PLA / HTPLA | Crystallinity | 90–110 (HTPLA); 80–100 (standard PLA) | 0.5–2 | Flat plate; packed salt/sand for critical geometry | 1–3% shrink, anisotropic (§6.6) | HDT ~55 → ~120 °C (§3.6) |
| PETG / PCTG | Stress relief | 60–70 | 2–4 | Flat plate | minimal | Small creep/stability gain; no HDT change (§3.6) |
| ABS / ASA | Stress relief | 60–80 | 2–4 | Flat plate | minimal | Reduced delayed cracking at stress concentrators |
| PC blends | Stress relief | 85–100 (PolyMax 90 for 2 h; Bambu 85–100 for 6–12 h) | 2–12, per vendor | Flat plate | 0.3–0.5% shrink | Delayed-crack resistance (§15.11) |
| PC-PBT | Crystallinity (PBT phase) | per vendor TDS | per vendor TDS | Flat plate | grade-dependent | HDT and stiffness beyond stress relief (§15.11) |
| PPA-CF / GF (consumer) | Crystallinity | 100–120 (Bambu PPA-CF: 120–140) | 4–6 (Bambu: 6–12) | Packed salt/sand | grade-dependent; measure | HDT and Z-strength gains (§14.9) |
| PPA unfilled (Siraya) | Crystallinity | 80–100, natural cool-down | 4–8 | Packed salt/sand; robust geometry | higher warp risk than CF grades | HDT, mechanical, stability gains (§14.9) |
| PPS-CF | Crystallinity | 200 | 2–4 | Packed sand | grade-dependent; measure | Holds the 180–200 °C heat envelope (§18.1) |
| PEEK | Crystallinity | 140–200 | 2–4 | Packed sand | 1–3% shrink; thin-wall warp | Develops full crystallinity below industrial chamber temps (§19.1) |
Table 26.2 — Consolidated annealing schedules, gathered from §3.6, §6.6, §14.9, §15.11, §18.1, and §19.1. The semi-crystalline rows change the part's engineering envelope; the amorphous rows only relax it. Always anneal a dimensional test piece before production parts — shrinkage is geometry-, infill-, and orientation-dependent — and cool slowly (switched-off oven) to avoid trading old stress for new.
Solvent welding dissolves both mating surfaces in a compatible solvent, then evaporates to leave a polymer-continuous bond. Works for ABS (acetone), PC (DCM), and PMMA (DCM or specialty acrylic solvents) — noting that EPA's 2024 TSCA rule has ended lawful US consumer sale of DCM (see the Table 26.1 regulatory note). Adhesive bonding covers everything else: cyanoacrylate (CA) for fast assembly of most polar polymers; epoxy (2K) for high-strength structural bonds; polyurethane for elastomer-to-rigid bonds; specialty polyolefin systems (Loctite Plastics Bonding System — SF 770 polyolefin primer + CA; 3M 4298UV primer + CA; 3M Scotch-Weld DP8010 structural acrylic) for PP and PE bonding. Mechanical fastening avoids the adhesive problem entirely: heat-set threaded inserts (brass inserts with a knurled exterior that engages the printed plastic) work on PLA, PETG, PCTG, ABS, ASA, PC, nylons — any polymer with a reasonable melt window. Self-tapping screws work on fiber-reinforced filaments where the matrix grips threads reliably. Press-fit and snap-fit assemblies leverage polymer-specific elasticity: PETG, PCTG, PP, and nylons hold snap-fits through many cycles; PLA fractures, and printed PC snap-fits fail at layer lines and sharp internal corners despite the resin's ductility — radius the hinge root and orient the flexure in-plane. Chapter 32 consolidates these design rules — fits, inserts, compliant features, and orientation — by material.
Filament cost spans a roughly 30× range across the polymers in this volume — from $15/kg commodity PLA to $500/kg PEEK-CF. The cost tier drives procurement strategy as much as the technical specifications do.
| Tier | Range ($/kg) | Representative products |
|---|---|---|
| Commodity | 15–25 | PLA, PETG, generic ABS, generic ASA, HIPS, generic CoPA, eSun/Sunlu budget brands |
| Mid-engineering | 25–50 | PCTG (Spectrum, 3D-Fuel, Fiberlogy), PolyMax PETG, PolyMax PC, Bambu PC, Prusament PETG, ASA premium brands, unfilled PA12 |
| Engineering | 50–100 | Prusament PC Blend, PA-CF (Polymaker, Bambu), TPU 95A premium (NinjaTek, Polymaker) |
| Specialty | 100–250 | PA-CF premium, PPA-CF (Bambu, 3DXTech), PPS-CF, ESD-PC, FR-PC, PEBA (3DXTech 3DXLABS, Fillamentum), PVDF specialty grades |
| Ultra | 250–500+ | PEEK, PEKK, PEI 1010-CF, Prusament PC Space Grade Black, ThermaX PSU/PPSU, specialty industrial-grade products |
Table 27.1 — Filament price tiers (early 2026 retail; ±20% drift typical across regions and bulk quantities). The cost step from commodity to mid-engineering is roughly 2×; from mid-engineering to engineering is another 2×; from engineering to specialty is roughly 2–2.5×; from specialty to ultra is another 2–4×. Procurement decisions should map the cost tier to the application's binding constraint.
Three observable signals distinguish quality manufacturers from budget-tier producers, often more reliably than price alone:
Diameter tolerance. ±0.02 mm is the engineering-grade target for 1.75 mm filament. ±0.03 to ±0.05 mm is acceptable for commodity use; looser than ±0.05 mm produces visible extrusion-rate variation in the printed bead. Batch consistency. Vendors that publish lot codes and provide consistent color and mechanical properties across batches are the engineering-grade choice; budget brands often show batch-to-batch color drift and mechanical variance. TDS depth. Published TDS data — particularly printed-specimen tensile values, not just resin-grade values — distinguishes engineering-oriented vendors from marketing-oriented ones. The depth and specificity of the TDS correlates strongly with the underlying product consistency.
A growing set of vendors offer recycled-content products that the makers position as close to virgin material on the headline engineering metrics, though independent comparison is limited and reprocessing history affects any given batch. 3D-Fuel ReFuel reprocesses post-industrial PCTG, with the vendor reporting retention of the ISO 527 tensile envelope of virgin Pro PCTG. Fiberlogy R PP is 100% post-consumer/post-industrial polypropylene with vendor-documented mechanical properties matching virgin Fiberlogy PP. Braskem FL900PP-CF uses 100% recycled carbon fiber feedstock in the PP matrix. Fishy Filaments Porthcurno (sold in partnership with Fillamentum) is 100% ocean-recovered PA6 from end-of-life fishing nets (Part I §5.4). An adjacent sustainability program — not a recycled-content product — is Polymaker PolyTerra PLA: virgin bio-based PLA with carbon offsetting, shipped on a cardboard rather than plastic spool. These products are real progress over straight virgin filament; treat them as marginal improvements to mainstream procurement rather than as license to print wastefully.
Hobbyist or casual user: stick to commodity and mid-engineering tiers (PLA, PETG, PCTG, basic PC blend, TPU 95A) and one or two budget-tier brands plus one engineering-tier brand for reference. Buy in 1 kg spools; bulk procurement doesn't pay back unless the print volume supports it. Maker space or prototyping shop: standardize on engineering-tier brands (Prusament, Bambu, Polymaker, Spectrum) for predictable results; budget tier as a secondary option for cosmetic work. Bulk 2.5 kg spools and 5 kg refills are cost-effective. Engineering qualification or production work: specialty-tier and engineering-tier brands with published TDS data are mandatory; bulk procurement of specific batch codes preserves repeatability. The cost step up the tier ladder is usually worth it for parts that will be tested or certified.
Tribological filaments are engineered for parts whose binding constraint is wear at a moving interface — gears, bushings, cams, slides, bearing surfaces, guides. The relevant failure mode is not fracture under load or creep at temperature but the gradual loss of material and dimensional accuracy as two surfaces rub. The polymers covered in this chapter appear elsewhere in the volume under their chemistry families — iglidur tribo-grades are commonly inferred to be PA-based, though igus does not disclose the base polymer (Chapter 13), PC/PTFE is a polycarbonate composite (Chapter 15), POM is its own family (Chapter 17) — but the tribological use case cuts across all of them, which is why it earns a cross-cutting chapter rather than living inside any single polymer family. This chapter consolidates the wear-grade options and gives the vocabulary needed to choose between them.
Four parameters govern whether a printed part survives in a wear application. Specifying a tribo filament without understanding them is guesswork.
Coefficient of friction (COF). The ratio of friction force to normal force at a sliding interface, reported as a dynamic value (sliding) and a static value (breakaway). Low COF means less resistance, less frictional heating, and less energy lost in the mechanism. Engineering tribopolymers reach dynamic COF values of 0.10–0.25 against steel dry; unfilled engineering polymers like PA6 or PETG run 0.30–0.45. COF is not a fixed material constant — it depends on the counterface material, surface finish, load, sliding speed, and temperature — so vendor COF values are comparative indicators, not design allowables.
Wear rate (specific wear rate, k). The volume of material lost per unit of sliding distance per unit of normal load, with SI units of mm3/(N·m). It is the single most direct measure of how long a wear part will last. Tribologically optimized polymers report k values one to two orders of magnitude below their unfilled base resin. A part that loses 0.05 mm of wall over a service life is functional; the same part in a high-k material loses 0.5 mm and fails on clearance or backlash.
PV limit. The product of contact pressure (P) and sliding velocity (V) defines the thermal-mechanical envelope of a sliding part. Every tribopolymer has a PV limit above which frictional heat generation outruns heat dissipation, the contact surface softens, and wear accelerates catastrophically. The PV limit is the reason a bushing that runs cool at low speed fails at high speed under the same load. Design wear parts to operate well below the published PV limit; FDM-printed parts should be derated further because layer-line porosity reduces thermal conductivity and the effective contact area is lower than a molded equivalent.
Counterface and lubrication regime. Tribological performance is a property of the pair, not the polymer alone. A polymer sliding against hardened steel behaves differently than the same polymer against aluminum, against another polymer, or against itself. Dry-running (self-lubricating) tribopolymers are formulated with solid lubricants — PTFE, silicone, graphite, or molybdenum disulfide dispersed in the matrix — so they need no applied grease or oil; this is the dominant use case for printed wear parts, because applied lubricant attracts grit and many printed mechanisms cannot be relubricated in service. Lubricated service (with grease or oil) lowers wear further but is rarely the design intent for FDM parts. Polymer-on-polymer pairs should mix chemistries — POM against PA, for example — because identical polymers in sliding contact tend to gall and adhesively transfer material.
An anisotropy caveat specific to FDM. A printed wear surface is not isotropic. Layer lines create a directional texture; sliding across the layer lines wears differently than sliding along them, and a surface formed by the top or bottom layer differs from one formed by the perimeter walls. Wherever possible, orient the part so the wear surface is formed by smooth perimeter extrusion rather than stepped layer banding, and expect a break-in period during which high asperities are worn flat before the steady-state wear rate is reached.
The hobbyist-accessible tribological filament market clusters into four groups: dedicated tribopolymer compounds (the igus iglidur filament family), PTFE-modified composites of otherwise-conventional polymers, neat POM, and the carbon-fiber-reinforced engineering grades that deliver wear resistance as a side effect of stiffness.
| Filament | Base / type | Tribological character | Best use |
|---|---|---|---|
| igus iglidur I150 | Tribopolymer with solid lubricant (base polymer undisclosed; commonly inferred PA-class) | Dry-running; low wear against steel and against itself; 65 °C continuous limit; not the food-contact grade — igus positions i151 as the FDA / EU 10/2011-compliant, optically detectable grade | Gears, bushings, sliding parts in general service; specify i151 for food-adjacent work |
| igus iglidur I180 | Tribopolymer, higher-temperature grade | Dry-running; wear-rated similarly to I150; 100 °C continuous service vs I150's 65 °C; enclosure recommended for printing | Wear parts at service temperatures beyond I150's 65 °C limit |
| igus iglidur RW370 / J-class | High-temperature and specialty tribopolymer grades | Dry-running; extended PV envelope or temperature range per grade. RW370 is industrial-tier — ~350–360 °C nozzle, bed >180 °C, chamber ≥160 °C — far beyond the prosumer envelope, like the PEI/PEEK rows | Wear parts at elevated temperature or higher PV than I150 tolerates; RW370 only via industrial hardware or outsourcing |
| PC/PTFE (e.g. Spectrum) | Polycarbonate matrix + dispersed PTFE | Low COF from the PTFE phase; PC matrix carries structural load and heat | Load-bearing wear parts that also need PC-class stiffness and HDT |
| POM (acetal — Gizmo Dorks) | Neat polyoxymethylene | Low COF, excellent fatigue resistance, low and stable wear; no solid-lubricant additive needed | Gears, cams, low-friction mechanisms; the default printed-gear material |
| PETG-PTFE | PETG matrix + dispersed PTFE | Lower COF than neat PETG; modest wear improvement; easy to print | Light-duty sliding parts where PETG printability is wanted and loads are low |
| PA-CF wear grades | Carbon-fiber-reinforced nylon (PA6-CF, PA12-CF) | Wear resistance as a by-product of fiber stiffness; abrasive to the counterface | Stiff structural parts with incidental wear; not a first choice for pure bushings |
| PA + solid-lubricant blends | Nylon compounded with PTFE, MoS2, or graphite | Dry-running; between neat PA and dedicated tribopolymers on wear | General wear parts where a dedicated tribopolymer is unavailable |
Table 28.1 — Hobbyist-accessible tribological filaments. The igus iglidur filament family is the only group engineered specifically and solely for wear; the others deliver tribological performance as either a PTFE-additive modification (PC/PTFE, PETG-PTFE) or an inherent property of the base polymer (POM) or the reinforcement (PA-CF). Picking a CF-reinforced grade for a pure bushing application is a common error — the fiber that provides stiffness also abrades the metal shaft running inside it.
The iglidur filament family is igus's adaptation of its molded tribopolymer bushing materials to FDM filament form. The grades carry the same naming as the molded-part and bar-stock lines: I150 is the general-purpose dry-running grade and the most widely stocked (the igus i151 page explicitly says to choose I150 only if FDA food conformity is not necessary — i151 is the food-compliant, optically detectable blue grade); I180 extends continuous service temperature to 100 °C versus I150's 65 °C at the cost of more demanding processing (enclosure recommended); specialty grades extend the temperature or PV envelope further. All are formulated to run dry — the solid lubricant is dispersed through the polymer, so a freshly printed part is self-lubricating with no break-in grease required. iglidur filament is the engineering default when wear is the primary design constraint and the part does not also need to carry a heavy structural load.
PTFE-modified composites — PC/PTFE and PETG-PTFE — take a conventional matrix polymer and disperse PTFE through it to lower the coefficient of friction. They are not equivalent. PC/PTFE keeps polycarbonate's stiffness, impact toughness, and ~110–140 °C service envelope, so it suits wear parts that also carry load or see heat; it demands an all-metal hotend and the PC-class process discipline of §15.7. PETG-PTFE is a light-duty material — it prints as easily as PETG and lowers friction usefully, but the wear improvement over neat PETG is modest and the PETG matrix limits it to low loads and near-room-temperature service. Treat PETG-PTFE as a friction-reduction convenience, not a true bearing material.
POM deserves its standing as the default printed-gear material. It delivers a low and stable coefficient of friction, excellent fatigue resistance under cyclic tooth loading, and a low wear rate without any solid-lubricant additive — the tribological behavior is intrinsic to the acetal chemistry. The trade-offs are covered in §17.2: difficult bed adhesion and a genuine formaldehyde-emission hazard that mandates ventilation. Where those can be managed, POM outperforms every PTFE-modified composite on gear and cam duty. POM against PA — acetal gear on a nylon gear, or acetal on an iglidur bushing — is a particularly good polymer-on-polymer pairing because the dissimilar chemistries resist galling.
Match the material to the dominant duty. For gears and cams — cyclic tooth contact, moderate load, the failure mode is tooth wear and rising backlash — POM is the first choice, with iglidur I150 the alternative when POM's bed adhesion or ventilation requirements cannot be met. For plain bushings and sliding bearings — continuous rubbing against a shaft, the failure mode is bore wear and growing clearance — a dedicated iglidur grade is the engineering answer; the dry-running formulation is exactly the design intent. For load-bearing wear parts — a slide or wear pad that also carries structural load or sees heat — PC/PTFE is the choice because the polycarbonate matrix supplies stiffness and thermal headroom that neither POM nor iglidur matches. For light-duty low-friction parts — a drawer slide, a low-load guide, a part where smooth motion matters more than service life — PETG-PTFE is adequate and the easiest of the group to print. Avoid CF-reinforced grades for pure bushings: the exposed carbon fiber abrades a metal counterface, and the application is better served by a dedicated tribopolymer.
Three design practices materially extend printed wear-part life. Print the wear surface from perimeter walls, not layer banding — orient the part so the sliding face is formed by smooth perimeter extrusion, which gives a more uniform surface and a lower steady-state wear rate. Use high wall and infill density at the contact zone — internal porosity beneath a wear surface reduces both load capacity and thermal conduction, pushing the part toward its PV limit sooner; solid or near-solid infill under bearing surfaces is worth the material. Allow for break-in — printed wear surfaces shed high asperities during the first hours of service before reaching steady-state wear, so size clearances expecting a small initial dimensional change, and where the application is critical, run a deliberate low-load break-in before putting the part into full service.
Cross-references: iglidur calibration values appear in Appendix B and the PA-family process guidance in §13.5; PC/PTFE process and the mandatory all-metal-hotend requirement are in §15.7; POM bed adhesion, the formaldehyde-emission hazard, and ventilation requirements are in §17.2 and §5.3. The design practices above generalize beyond wear parts in Chapter 32.
Fiber reinforcement is the volume's most-repeated theme — it appears in the disclosure discussion of §2.5, the warp physics of §3.2, the nylon and PPA chapters (§13.4, §14.11), and nearly every reinforced-grade survey in between. This chapter consolidates the underlying mechanics so those sections can stay application-focused. One scope statement up front, restating §1.2: everything here concerns chopped short fiber compounded into filament. Continuous-fiber systems (Markforged-class hardware that lays unbroken carbon, glass, or aramid strand inside an FFF shell) are a different process in a different stiffness class with different design rules, and chopped-fiber intuition does not transfer to them; they are out of this volume's scope.
A short fiber stiffens a polymer by carrying load the matrix hands to it: stress transfers into the fiber through shear along the fiber–matrix interface. That transfer needs length to act on — below a critical fiber length (a property of the fiber, the matrix, and the interface strength), the fiber pulls out or simply rides along rather than carrying its share. This is the quiet limitation of every chopped-fiber filament: fibers enter the compounder as millimeters and leave the nozzle as fragments commonly in the tenths-of-a-millimeter class, shortened at every processing step — compounding, filament extrusion, and the final trip through a 0.4 mm nozzle — so much of the fiber content in a printed part works below full efficiency. It is one reason printed composite stiffness lands so far below what hand-calculated fiber fractions would predict, and why two products with identical stated loadings can perform differently (fiber length distributions are never disclosed).
Orientation is the second governing fact. Shear flow in the nozzle aligns fibers with the extrusion direction, so the stiffness gain concentrates along the printed bead path — this is why fiber-filled parts are even more anisotropic than unfilled prints (§3.1: Z at 40–60% of XY), and why slicer wall/infill direction choices matter more for composites. The fibers contribute essentially nothing across the layer interface. Voids complete the picture: fiber ends and fiber–matrix debonds nucleate micro-porosity that no slicer setting fully removes.
The process effects follow from the mechanics, and most of them are favorable — which is why fiber grades exist for polymers that barely print unfilled. Shrinkage suppression is the headline (§3.2): fibers neither crystallize nor contract like the matrix, and they mechanically restrain it, cutting in-print shrinkage by half to two-thirds at common loadings — the entire printability story of PP-GF (§11.4), PET-CF (§9.3), and PPA-CF (§14.4). Melt viscosity rises, so fiber grades print hotter and hit volumetric-flow ceilings sooner than their unfilled bases; there is also a floor — fiber-loaded melt back-pressure jams at very thin layer heights (3DXTech documents a 0.25 mm floor for PP+CF, §11.5). The stiffer bead bridges and tolerates fan-off printing better, which suits the layer-adhesion-sensitive engineering families. Surfaces come out matte, hiding layer lines — a real cosmetic advantage that also hides the stringing and micro-roughness moisture causes, so do not let the finish substitute for drying discipline.
| Axis | Chopped carbon fiber (CF) | Chopped glass fiber (GF) |
|---|---|---|
| Stiffness per loading | Highest; the specific-stiffness leader | Roughly half CF's gain at the same wt% |
| Density penalty | Small (CF ~1.8 g/cm3) | Larger (glass ~2.5–2.6 g/cm3) |
| Impact retention | Poor — brittle composite behavior | Substantially better than CF |
| Color | Black only, matte | Colorable; natural/white available |
| Nozzle wear | Hardened mandatory | Hardened mandatory |
| Electrical | Conductive fragments — a consideration around exposed electronics and for ESD behavior | Insulating |
| Cost | Comparable to GF at consumer tier | Comparable; historically cheaper |
Table 29.1 — CF vs GF selection axes, generalizing the §13.4 nylon discussion to every matrix in the volume. The selection logic is stable across families: stiffness-to-weight as the binding constraint favors CF (drone frames, end-effector tooling); impact retention, color, or cost favors GF (brackets, housings, snap-fit adjacent parts). Neither is safe on a brass nozzle (§4.1).
Every commercial reinforcement fiber is coated at manufacture with a sizing — a proprietary surface treatment (epoxy-, urethane-, or silane-class chemistries) that protects the fiber in handling and, critically, couples it to a specific matrix family so interfacial shear can transfer load. Sizing is invisible on a TDS and never disclosed, but it is a first-order variable: the same nominal fiber loading in the same nominal matrix performs differently across brands partly because the fiber–matrix interface differs (§13.6, §22.4). Recycled-CF feedstocks (Prusament PP-CF, Braskem FL900PP-CF, §11.5) add another layer of variance — reclaimed fiber arrives with varied length distributions and legacy sizing. None of this is readable from a datasheet; printed-coupon testing (Chapter 31) is the only proxy available to a practitioner.
Measured stiffness rises with fiber loading only up to a point: past roughly 20 wt%, printed-part stiffness tends to plateau while brittleness keeps rising (§14.11). Packing density, fiber–fiber interaction, and breakage during compounding all work against the marginal fiber. Meanwhile the trade is monotonic: elongation collapses to a few percent, unnotched impact drops sharply (notched values may hold — the fiber blunts crack initiation while embrittling everything else, §13.4), and cyclic flexural loading finds the fiber-aligned planes and fails along them (§14.4). The practical rules: a higher headline loading is not by itself a reason to prefer one product; never specify a fiber-filled grade for a flexure, hinge, or snap-fit; and treat the composite as a stiffness-and-stability material, not a toughness material.
Fiber-filled melt is an abrasive slurry. It erodes the nozzle bore — rounding the orifice until extrusion width drifts and calibration silently rots — and a brass nozzle wears measurably within a single large print (§9.4). Hardened steel is the practical minimum; ruby, tungsten carbide, and PCD tips extend life from hundreds toward thousands of hours under heavy loading (§4.1). Wear does not stop at the nozzle: extruder gears, PTFE path liners, and multi-material buffer internals all see it, and buffer systems not rated for abrasives should not feed CF/GF filament (§11.9). Handling adds one more failure mode: high-loading filaments are brittle on the spool, and bent filament paths can snap them inside PTFE tubes — §14.10 documents the PPA-CF case and its mitigations, which apply in degree to every high-loading composite.
Two exposure paths matter. Machining and sanding printed composites liberates respirable fiber fragments — wet sanding and low-speed technique are the controls (§26.1). Printing fiber grades shifts the emission profile relative to the base polymer (§5.1, Table 5.1); enclosure and filtration recommendations follow the matrix polymer's row. End-of-life is the third, slower problem: filled grades are essentially non-recyclable because the fiber prevents clean melt reprocessing (§5.4).
Part I established which polymers care about moisture and why (§3.4) and set the drying protocols (§3.5, Table 3.1). This chapter covers the other half of what is arguably the most-used workflow in engineering FDM: the hardware that executes those protocols — dryers, ovens, dry boxes, desiccants, and the verification that closes the loop. The polymer chapters lean on this constantly (§13.5's "skipping the drying row is where most first-time nylon prints fail"; §14.8's dryer-ceiling caveat; §20.1's print-from-dryer requirement); here it is stated once.
| Class | Realistic ceiling | Right for | Limits |
|---|---|---|---|
| Consumer filament dryers (single/dual-spool heated enclosures) | ~80–90 °C in practice (§14.8), many models lower | PLA-class through PA12/PA612 protocols; holding a spool dry while printing | Often run below setpoint; limited air exchange — moisture must leave the enclosure, and vented models dry faster than sealed ones; soft TPU deforms near the top of the range (§16.7) |
| Food dehydrators | ~70 °C | Budget multi-spool drying at the PETG/TPU/PA12 tier | Spool diameter fit; no nylon-6/PPA-class temperatures |
| Convection ovens | 150 °C+ | The 100–140 °C schedules: PA6 at the top of its range, PPA-CF, PPS-CF, PEI (§3.5, §18.1) | Setpoint overshoot and hot spots — verify with an independent thermometer before trusting a spool to it; spool substrate limits (§30.4) |
| Vacuum and industrial dryers | per spec | Production throughput; fastest dry times | Cost; industrial tier |
| Heated chamber, passively (§4.3) | chamber temp | Opportunistic top-up during long prints | Not a substitute for a real dry; other spools in the same enclosure may soften |
Table 30.1 — Drying hardware classes. The single most common mismatch: a consumer filament dryer nominally set to 90 °C executing a 100–140 °C PPA or PPS protocol — it cannot (§14.8), and the under-dried spool then fails in exactly the ways §14.8 describes. Match the hardware class to the protocol temperature in Table 3.1 before matching anything else.
Storage is cheaper than re-drying. The standard tiers: sealed dry boxes (gasketed containers with bulk desiccant) hold opened spools at low humidity between prints and, fitted with PTFE pass-throughs, feed the printer directly; vacuum bags with desiccant archive spools that will not be used for months; as-shipped packaging with the vendor desiccant is fine only until first opening. Practical humidity targets, from community and vendor practice: under ~20% RH for the nylon/PVA class, under ~30% for general engineering storage. Multi-spool buffer systems (AMS-style enclosures) occupy a middle ground — most include desiccant bays and stay usefully dry for the forgiving materials, but they are storage, not drying, and the moisture-critical polymers still want an active dry before a serious print (§3.5). Thermally, chamber-adjacent storage cuts both ways (§4.3): spools above 45 °C dry passively, but soft materials can deform.
Indicating silica gel is the workhorse: cheap in bulk, safe, regenerable in a ~100–120 °C oven for one to two hours (follow the packaging — microwave regeneration risks hot spots that melt beads), and the orange indicating variants show their state at a glance. Avoid legacy cobalt-chloride blue-indicating gel; the indicator is a regulated toxicity concern and the orange chemistry replaced it for good reason. Molecular sieve pulls humidity far lower than silica gel and holds it at elevated temperature — better equilibrium for bone-dry nylon storage — but regeneration needs 200 °C+, beyond practical kitchen-oven cycling, and it gives no visual state. Calcium-chloride tubs (DampRid-class) absorb into liquid brine: effective for cabinets, wrong for print-adjacent boxes — brine near spools and electronics is its own hazard. The practical default is bulk indicating silica gel on a regeneration rotation, with molecular sieve reserved for the driest long-term storage.
The ceiling on drying temperature is usually the spool, not the polymer (§3.5's PA6 note, §14.8). Ordinary plastic spools deform above roughly 80–90 °C; vendors specifying 100–140 °C protocols ship on heat-rated cardboard (Bambu's are rated to 145 °C) or expect the user to re-spool onto metal for oven cycles. Cardboard tolerates the heat but sheds edge debris that wears feeders and buffer systems (§14.8) — a real trade, not a free upgrade. For oven schedules above ~90 °C on a plastic spool, re-spooling is the honest options list.
Three verification layers, cheapest first. Hygrometers in the dry box: place the sensor at spool level rather than the lid, expect ±5% RH from the inexpensive units, and read the trend over hours rather than the instant value — a freshly closed box takes time to equilibrate. Weighing is the most direct evidence a home workflow can produce: weigh the spool before and after drying on a 0.1 g-resolution scale; the mass lost is water driven off, and a spool that stops losing mass between cycles is as dry as that protocol will get it. Cardboard spools confound both methods — the spool itself absorbs and releases water. The print is the final arbiter: the symptom cluster of §14.8 (fine stringing through tuned retraction, surface roughness, audible popping, micro-bubbled beads) means wet, whatever the hygrometer said.
For the polymers that reabsorb faster than a print finishes — PVA and BVOH by design (§20.1), PA6-class nylons and PPA in humid shops, hygroscopic elastomers — drying before the print is necessary but not sufficient: the spool must stay dry while it feeds. The standard arrangement runs PTFE from a heated dryer or desiccant dry box straight to the extruder. Two frictions to plan around: the added path length and bends raise drag, which flexible filaments tolerate poorly (§16.5), and abrasive filaments wear pass-through fittings and buffer internals not rated for them (§11.9). For multi-day engineering prints, a dryer running at holding temperature (50–70 °C) with the print feeding from it is the difference between the first layer's material state and the last layer's.
"Bracket-test on the actual machine" is this volume's most repeated advice — it closes §1.4's anisotropy warning, §8.11's Z-strength caveat, §13.7's datasheet critique, and the §23 calibration workflow. This chapter turns it into a procedure. The scope is practitioner-grade qualification: producing defensible numbers for your machine, spool, and geometry. It is not accredited-laboratory testing, and nothing here substitutes for the certification pathways regulated applications require (Appendix E).
A datasheet value is a ceiling measured on someone else's best-case specimen (§22.4). The design question is different: does this part, printed this way, survive this load in this environment, with margin. Between the two sit every variable the TDS holds constant and a print shop does not: orientation and anisotropy (§3.1), layer-line porosity, moisture state (§13.2), anneal state, per-spool drift (§23), and the geometry's own stress concentrations. Testing collapses all of them into one observable.
Printed test coupons — ISO 527-2 / ASTM D638-style tensile bars, ISO 178 flexural bars — are cheap and standard, but printed coupons are not molded coupons, and their absolute numbers inherit every process variable above. Use them where they are strong: relative comparisons under controlled change. Brand A vs brand B, dry vs conditioned, 240 °C vs 260 °C, annealed vs not, three walls vs five — printed on the same machine in the same session, coupons rank these reliably even when their absolute values would not survive an audit. When an absolute number is the requirement, test the part itself (§31.4).
Upright-printed tensile bars under-read — layer steps and grip-zone stress concentrations initiate failure early — so treat any Z coupon number as a lower bound, or use geometries designed for the question: waisted (dog-boned in the round) cylinders, or printed hooks loaded to failure. Report Z strength as a fraction of XY from the same print session and compare against the §3.1 bands (60–85% unfilled, 40–60% fiber-filled): a healthy process lands in-band, and a low outlier is a process problem — temperature, cooling, moisture — before it is a material property. Z response to nozzle temperature and fan is steep, which makes it the single most valuable axis to bracket during §23.2's temperature-tower step for load-bearing work.
The procedure the volume keeps pointing at, in full: print 3–5 replicates of the actual geometry at the candidate settings; mount each in a fixture that reproduces the service load path and direction; load progressively to failure — dead weights, a luggage/crane scale in a pull rig, or an inexpensive digital force gauge all produce a number; record failure load and failure mode (layer separation vs bulk fracture vs feature pull-out — the mode tells you what to fix); change one variable at a time and repeat. The pass criterion is stated up front: the worst specimen carries at least the safety factor times the service load. Scatter is the diagnostic: a spread wider than ~20–30% of the mean means the process, not the design, is the problem — go back to drying (§30), adhesion (§24), and calibration (§23) before trusting any average.
Test specimens in the state the part will serve in, because state moves the answer more than most design changes. Moisture: nylons tested dry-as-printed overstate wet service badly (Table 13.2's dry→wet gaps; Table 14.6's 61% PA6-CF stiffness loss) — condition coupons to service humidity, by weeks at ambient or an accelerated warm-water soak, and expect conditioned nylon parts to have grown slightly (§13.2). Anneal state: test what production will ship — annealed if annealed, as-printed if not (§26.4). Temperature: if the part serves warm, test warm (§31.6). A part qualified in the wrong state is qualified for a different application.
HDT is a minutes-long test; service is not (§1.4, §22.4). The practitioner's creep screen: fix a printed bar as a cantilever (or in three-point bending) under a service-representative stress, hold it in an oven at service temperature, and measure deflection at 1 hour, 24 hours, and a week. Amorphous polymers loaded near Tg fail this screen quickly and visibly — exactly the §1.4 warning made observable — while a part that holds deflection through a week at temperature has earned real confidence no datasheet line provides. For hot, load-bearing applications the week of oven time is the cheapest insurance in this chapter.
Print-to-print variance is real, so single specimens prove nothing. Working rules: n ≥ 5 for anything load-bearing; report the minimum and the spread, not just the mean; design against the lower bound. A coefficient of variation around 5–10% on XY tensile indicates a healthy process (Z runs higher); a CV well beyond that is a process alarm (§31.4). Safety factors are the designer's call, but common practice for printed parts runs ~2× on static ultimate for non-critical duty and substantially higher under fatigue, impact, or moisture cycling — treat these as conventions, not standards. Finally, a qualification is only as durable as its record: log spool brand and batch code, drying state, the full calibrated profile (the Appendix B pattern), calibration date, print orientation, and conditioning — and re-verify on each new spool per §23.8, because batch drift of 5–10% is normal.
Calibration (Chapter 23) makes the machine dimensionally honest; this chapter makes the geometry live with what the polymer does anyway. It is not a general design-for-additive tutorial — the preface excludes that — but the material-dependent design deltas the polymer chapters keep gesturing at, consolidated: fits by shrinkage class, fastening by family, compliant features by family, and load-path orientation.
| Shrinkage class | Polymers | Typical XY compensation (§23.6) | Design consequence |
|---|---|---|---|
| Low — amorphous | PETG, PCTG, ABS, ASA, PC blends; PLA (slow-crystallizing, §6.1) | 0.2–0.5% | Sliding fits of 0.15–0.3 mm per side are achievable on a calibrated machine; tight bores stay round |
| High — semi-crystalline, unfilled | PP, PE, PA6, PA66 | 1–2%+ | Design generous clearances (0.3–0.5 mm+); avoid long tight bores and large precise flats; expect §32.6 warp geometry rules to apply |
| Fiber-tamed semi-crystalline | PP-GF/CF, PA-CF/GF, PPA-CF, PET-CF, PPS-CF | 0.2–0.5% | Near-amorphous dimensional behavior (§3.2); the composite premium buys precision as much as stiffness |
Table 32.1 — Fit design by shrinkage class. Compensation values are the §23.6 calibration outputs; the classes, not the exact numbers, are the design input. Two universal FDM effects sit on top: holes print undersized (bead rounding plus shrinkage — compensate in the slicer or design holes 0.1–0.3 mm over, and prefer teardrop profiles for horizontal holes), and first layers flare outward (elephant foot — chamfer bottom edges 0.3–0.5 mm). One material-specific effect: nylon parts grow measurably as they condition toward service humidity (§13.2), so dimension-critical nylon assemblies should be measured at equilibrium, not off the plate (§31.5).
Perimeters carry load; infill mostly stabilizes them. Engineering parts default to 3–5 walls (§3.1) with infill chosen for the failure mode: solid or near-solid under inserts, bearing faces, and wear surfaces (§28.3); 20–25% cubic elsewhere, remembering the top-surface dishing failure on sparse infill that §8.8 and §11.6 document. Where Z-loading is unavoidable, more walls help more than more infill.
Heat-set inserts are the default engineering fastener: size the boss at roughly twice the insert OD, give it the insert's length plus 2 mm of depth, and back it with local solid infill and at least two perimeters. PLA, PETG, PCTG, ABS, ASA, PC, and the nylons all take inserts well (§26.5); PLA is the easiest install of all (§6.8). PP and PE take neither adhesives nor inserts well — the low-energy surface defeats glue and the creep-prone matrix relaxes around inserts, so design through-bolts and mechanical interlocks instead (§11.8). Printed threads work from roughly M8 and coarser; below that, tap or insert. Tapped threads hold best in the ductile and high-crystallinity materials — PCTG (Table 8.7) and POM (§17.2) — and worst in brittle PLA. Sustained bolt preload relaxes by creep in every printed polymer, worst in PLA (§6.8): re-torque schedules, shoulder screws, or through-bolted metal stacks are the fixes for joints that must stay tight.
Snap-fits, clips, and latches: PETG, PCTG, PP, and the nylons survive many cycles; PLA fractures on early cycles; printed PC fails at layer lines and sharp internal corners despite the resin's ductility — radius the flexure root and orient the flexure in-plane (§26.5). Living hinges: unfilled PP copolymer is the only fully credible printed living hinge (§11.10), with PA12/PA612 serviceable at lower cycle counts and TPU/PEBA covering the soft end (§16.11). The absolute rule from §29.5: no fiber-filled grade in any flexure, hinge, or snap-fit — the fibers create the fatigue planes the feature will fail along (§14.4).
Orient so principal stress runs in XY (§3.1); where a feature must load in Z, redesign rather than hope — dovetails, through-bolts, or splitting the part and bonding across the joint move the load off the layer interface. Wear and sliding faces belong on perimeter walls, not stepped top/bottom surfaces (§28.3). Overhang planning interacts with material: fiber-filled beads bridge better (§29.2), elastomers droop early, and the fan-restricted engineering polymers (§10.6, §15.9) cannot buy overhang quality with cooling the way PLA can — design the overhangs out instead.
For the high-shrink families (§3.2, Chapters 10–11): radius external corners generously — shrinkage stress concentrates at sharp corners and starts the lift there; break up large uninterrupted flat slabs with ribs, cutouts, or crowned surfaces rather than fighting them with adhesive alone; keep first-layer contact area high where grip is needed and design brim or mouse-ear allowances into the part's finish expectations; and favor several smaller bonded parts over one warp-prone monolith when the geometry allows. The bed-adhesion chapter (§24) supplies the process half of this contract; the geometry half is cheaper.
← Contents · ‹ Part XI — Support and niche polymers · Appendices ›
FDM Polymers — A Technical Reference
- Part I — Foundations
- Part II — PLA Family
- Part III — Polyester Family
- Part IV — Styrenics Family
- Part V — Polyolefins
- Part VI — Polyamides
- Part VII — Polycarbonates
- Part VIII — Thermoplastic elastomers
- Part IX — Specialty engineering thermoplastics
- Part X — High-temperature polymers
- Part XI — Support and niche polymers
- Part XII — Cross-cutting workflows
- Appendices
- Source manifest