A 3D printer turns a CAD file into a physical part for somewhere between $5 and $200 in 4 to 48 hours. That changed the economics of physical prototyping more than any other shift in the last 30 years. It also leads inventors to print things that should not be 3D-printed at all, and, more often, to reach for a printer before they have asked whether a physical part is needed at all.
For a licensing-track inventor, it usually is not, at least not first. Companies evaluate inventions from photorealistic renderings, CAD, and product animation. A virtual prototype carries the design while it is still in flux, communicates the product to a licensee, and revises in an afternoon. A 3D print is a physical step that comes after the design is settled, and only when a specific reason calls for a physical part. The three types of invention prototypes split into proof-of-concept, looks-like, and works-like stages, and 3D printing serves the last two. Note that everything a 3D printer makes starts as a CAD file, the same CAD file the virtual prototype is built from.
This article explains where 3D printing fits, which process to use, and where it earns its place, so an inventor knows what the work involves and what they are paying for. Enhance Innovations has run invention design from its office in Champlin, Minnesota since 2010, and works virtual-first: the core deliverable is a virtual prototype, with physical builds scoped per project when one is needed. The decision framework below assumes that order.
When 3D printing is the right call
Once a project has a reason for a physical part, 3D printing earns its place when one or more of the following is true.
The geometry is complex. Internal channels, organic curves, undercuts, and integrated features that would require multi-axis machining or multiple manufacturing operations get printed in a single build. This is where 3D printing has no peer.
The quantity is low. One to twenty parts. Above that quantity, traditional manufacturing methods (machining, casting, molding) start to win on per-part cost.
The part needs to be handled. Once the design is settled in CAD, a printed part lets a manufacturer or retail buyer hold the form, feel the weight, and judge the ergonomics. This is the looks-like model use case, and it applies when a buyer specifically asks to handle a sample.
The part needs to fit physical hardware. Holding a printed enclosure against a real component to verify clearance catches issues the CAD model already flagged but a person wants to confirm by hand. This belongs to the works-like stage, and matters most when the inventor is engineering toward self-manufacturing. The engineering and prototyping work that produces a functional unit scopes a physical build only when the project reaches that point.
You need a master pattern for casting. SLA prints make excellent masters for silicone mold-making. The print itself is the precursor, not the production part.
What 3D printing does not do is replace the virtual prototype. Geometry exploration and the licensing pitch happen on a screen, where a change costs an afternoon. The printer enters once that work is done.
When 3D printing is the wrong call
3D printing is the wrong choice in a handful of situations that come up enough to be worth naming.
The part needs to bear sustained mechanical load. FDM and SLA prints are weaker than equivalent machined or molded parts. Anisotropy (different strength in different directions) is a particular concern with FDM. SLS and engineering-grade SLA close some of this gap but do not eliminate it.
The part is mostly flat. A flat panel, plate, bracket, or sheet-form part should be laser-cut from acrylic, polycarbonate, or sheet metal. The build time on a 3D printer for a flat part is wasted compared to the few seconds a laser cutter takes to do the same work.
The geometry is simple but the volume is high. If you need 50 identical parts and the geometry is a square block with four holes, machining 50 blocks from a bar of stock is cheaper and faster than 50 prints.
The surface finish has to look injection-molded. Even the best SLA prints carry a subtle texture and parting line that experienced eyes catch. For a presentation-grade looks-like model that needs to sit next to actual injection-molded parts, the print is a starting point but not the deliverable. Expect 4 to 12 hours of hand finishing per part on top of the print time.
The part has to handle heat above 60 to 70 degrees Celsius. Standard PLA softens around 60 degrees. ABS and PETG go higher. Engineering resins and SLS nylon go higher still. But for parts that live near a heat source, machined polycarbonate or PEEK outperforms any 3D-printed equivalent. The trade-offs here are the reason choosing materials for an invention prototype is a decision in its own right.
The three main 3D printing technologies for prototypes
Three technologies cover 95% of prototype work. Each has a sweet spot.
FDM (fused deposition modeling)
FDM is the desktop printer most inventors picture. A heated nozzle extrudes molten plastic in layers. Filament feeds in from a spool.
Strengths: cheap, fast for medium-sized parts, wide material selection (PLA, ABS, PETG, TPU, nylon, polycarbonate, glass-filled and carbon-filled variants).
Weaknesses: visible layer lines, anisotropic strength (weaker between layers), limited fine-detail capability, support material on overhangs leaves witness marks.
Best for: structural test parts, large enclosures, jigs and fixtures, internal components that nobody will see, fast iteration cycles where geometry matters more than finish.
Typical cost per part: $5 to $80 from a service bureau. Self-printed on an in-house FDM machine, materials cost is $1 to $15 plus labor.
Common machines: Bambu Lab X1, Prusa MK4, Ultimaker S5, Markforged Mark Two (for carbon-fiber reinforced parts).
SLA (stereolithography)
SLA cures liquid resin layer by layer using a UV light source. The resulting parts have fine surface detail and isotropic strength.
Strengths: fine feature resolution (down to 0.05mm), smooth surfaces, accurate dimensions, wide range of resins from rigid to rubbery to clear.
Weaknesses: parts get brittle over time when exposed to UV, post-processing is messy (alcohol wash plus UV cure), build volume usually smaller than FDM, material cost higher than filament.
Best for: cosmetic looks-like models, fine-detail parts, master patterns for silicone molds, jewelry-scale features, transparent parts, ergonomic models.
Typical cost per part: $40 to $250 from a service bureau. Self-printed on a desktop SLA machine, resin cost is $4 to $25 per part plus labor.
Common machines: Formlabs Form 4, Anycubic Photon, Phrozen Sonic Mega, Carbon (production-scale).
SLS (selective laser sintering)
SLS uses a laser to fuse powdered nylon (or other thermoplastic powders) into solid parts. No support material is required because the unfused powder bed supports the build.
Strengths: production-grade nylon strength, isotropic mechanical properties, complex geometries with no support marks, good chemical and thermal resistance.
Weaknesses: rough matte surface finish, limited material selection (mostly PA12 nylon and variants), higher per-part cost than FDM or SLA at small volume, difficult to do in-house (industrial machines only).
Best for: works-like functional parts, end-use bridge production, parts with living hinges or snap-fits, parts that need to flex or take load.
Typical cost per part: $50 to $500 from a service bureau. SLS is service-bureau-only for most design firms.
Common machines: EOS, HP Multi Jet Fusion (technically a different process but lands in similar use cases), Formlabs Fuse.
How to pick between FDM, SLA, and SLS
A useful rule. If you do not know which to use, default to FDM for early stages because the iteration cost is lowest, then switch to SLA or SLS when geometry stabilizes and the question shifts to finish or function. This mirrors how prototype iterations move from rough and cheap toward refined and final.
Service bureaus: where 3D printing work gets sent
Design firms typically split 3D printing between fast in-house prints for quick physical checks and service-bureau work for anything that needs to be presentation-grade, larger than 12 inches, or in a material outside an in-house machine's capability.
The main prototype service bureaus, with notes on what each one is good for:
Xometry. Online quoting, fast turnaround, very wide material selection across FDM, SLA, SLS, MJF, and traditional machining and molding. Good for one-stop coverage when a project needs multiple processes.
Protolabs. Higher quality control, faster lead times on rush work, slightly more expensive. Strong on 3D printing through machined parts and into low-volume injection molding. Their cost calculator inside the design feedback workflow catches cost-driving features early.
Shapeways. Originally a hobbyist platform, now serves the design and jewelry community. Strong material variety on the artistic side (full-color sandstone, precious metal casting from prints).
JLCPCB and PCBWay. Primarily PCB shops but both offer 3D printing alongside. Lowest cost on simple parts. Slower lead times due to international shipping.
Sculpteo. European-headquartered service with a strong materials portfolio and good US fulfillment.
Local print shops and university maker spaces. Many cities have a local print shop that runs FDM and SLA on a queue. Cost is comparable to service bureaus but delivery is same-day or next-day. Twin Cities has several. University maker spaces (St. Thomas, U of M) have walk-up FDM access at low or no cost for community members.
Public libraries. A growing number of public libraries (including Hennepin County in the Champlin area) offer free or near-free FDM printing for cardholders. Great for proof-of-concept rough prints. Quality and material selection are limited but the price is right.
Self-print or service bureau?
Inventors sometimes ask whether they should buy their own printer. For most independent inventors pursuing a license, the honest answer is that they will not print enough to justify a machine, and that geometry exploration belongs in CAD rather than in a stack of physical test prints. The question matters more for design firms and for inventors who have moved into self-manufacturing.
A desktop FDM printer in the $400 to $1,200 range covers fast, rough physical checks. The Bambu Lab X1 line, the Prusa MK4, and the Creality K1 are common at this price point. A desktop SLA printer in the $400 to $1,500 range adds cosmetic-surface capability; the Anycubic Photon Mono and the Formlabs Form 4 cover most needs, though resin handling needs a separate ventilated space.
For functional SLS work, buying a machine rarely makes sense. SLS printers start around $20,000 for entry-level (Formlabs Fuse) and run past $200,000 for industrial machines. Service bureau pricing wins unless someone is running 100-plus SLS parts per month.
The rule of thumb: occasional physical needs go to a service bureau, steady in-house volume that fits one machine justifies owning, and a high-volume multi-process workload justifies both. For a licensing-track inventor the realistic answer is usually none of the above, because the design work that matters is virtual.
File preparation: what to send the printer
Three file formats cover almost all printing.
STL. The mesh format that has been the printing default since the 1980s. Every printer accepts it. Loses metadata (units, tolerances, surface labels). Best for FDM and SLA where the print is the final form.
STEP. Solid model format that preserves geometry, tolerances, and assembly relationships. Service bureaus increasingly accept STEP and prefer it for complex parts. Use STEP when the part will be re-machined or re-tooled later.
3MF. The newer multi-component format from Microsoft and the 3D printing industry. Carries color, material, and orientation data with the geometry. Bambu Lab printers prefer 3MF. Adoption is still mixed across service bureaus.
CAD packages that export cleanly to all three: SolidWorks, Fusion 360, PTC Creo, Onshape, Rhino. Free options that get the job done: Tinkercad (browser-based, simple), FreeCAD (open source, full-featured), Onshape (free for hobbyists).
The most common mistake on file prep: not setting the units. A part designed in millimeters and exported as inches arrives at the printer 25.4 times too small. Always confirm units before submitting.
What 3D printing costs when a project needs it
A worked example shows where printing sits, assuming the virtual prototype is already built and the project has a reason for physical parts. A handheld product with a custom enclosure, 5 inches by 3 inches by 1.5 inches.
Looks-like model. SLA print at a service bureau in a tough resin, taken from the CAD file the virtual prototype already produced. 36-hour print and post-process. Roughly $185 per part for a single unit, plus several hours of hand finishing to reach a presentation surface. This step applies when a buyer asks to handle a sample.
Works-like unit. SLS print in PA12 nylon with functional internal geometry, snap-fit features, and mounting bosses for real hardware. Roughly $340 per part for a single unit, plus assembly and instrumentation. This step applies when the inventor is engineering toward self-manufacturing.
Across both physical builds, the print and finish cost lands in the high hundreds of dollars. Equivalent CNC-machined parts at the same stages run higher for this organic geometry, which is why 3D printing wins here; for a flat-paneled product the math reverses toward laser cutting and machining. Either way, both physical builds come after the virtual prototype, and a licensing-track inventor often needs neither. A fuller breakdown of what it costs to prototype an invention puts these print figures in the context of the whole package.
Limitations to know about
Three limitations come up over and over and trip up first-time prototyping inventors.
Dimensional accuracy. FDM holds about +/- 0.2mm on most features. SLA holds about +/- 0.1mm. SLS holds about +/- 0.3mm due to thermal shrink. If your design needs +/- 0.05mm tolerance, plan to machine, not print.
Material strength. A 3D-printed part is generally weaker than the same geometry machined from solid stock or molded from production resin. The closest 3D-printed materials to molded performance are SLS PA12, MJF nylon, and carbon-fiber-reinforced FDM. None match production injection-molded glass-filled engineering resins for ultimate strength.
Surface finish for cosmetic models. A raw print rarely passes as a finished product. Plan for sanding, priming, painting, and clear-coating. The labor on finish often exceeds the print cost.
FAQ
Can I print a working prototype that I can sell?
For some products, yes. Functional SLS nylon parts ship on production printers (HP Multi Jet Fusion bridge production runs into the thousands of units). For most consumer products with the volume to justify tooling, though, 3D-printed parts are prototypes, not production parts.
How long does a typical print take?
A small FDM part runs 2 to 6 hours. A medium FDM part runs 8 to 24 hours. SLA prints run 4 to 16 hours plus 30 to 90 minutes of post-processing. SLS prints run 16 to 48 hours including powder cooling.
Do I need a CAD model to print?
Yes. A 3D printer needs a 3D model, and the same CAD model also drives the renderings and animation that make up a virtual prototype. An inventor does not need to build the CAD themselves. A firm with industrial designers and engineers on staff can take a sketch to a CAD file in 1 to 3 weeks depending on complexity, and that file then serves the virtual prototype and any later printing. The fuller picture of how to make an invention prototype walks through that sequence from sketch to deliverable.
Will my idea be safe at a service bureau?
Service bureaus operate under standard NDAs and most will sign one on request. Their business model depends on confidentiality. The risk is low but non-zero. A firm that works with a known supplier on an ongoing basis, and that signs an NDA before the first technical conversation, adds an extra layer of trust.
What is the largest part I can 3D print?
FDM machines like the Modix BIG-Meter print parts up to a meter in any dimension. SLA caps around 30 centimeters in most desktop machines. SLS varies by service bureau. For larger geometry, parts get printed in sections and assembled.
What material best mimics an injection-molded ABS part?
For mechanical behavior, machined ABS rod or plate is closer than any 3D-printed equivalent. For visual approximation, SLA in a tough resin (Formlabs Tough 1500 or Tough 2000) with a primer-and-paint finish reads as injection-molded ABS at arm's length.
3D printing has moved from exotic to routine inside the prototype workflow, but it is a tool for the physical stages, not the starting point. For an inventor pursuing a license, the design work that matters happens in a virtual prototype: photorealistic renderings, CAD, and product animation that communicate the invention to a licensee and revise in an afternoon. A 3D print enters later, when a buyer asks to handle a sample or the inventor moves toward self-manufacturing. Enhance Innovations runs this work from one office in Champlin, Minnesota, with industrial design, engineering, marketing, and licensing under one roof, and starts most projects with a $399 patent search to confirm the idea is clear to pursue. A search reads the existing record at the USPTO patent database for anything that would block the claims. If you have an invention idea, that search is the first paid step, and the virtual prototype is what carries it forward.