There are at least three ways to prototype plastic parts. Which way is best? Each of our company's two divisions focuses on a separate method: Protomold provides rapid injection molding, while First Cut Prototype provides CNC prototyping. Also important are additive technologies that build-up parts layer-by-layer, such as stereolithography and fused deposition modeling (FDM).

Before discussing each method, it's interesting to note there is no clear line between prototype and production volumes. For example, about half our customers require what they call production volumes, under about 20,000 parts/year. But medical-device companies usually require production volumes of only 500 to 5,000 parts/year.

When it comes to millions of parts/year though, rapid injection molding is not cost-effective. In fact, carefully consider costs for anything over 10,000 parts. For example, our aluminum mold-tooling costs are low, but prices for parts at these volumes are typically higher than production with multi-cavity steel molds.

That said, some companies buy tens of thousands of rapid-injection molded parts because they can change designs and still get parts in days, not the months it typically takes with traditional molding. It relies on large, expensive molds made of hardened steel, while our method, for example, uses aluminum molds that are fast to make and easier to manipulate.

CNC prototyping makes sense when about one to ten parts are needed and the material required won't work with additive processes. For example, it would make sense to machine, say, three nylon enclosures for testing in operating conditions.

It's easy to claim one method produces stronger parts than another so we recently used Stork Materials Technology, a group of independent, accredited laboratories, to test the relative strengths of cut, molded, and built-up prototypes. This information is important because prototypes used for functional testing should meet the performance requirements of production parts.

All parts the lab tested were made from the same CAD model using ABS for molded and cut parts, and ABS-like plastic for built-up ones. FDM, an additive method, was used because it makes the strongest parts of any major RP technology. Molded samples had three different gate configurations because gating can affect part strength.

Standard ASTM D790 testing measured flexural properties and ASTM D638 testing measured tensile properties. The average strain at break and average break stress were much higher for machined and molded samples compared to FDM parts. Also, molded parts with one end gate showed higher average strain at break compared to samples with two end gates or a center gate. Molded parts with a center gate showed the highest average flexural strength. FDM parts rated lowest in these tests.

In terms of finish, prototypes for market testing should approximate production part finishes. Rapid injection molded parts have smooth finishes nearly identical to those found on production parts. CNC machined prototypes have a finish similar to molded parts, and FDM prototypes have a rough surface, reflecting the layering process that made them.

When it comes to delivery, additive methods have the advantage of speed because they make parts directly from 3D CAD models. Users can purchase industrial FDM printing equipment for around $20,000 that produces prototypes in just hours.

And lastly, when it comes to costs to make a single part, 3D printing is somewhat less expensive than CNC prototyping, and rapid injection molding is the most costly because of tooling costs. Of course, exact figures depend on materials being used and part complexity.

Because they have low fixed costs, 3D printing and CNC-prototyping costs start low and climb steadily with quantity. The last copy costs the same as the first, with no economies of scale. Typically, the cost of a machined part is slightly higher than that of a similar additive part. Lastly, the piece-part cost of rapid injection molding is high for the first 25 parts, but becomes much more cost-effective than the other methods at higher volumes.

Typical development scenarios include using RP in the earliest stages when the likelihood of change is great. And it might make sense to switch to CNC in early phases of functional testing when material properties are critical. Rapid injection molding is the most cost-effective when more parts are needed for extensive lab or market testing. Rapid injection is also affordable for moderate-volume production and for bridge tooling while steel molds are built.

ABILITY TO RESIST DEFORMATION

Sample	Average Flexural Strength, ksi
Machined	9.3
Molded, 2 end gates	9.3
Molded, center gate	9.6
Molded, 1 end gate	9.2
FDM, X axis	4.9
FDM, Z axis	3.6

IMPORTANCE OF GATE PLACEMENT

Sample	Average Strain Average Break at Break, %	Stress, psi
Machined	16.2	4562
Molded, 2 end gates	2.0	5117
Molded, center gates	2.2	5118
Molded, 1 end gate	14.5	5399
FDM, X axis	1.6	2563
FDM, X-Y axis	1.5	2529
FDM, Z axis	1.5	2008
The colored area demonstrates the importance of gate placement.

The tables compare the average flexural strengths (top) and the average strain at break and average break stress for cut, molded, and built parts.

HOW PLASTIC PROTOTYPES STACK UP

METHOD	PROS	DESCRIPTION	CONS	DESCRIPTION
Molding with rapid techniques	Material choices Production support Surface finish Dimensional accuracy	Most thermoplastics Up to 10s of thousands of parts Production quality Production quality	Tooling costs Geometry constraints	Protomold's minimum tooling cost is $1795 Internal undercuts are not supported
	Example ideal application: Medical device enclosures
METHOD	PROS	DESCRIPTION	CONS	DESCRIPTION
Cutting or milling	Material choices Tooling costs Surface finish Dimensional accuracy	Most thermoplastics No tooling required Mimics injection molded surface Comparable to injection molded accuracies	Production support Complexity constraints	Parts are machined one at a time Parts are currently machined with 3-axis technology.
		Example ideal application: Prototype medical instruments
METHOD	PROS	DESCRIPTION	CONS	DESCRIPTION
Fused Deposition Modeling	Tooling costs Geometric constraints	No tooling required Builds almost anything that can be imagined	Material limitations Production support Surface finish	Not many sole-source materials are available Parts are built one at a time Parts have a “stair step” surface finish
	Example ideal application: Replica of internal human skeletal structure