Each of these five rapid prototyping mistakes wastes a build, and each costs far less to avoid before printing than to discover afterward. They are the wrong process for the question, a print loaded in its weakest direction, zero clearance, post-processing left out of the schedule, and threads tapped into printed plastic. All five sit downstream of the question of what the build must prove, so if that is still undecided, begin with getting started on prototyping.
Rapid Prototyping Mistakes Start With the Process Choice
Every process answers some questions well and others poorly. SLA (stereolithography) gives the smoothest surface of the three common plastic processes. SLS (selective laser sintering) comes out slightly grainy and FDM (fused deposition modeling) shows layer lines, so a part built to judge surface finish should not be an FDM print.
Cost runs the other way. FDM is the most cost-efficient choice for simple parts in small numbers. SLA costs more because of its resin, and SLS has the highest setup cost but needs no supports and nests parts in the build, which makes it the cheapest route for complex geometry and larger batches. For a one-off, printing usually costs less than machining, because the setup time and programming fall on one piece, though a large, solid, simple part can reverse that.
We would pick the cheapest process whose output can still answer the question, and move to real stock material only when the question demands it, as machining in rapid prototyping explains. We print in all three plastic processes through our 3D printing service, and the full comparison of methods and tolerances is in how rapid prototyping works.
Which Way Does the Load Run Through the Print?
FDM parts are weakest across their layers. Stratasys data for ABS-M30 from its industrial machines puts tensile strength across the layers at 71 to 89 percent of the strength along them. Elongation at break falls to about a quarter and notched impact to about a third, and the gap differs between its F900 and F770 machines.
Orientation fixes much of it: a hook printed lying flat keeps the main load from pulling layers apart. Wall count matters more than infill, since two to four walls usually adds more strength than raising infill from 50 to 80 percent. In bending, the maximum stress sits at the outer surface, which is where the walls are. Formlabs puts the loss across layers at 30 to 50 percent in many parts, so designing for about half the strength sits at the cautious end.
SLS and SLA behave differently. EOS PA 2200 nylon 12 shows tensile strength of 48 MPa in the build plane against 47 MPa vertically, though elongation at break halves from 20 to 10 percent, while SLA parts have no X, Y, or Z difference in their chemical bonds, though some resins are brittle.
Geometry matters as much as direction. Sharp inside corners concentrate stress and are a leading cause of plastic part failure, and DuPont’s rule puts the fillet at about half the wall thickness where ribs meet walls.
We would settle orientation before printing, then run the first load test with the load running along the layers. A part that fails there has a design problem, and a part that fails only across the layers has a print-orientation problem.
Why Won’t Parts Drawn at Zero Clearance Fit?
Zero clearance on a drawing will not survive printing, because no printer lands exactly on the drawn size. Guidance from Prusa starts at 0.3 mm for movable parts, and UltiMaker allows 0.6 mm between components that print together.
Printer accuracy explains why. Formlabs’ measured accuracy for the Form 4 with Grey Resin V5 is ±0.15% on features up to 30 mm, ±0.2% from 31 to 80 mm, and ±0.3% from 81 to 150 mm, with lower limits of 0.02, 0.06, and 0.15 mm. On a 100 mm feature that is 0.3 mm before any other error is counted, so a clearance that works at 30 mm can fail at 100 mm.
For holes, UltiMaker’s advice is to design them smaller than needed and drill them to size afterward. We would draw the clearance as a named parameter in the model rather than a fixed gap, so the next round is one number to change.
Where Does Post-Processing Fit in the Schedule?
A rapid prototype is not finished when the printer stops. SLA parts leave the printer in a semi-reacted green state, so excess resin is removed with an alcohol or ether wash and the part is post-cured to reach its intended properties. Supported SLA parts also need light sanding to remove support marks, and some SLA materials are sensitive to long UV exposure, so a cured part left in direct sun is not a stable test specimen.
SLS builds have to cool before parts come out, and Formlabs’ Fuse 1+ timing examples show about 6 to 14 hours of cooling after printing. We would quote the print time and the finishing time as separate lines, because a test on a part that skipped its cure measures a part that never reached its properties.
Should You Tap Threads Into Printed Plastic?
Printed plastic is a poor place to cut a thread. Printing guides recommend inserts over tapped threads for a stronger thread, with brass heat-set inserts for SLS and FDM parts and screw-in or epoxied inserts for SLA, because SLA resins are heat-sensitive. Metal inserts also give threads that survive frequent disassembly, which matters because a worn-out thread tells you nothing about the design. We would design the insert pocket into the first print instead of tapping later, since a prototype that gets assembled and taken apart repeatedly will test the thread long before it tests the design.
If a build keeps failing the same way, check it against these five rapid prototyping mistakes before changing the design. Does the process suit the question, which way does the load run, what clearance did you draw, when does the finished part actually arrive, and are the threads inserts? Use the contact page to tell us what the part has to do and which process you were leaning toward, and we will reply with the process and orientation we would choose. The phone line, 718-557-9578, works too.
