Additive Manufacturing (3D Printing) for Production
3D printing stopped being a prototyping toy years ago. Here is how to tell when it is the right way to make a real production part - and when it absolutely is not.
The word "production" changes everything
Every factory has a 3D printer in a corner. Most of them print prototypes, the occasional bracket, and a steady stream of desk toys. That is not what this is about. The moment you say a printed part will go into a product - sold to a customer, bolted into a machine, implanted in a patient, flown on an aircraft - the conversation changes completely. Production is not about whether you can print the part once. It is about whether you can print it the thousandth time, prove it is right, and do it at a cost that beats the alternative.
The technology cleared that bar a while ago. Aerospace fuel nozzles, dental aligners by the million, titanium spinal implants, and end-use spare parts for trains and tanks are all made additively today. The interesting question is no longer "can AM make production parts" but "which parts, and how do I know."
Complexity is free, but everything else costs money
The one genuinely magical property of additive manufacturing is that geometric complexity is free. A solid cube and an intricate, lattice-filled, internally-channelled version of the same cube cost roughly the same to print, because the machine only cares about volume, not cleverness. This is why AM produces parts that no mill or mold could ever make: conformal cooling channels snaking through the inside of an injection mold, organic load-following brackets spat out by topology-optimization software, twenty welded pieces consolidated into one monolithic part.
Everything else, though, costs real money, and beginners consistently underestimate it. Metal powder is expensive and partly wasted. Support structures cost material, machine time, and labour to remove, and they scar the surface. Post-processing - stress relief, support removal, hot isostatic pressing to close porosity, heat treatment, and machining of every mating face - is frequently the largest single line item on a metal part, dwarfing the print itself. A CT scanner to verify the inside of a critical part costs more than the printer. The print button is the cheap part of additive manufacturing.
Knowing where the curves cross
The cleanest way to think about AM economics is as a different cost curve. Injection molding has an enormous fixed cost - the mold - and a tiny cost per part. Additive manufacturing has almost no fixed cost and a high, flat cost per part. Plot them and they cross somewhere. Below the crossover, AM wins because there is no tooling to pay off. Above it, molding wins because the marginal cost dominates. CNC machining lives in between: no tooling, decent marginal cost, superb tolerances, and it beats AM on any simple prismatic part.
So the production engineer's real job is locating the crossover for this part. Low volume or one-per-customer customization pushes toward AM. Crushing geometric complexity pushes toward AM. A consolidation opportunity - turning an assembly into a single part and deleting the assembly labour, the fasteners, and the inventory line - pushes hard toward AM. Simple, high-volume, tight-tolerance parts push the other way, and no amount of enthusiasm should override that. The discipline to say "this should be molded" is what makes the rest of your AM recommendations credible.
Design is not optional
You cannot take a part drawn for a milling machine, press print, and get a good result. Additive parts are anisotropic - they are weakest between the layers - so orientation on the build plate quietly decides whether your bracket survives its load or peels apart like a deck of cards. Overhangs steeper than about forty-five degrees need supports, so good designers chamfer, teardrop, and reorient until the supports largely disappear. Critical holes come out undersized or oval, so you model them small and ream them to size afterward. Mating faces need machining stock because additive is a near-net-shape process, not a finishing one. This body of rules has a name - Design for Additive Manufacturing - and ignoring it is the single most common reason production AM programs produce expensive scrap.
The part that nobody talks about: proof
The hardest part of production AM is not the printing or even the designing. It is proving the part is right, every single time, to someone who will not take your word for it. Regulated industries demand a frozen process: once a part is qualified, you may not change the machine, the powder supplier, the parameters, or the post-processing without starting qualification over. The very flexibility that makes AM exciting - tweak a setting, try another orientation - is exactly what certification forbids. Repeatability comes from controlling powder lots, machine calibration, humidity, and operator decisions, and from documenting all of it into a traceable build record. Witness coupons printed alongside the real part get pull-tested to confirm the build behaved. A CT scan hunts for the porosity you cannot see.
This is unglamorous work, and it is where most programs stall. But it is also the whole game. Anyone can print one good part. A manufacturer prints a thousand good parts and can prove each one. Treat additive manufacturing as a precision tool in the production toolbox - reach for it when complexity, customization, consolidation, or speed create genuine value, qualify it without shortcuts, and leave it on the shelf the rest of the time. That selective discipline, far more than the printer itself, is what turns 3D printing into production.
No comments:
Post a Comment