Latest / Hardware Startups with Fexingo: Devices, Robotics, and Manufacturing Tech Companies / Why Hardware Startups Should Use Stereolithography for Precision Prototypes
Transcript
- Lucas: If you're a hardware startup trying to get a functional prototype into your hands fast — not just a look-alike, but something that actually fits, seals, or flows fluid — then stereolithography is probably your best friend right now. Luna: SLA. The OG 3D printing technology. I feel like everyone's heard of it, but most founders I talk to still default to FDM or SLS without really knowing what SLA can do. Lucas: Right, and that's the gap we want to close today. Because SLA gives you sub-50-micron accuracy — that's about half the thickness of a human hair. For things like fluidic manifolds, optical housings, or snap-fit enclosures, that kind of precision can save you weeks of rework. Lucas: And look, if these kinds of practical engineering conversations are useful for what you're building, I'll mention — there's a simple way to keep the show ad-free and independent. A couple of dollars a month genuinely makes a difference. It's buy me a coffee dot com slash fexingo. Luna: Yeah, and we mean that. It's how we keep digging into topics like this without any sponsor noise. Lucas: So back to SLA — let's talk about a specific case. I was talking to a medtech startup last quarter. They were prototyping a fluidic manifold for a diagnostic device. The manifold had channels that needed to be exactly 0.1 millimeters wide, with no burrs, no warpage, and they had to be transparent so the optics could read the flow. Luna: That's a tough spec. FDM would give you layer lines and probably leak. SLS would leave a powdery surface that's not transparent. Lucas: Exactly. They tried CNC machining first — $8,000 for a single part, two-week lead time, and the tooling broke on the second run. Then they switched to SLA on a Formlabs Form 3. Printed the manifold in eight hours, total material cost about fifty bucks, and the channels came out at 0.09 millimeters — within tolerance. Luna: And that's the killer app for SLA: when you need dimensional accuracy and a smooth surface finish, but you don't need production-grade mechanical strength. Because SLA resins are brittle compared to injection-molded plastics. Lucas: Brittle is the big trade-off. Standard SLA resins have an elongation at break of maybe 5 to 10 percent. So if your prototype needs to survive a drop test or flex repeatedly, SLA is not your answer. But for fit-checks, fluidics, and optical paths — it's perfect. Luna: And you've got options now. There are engineering resins — like from Carbon or Loctite — that are tougher. Carbon's RPU 70 has elongation at break around 70 percent. But those require a different kind of SLA, usually with a heated build chamber and more expensive materials. Lucas: Right, and that brings up the distinction between desktop SLA and industrial SLA. Desktop machines — Formlabs, Peopoly, Elegoo — use a laser that cures resin point by point. They're great for one-off prototypes. Industrial systems like Carbon's Digital Light Synthesis use a digital light projector to cure whole layers at once, which is faster and can use tougher materials, but you're looking at $50,000-plus for the machine. Luna: So for a startup on a bootstrap budget, you start with a desktop SLA. But you have to think about post-processing. SLA parts come out with a sticky resin coating that you have to wash in isopropyl alcohol and then UV cure. That's an extra step people forget. Lucas: It's not a one-button process. And the wash and cure stations add another couple hundred bucks. But compared to the weeks of CNC lead time or the surface roughness of SLS, it's a no-brainer for certain geometries. Lucas: Let's talk about geometry. SLA can do internal channels and overhangs that would be impossible to CNC. That medtech manifold had a curved internal channel that changed diameter along its length. No way to cut that with a drill bit. Luna: And the surface finish is good enough that you can actually test fluid flow without having to sand or polish. That's huge for development speed. Lucas: Another example: I know a robotics startup that used SLA to print the housing for a camera module. They needed the lens mount to be perfectly perpendicular to the image sensor within 0.05 degrees. SLA held that tolerance. They iterated through five design revisions in one week, each costing less than a hundred dollars in resin. Luna: Compare that to aluminum CNC: five revisions would have been five grand and a month of waiting. So SLA compresses the iteration loop dramatically. Lucas: It does. And that's the real value for a startup — not just the cost per part, but the speed of learning. You can go from CAD to physical part in a day. Test it, find the flaw, fix the model, print again tomorrow. Luna: But you have to design for SLA. You can't just take an injection-molded part geometry and print it. You need to think about drain holes for uncured resin, support structures, and orientation on the build platform. Lucas: Absolutely. Orientation matters because layer adhesion is weaker in the Z direction. If you're printing a snap-fit arm, you want the arm oriented so the bending stress isn't parallel to the layers. Otherwise it'll snap right along a layer line. Luna: And supports leave marks. You'll have little nubs where the supports touched the part. Those need to be sanded off, and if it's a sealing surface, that can be a problem. Lucas: That's why some startups use SLA for master patterns — they print the SLA part, sand it smooth, and then use it to make a silicone mold for urethane casting. That gives you a production-like part in a tough material without the tooling cost. Luna: That's a smart bridge strategy. You get the speed of SLA with the material properties of cast urethane. And the mold costs maybe a few hundred bucks. Lucas: Let's zoom out. SLA is not for production. The resins degrade with UV exposure over time. They creep under constant load. But for prototypes — especially functional prototypes that need precision — it's unmatched. Luna: And the industry is moving fast. There are now biocompatible SLA resins for medical devices, high-temp resins that can withstand 200 degrees Celsius, and even flexible resins that behave like rubber. Lucas: So the question for a hardware founder is: when should you reach for SLA instead of FDM or SLS? My rule of thumb — if the part has features smaller than one millimeter, or needs a smooth surface for sealing or optics, or you need to iterate quickly on a complex internal geometry — go SLA first. Luna: And if you need strength, go SLS with nylon. Or if you just need a quick visual model, FDM is fine. But SLA hits that sweet spot of precision and speed for functional prototypes. Lucas: One last thing — don't sleep on the cleaning and curing. A lot of startups buy an SLA printer and then wonder why their parts are tacky or brittle. You need to wash thoroughly and cure at the right temperature and time. Every resin has a datasheet. Follow it. Luna: And handle the liquid resin with gloves. It's a skin irritant. Basic safety goes a long way. Lucas: Good point. So to recap: SLA gives you sub-50-micron accuracy, smooth surfaces, and fast iteration. It's ideal for precision prototypes like fluidic manifolds, optical housings, and snap-fit parts. The trade-offs are brittleness, post-processing, and limited material strength. But as a tool in the hardware startup toolbox, it's indispensable. Luna: And with printers under $3,000 now, there's really no barrier to getting started. I'd say every hardware founder should have access to one. Lucas: Absolutely. Next time you're facing a two-week CNC lead time, ask yourself — can I SLA this instead? You might save a month.