Latest / Hardware Startups with Fexingo: Devices, Robotics, and Manufacturing Tech Companies / How Hardware Startups Use Micro Injection Molding for Tiny Parts
Transcript
- Lucas: So you're designing a wearable health sensor, and it needs a tiny housing with a snap-fit latch that's maybe two millimeters wide. You look at CNC machining and the quote is brutal — then you wonder, can you injection mold something this small? Luna: I think most hardware founders assume injection molding is for parts you can hold in your palm, not for something the size of a grain of rice. Lucas: Right, and that's where micro injection molding comes in. It's not just smaller machines — it's a different engineering discipline. Parts under a gram, often with wall thicknesses below a tenth of a millimeter, and tolerances in the microns. Luna: Is it a separate process, or can you just run a smaller mold on a standard press? Lucas: You can technically run a micro mold on a standard press, but the shot size is so tiny that the plastic degrades in the barrel before it gets injected. Dedicated micro molding machines have a screw diameter of maybe 14 millimeters — versus 30 or 40 on a standard press — so the residence time is short and the melt stays consistent. Luna: And the mold itself must be a nightmare. Tiny cavities, tiny ejector pins. Lucas: Absolutely. The tooling is where the cost and the art live. You need extremely precise machining — often with micro EDM or laser ablation — to create cavity details that are maybe fifty microns wide. And the surface finish has to be flawless, because any imperfection shows up magnified on the part. Luna: What about material flow? With such thin walls, the plastic cools instantly. Lucas: That's the core challenge. You need high-flow grades of material — liquid crystal polymers, or certain nylons — and you need to inject at very high speeds, sometimes over a meter per second, to fill the cavity before it freezes. The injection pressure can be two to three times what you'd use for a normal part. Luna: So it's not just shrinking the geometry; you're fighting physics the whole way. Lucas: Exactly. And one thing that surprises startups: you can't just take your CNC prototype geometry and scale it down. You need to design for the process — uniform wall thickness, generous radii, no sharp corners where the plastic can't reach. Every fillet matters. Luna: Is there a typical part size cutoff? Like, below what volume or weight do you switch to micro molding? Lucas: Most sources define micro molding as parts weighing less than a gram, or having features under a hundred microns. But practically, if your part fits on a fingernail, you're in micro territory. A good example is the gear train inside a miniaturized insulin pump — tiny planetary gears with teeth maybe two hundred microns wide. Luna: I actually read about a startup that consolidated a five-part gear assembly into a single micro-molded component. They eliminated the assembly step entirely. Lucas: That's one of the big wins — part consolidation. When you can mold complex geometry with living hinges, snap fits, and gear teeth all in one shot, you remove labor and tolerance stack issues. That startup saved something like forty percent on their unit cost after switching. Luna: But the mold cost must be steep. How does a cash-strapped hardware startup justify that? Lucas: It's a classic tooling amortization question. A micro mold might cost twenty to fifty thousand dollars — comparable to a small conventional mold. But the per-part cost can be pennies. So if you're planning to produce tens of thousands of units, the math works. The hard part is the first article iteration cycle, because modifying a micro mold is expensive. Luna: That's why you see startups doing a lot of prototyping with 3D printing or silicone molding before committing to steel. Lucas: Right. And some micro molders offer prototype inserts made from aluminum or even hardened steel using direct metal laser sintering. You can get a few hundred parts for validation before cutting the production tool. It's not cheap, but it's cheaper than discovering a flow issue after the hard tool is done. Luna: Are there industries where micro molding is becoming a must-have, not a nice to have? Lucas: Hearing aids and in-ear monitors have been using it for decades. But the big growth area right now is in implantable medical devices — think tiny drug delivery ports, catheter tips, and micro-optics for endoscopes. Also, the sensor packages for IoT devices: a temperature and humidity sensor needs a housing that doesn't block the sensing element but protects it. Luna: And those sensors are getting smaller every year. The package has to shrink with them. Lucas: Exactly. One startup I know is molding a micro reed switch housing that's just two by three millimeters, with a wall thickness of a hundred and fifty microns. They had to switch from a standard liquid crystal polymer to a specialty grade with better flow, and they added a hot runner system to keep the melt temperature consistent. Luna: The tolerances on something like that must be insane. What are we talking — plus or minus ten microns? Lucas: For critical features, yeah, ten microns is achievable. For non-critical dimensions, you might hold twenty-five. But the mold steel itself has to be machined to those tolerances, and then account for shrinkage — which is anisotropic in many filled polymers. So the mold designer has to predict how the part warps and compensate before cutting steel. Luna: So it sounds like a partnership: the startup designs the part, but the molder's engineers are really the ones who make it work. Lucas: Absolutely. And the best startups bring the molder in early, before the design is locked. If you wait until your prototype is finalized, you may have to redesign anyway because of draft angles or gate placement. It's much cheaper to simulate mold flow with a partner than to cut tooling twice. Luna: Are there any common mistakes you see first-timers make? Lucas: Three big ones. One: they design flat surfaces perpendicular to the mold opening — no draft angle — so the part sticks. Two: they put the gate in a spot that causes visible weld lines on a cosmetic surface. Three: they use a material with too much shrinkage, so the part doesn't hold the dimension they need. Luna: For listeners who are building something with tiny plastic parts, what's your advice on finding a good micro molding partner? Lucas: Look for a molder that specializes in micro, not just a big injection molder that says they can handle small parts. Ask about their screw diameter, their typical shot size, and their experience with your specific polymer. And ask for references from companies with similar part complexity. Luna: And expect longer lead times for tooling — maybe eight to twelve weeks for a production mold, versus four to six for a standard one. Lucas: And that's if everything goes smoothly. So plan your timeline accordingly. Micro injection molding isn't magic, but it is an enabling technology that lets you build products that were impossible to manufacture at scale just a decade ago. Luna: If today's breakdown of the process and the pitfalls was useful for what you're building or running, that's exactly why we do this show. The way these stay ad-free is listener support — buy me a coffee dot com slash fexingo. Lucas: Yeah, it genuinely makes a difference. No pressure, but it helps us keep digging into niche topics like this every week. Luna: And we love doing it. So with that said, let's get back to the practical side — Lucas, you mentioned mold flow simulation earlier. How accessible is that software for a startup that's not a molding expert? Lucas: There are affordable cloud-based tools now, like Moldex3D's lighter packages or even Autodesk Moldflow in simulation-only tiers. A basic fill analysis for a micro part might cost a few hundred dollars per run. It's not trivial, but it beats cutting steel and guessing. Luna: So you can test your part design virtually before you ever talk to a toolmaker. Lucas: Exactly. And that simulation can reveal things like trapped air, weld line strength, and cooling time. For a micro part, cooling time might be under a second, so the cycle time is dictated more by ejection and handling than by solidification. Luna: And handling those tiny parts after ejection — that's a whole other challenge. How do you even pick them up without losing them? Lucas: Some molders use a robot with a vacuum end-effector tailored to the part geometry. Others use a take-out plate with small cavities that the parts drop into. And then they go into a vibratory bowl or a tape and reel system for automated assembly. It's all custom. Luna: It really makes you appreciate the complexity behind something as simple as a plastic button. Lucas: Exactly. And that's the fun of hardware — the stuff that looks simple is often the hardest to make.