Latest / Hardware Startups with Fexingo: Devices, Robotics, and Manufacturing Tech Companies / Hardware Startups Mastering The Art Of Design For Manufacturing
Transcript
- Lucas: There is a specific moment in every hardware startup’s life where the dream meets the factory floor. You have a working prototype that looks beautiful on the desk, but now you need to make ten thousand of them without going bankrupt. That gap is where most founders get stuck. Luna: And it is rarely because they lack funding. It is usually because they designed something that is impossible to build efficiently. What is the one mistake you see over and over? Lucas: It is the obsession with perfect aesthetics at the expense of assembly logic. I remember speaking with a robotics team last month who spent months refining the external shell of their autonomous bot. They wanted seamless joints and hidden screws. But when they went to manufacturing, the assembly time per unit doubled because technicians had to reach around tight corners to snap pieces together. Luna: So the design looked good but cost more to put together? That seems like a classic case of vanity metrics overriding operational reality. Lucas: Exactly. And the cost compounds. If an assembly station takes three minutes longer per unit, you need more stations, more labor, and more floor space. Suddenly your margin evaporates before you even ship the first box. The fix is simple in theory: involve manufacturing engineers during the design phase, not after. Luna: But most founders are engineers themselves. They want to hold the final product in their hands before letting anyone else touch the specs. Lucas: Right. So we end up with what I call the 'hidden screw' problem. You see it everywhere in consumer tech. A designer decides a screw should be hidden under a rubber foot or behind a logo. It looks clean. But now the factory has to use adhesive bonding instead of mechanical fasteners. Adhesives require curing time, specialized equipment, and quality checks to ensure the bond holds. Mechanical fasteners just click. Click is fast. Click is cheap. Luna: I bet the adhesive bonds fail faster too. Temperature changes expand and contract materials differently than metal screws do. Lucas: You nailed it. Reliability drops, returns go up, and the brand takes a hit. Let us look at a specific example. Consider a smart home sensor company we worked with recently. They were trying to mimic the sleek look of high-end audio gear. Zero visible seams. The result? Their initial production run had a thirty percent defect rate because the casing warped during injection molding. The tolerances were too tight for the plastic material they chose. Luna: Thirty percent? That is catastrophic for a launch. Did they scrap the whole batch? Lucas: They did not have the capital to. They reworked them manually. Every single unit. It took two weeks and burned through their runway. After that, they changed their approach. They started designing for ease of assembly. Wider tolerances, standardized screws, and modular sub-assemblies. The next run was flawless. Luna: That is a harsh lesson. But it highlights why supply chain visibility matters so much. You cannot optimize what you do not understand. Lucas: Agreed. And this brings us to the broader topic of component standardization. Founders love custom parts. They think it gives them a unique edge. But custom parts mean custom tooling. Custom tooling means long lead times and high upfront costs. If you can use off-the-shelf connectors or standard battery cells, you save millions in tooling fees. Luna: It feels counterintuitive though. If everyone uses the same parts, how do you differentiate? Isn't that just commoditizing your product? Lucas: Differentiation comes from the software, the user experience, and the brand narrative. Not from the internal circuit board layout. Think about Apple. They use many standard components. Their magic is in the integration and the ecosystem. Hardware startups should focus their innovation budget on what the customer actually sees and feels, not on the invisible internals. Luna: That makes sense. The customer does not care if the PCB is custom-designed unless it enables a feature they love. Speed to market is often more valuable than perfection in the internals. Lucas: Precisely. And speed relies on having suppliers who are ready to go. If you design a proprietary connector, you are betting on a supplier who will produce it reliably. If you use a standard usb c port, you have a hundred vendors competing to sell you the same thing. Competition drives down price and improves service. Luna: So the advice is: keep the exterior unique, but make the interior boringly standard. Is that the golden rule? Lucas: Boringly standard is the best kind of standard. It reduces risk. It allows you to pivot quickly if a component goes obsolete. Remember how we talked about component obsolescence in earlier episodes? Standard parts have longer lifecycles and better support networks. Luna: I remember that. It is like insurance against the future. If your custom part disappears, you might have to redesign the entire device. With standard parts, you just swap vendors. Lucas: Exactly. Now, let us talk about testing. Many startups skip rigorous stress testing on the pilot units because they are eager to launch. They assume the prototype behavior will hold up at scale. It never does. Luna: Why is that? Are they rushing the process? Lucas: Partly. But also because they do not test the manufacturing process itself. They test the design. There is a difference. You might have a great design, but if the assembly line introduces vibration or static discharge, the product fails. You need to test the product as it comes off the line, not just in the lab. Luna: That is a subtle but critical distinction. Lab conditions are controlled. Factories are chaotic. How do you account for that chaos in your testing regime? Lucas: By building fixtures and jigs that simulate the factory environment. If a technician has to twist a cable at a forty-five-degree angle to connect it, test the cable at that exact angle. Do not test it straight out of the bag. Real-world usage includes real-world installation. Design for both. Luna: So ergonomics matter for the assembler too, not just the end user. It is human factors engineering applied to the factory floor. Lucas: Spot on. Repetitive strain injuries slow down lines. If a task is awkward, workers make mistakes. Mistakes create defects. Defects kill margins. It is all connected. Luna: This conversation is really opening my eyes to how interconnected everything is. From the first sketch to the final shipment, every decision ripples forward. Lucas: It does. And that is why I am so passionate about this topic. Most people think hardware is just about making things. It is really about managing complexity. Luna: If today was actually useful to you, the way these stay ad-free is listener support — buy me a coffee dot com slash fexingo. Lucas: We appreciate any help that keeps this show running. It lets us dig deeper into these technical details for you. Luna: Thanks for listening. Back to the factory floor. Lucas: Let us return to the topic of mold design. This is where the rubber meets the road, literally. Injection molding is the workhorse of hardware manufacturing. But it is also expensive. Steel molds can cost fifty thousand dollars or more. Luna: That is a huge upfront investment. How do founders decide if the mold is worth it? Lucas: By calculating the break-even point. If you only need five hundred units, a soft silicone mold might be cheaper. If you need fifty thousand, steel pays for itself quickly. But there is a third option: aluminum molds. They are cheaper than steel and faster to produce, but they wear out sooner. They are a sweet spot for mid-volume production. Luna: So it is not just about the number of units. It is about the timeline. If you need the product now, aluminum gets you there faster. Lucas: Correct. Time is money. Delaying production by four weeks because you waited for a steel mold can cost you more in lost sales than the mold difference. But you must plan for the eventual switch to steel if you expect longevity. Luna: What about the design features inside the mold? Things like draft angles and undercuts? Lucas: Draft angles are non-negotiable. Plastic shrinks as it cools. If the walls are parallel, the part sticks to the mold and breaks. You need a slight taper, usually one degree or more, so the part releases cleanly. Undercuts are tricky. An undercut is any feature that prevents the part from sliding straight out. Like a hole drilled through the side of a box. Luna: Those sound like nightmares for manufacturing. How do you handle them without complicating the mold? Lucas: You can use slides or lifters in the mold. These are moving parts that retract to allow the part to release. But each slide adds cost and maintenance to the mold. Every moving part is a potential failure point. The fewer moves, the better. Design around the geometry. Avoid undercuts if you can. Luna: So simplicity wins again. Fewer moves, fewer problems. It seems like the fundamental principle of good engineering is reductionism. Lucas: It is. Reduce complexity, reduce risk, reduce cost. When you simplify the design, you simplify the manufacturing. And when you simplify the manufacturing, you simplify the business. Luna: That is a powerful framework. Simplicity is not just aesthetic. It is economic. Lucas: Absolutely. And it extends to packaging too. Often overlooked, but packaging is part of the assembly line. If your product requires twelve steps to pack, you need more labor. If it snaps into a tray with one motion, you save time. Luna: I never thought about packaging as an assembly step. But it makes sense. It is still handling the product. Lucas: Exactly. Design the package to protect the product while enabling rapid assembly. Use molded pulp instead of custom foam inserts. Pulp is cheaper, recyclable, and easier to form. Foam often requires gluing or stapling, which adds steps. Luna: Sustainability and efficiency aligning. Another win for simplicity. Lucas: Yes. And consumers appreciate it. Clean, minimal packaging signals quality. Bulky, wasteful packaging signals laziness. You get brand value for free by making smarter choices. Luna: It is amazing how many layers of value you unlock just by thinking carefully about the process. What is the biggest takeaway for our listeners building hardware? Lucas: Involve manufacturers early. Design for assembly. Keep it standard. Test the process, not just the product. And never underestimate the power of a well-placed screw. Luna: A well-placed screw. I love that. Simple, effective, and essential. Lucas: It really is. Sometimes the smallest detail determines whether you succeed or fail. Luna: Do you think AI will change how we approach DFM in the next few years? Lucas: It already is. Generative design tools can suggest optimal geometries for weight and strength. But they still need human oversight for manufacturability. AI can propose a shape, but a human must decide if the factory can actually build it. Luna: So AI is a tool, not a replacement. It augments the engineer's intuition rather than replacing it. Lucas: Precisely. The human element remains crucial. Experience matters. Knowing what works in a factory is hard-won knowledge that algorithms struggle to replicate. Luna: Here is to combining old-school wisdom with new-school tools. Thank you for joining us today. Lucas: Thank you. Next time we talk, we will look at how to navigate global shipping logistics without losing your mind. Luna: Sounds like a challenge. Until then, keep building.