Latest / Hardware Startups with Fexingo: Devices, Robotics, and Manufacturing Tech Companies / How Hardware Startups Use Design for Manufacturability
Transcript
- Lucas: If these conversations are useful for what you're building or running, remember that small contributions keep this network ad-free and independent. A couple of dollars a month is genuinely what keeps these going — buy me a coffee dot com slash fexingo. Luna: Yeah, it really does help us dig deeper into the technical weeds without chasing sponsors who don't understand the hardware lifecycle. Thanks for supporting the show. Lucas: Let's get back to it. We've spent a lot of time talking about how hardware startups cut costs through panelization or sensor fusion, but there is one foundational discipline that kills more early-stage companies than supply chain disruption: Design for Manufacturability, or DFM. It is the gap between a perfect prototype and a profitable product. Luna: Right, because a prototype only has to work once, but a manufactured device has to work ten thousand times, consistently, and cheaply. Where do you see founders making the biggest mistake here? Lucas: It usually starts with geometry. I recently spoke with an engineer at a mid-sized robotics startup that was struggling to scale a precision arm. They had designed all their structural joints with sharp ninety-degree internal corners because it looked clean in the CAD model. But when they sent those files to a CNC machining shop, the cost spiked by nearly thirty percent overnight. Luna: That sounds counterintuitive. Why would a sharp corner cost more to machine than a rounded one? Lucas: Because cutting tools are cylindrical. You cannot drill a square hole with a round drill bit without leaving a fillet. To get a true ninety-degree internal corner, the machinist has to use a special form tool or perform multiple operations to clear out the material. Adding even a two-millimeter radius to that corner allows them to use standard end mills, which are faster, cheaper, and less prone to breaking. Luna: So a tiny geometric tweak completely changes the toolpath and the bill of materials. That feels like a lesson that doesn't stick until you pay the invoice. Lucas: Exactly. And it compounds. Once you introduce those standard radii, you also open up possibilities like injection molding for later volumes. Injection molds absolutely require draft angles and radii to release the plastic from the steel cavity. If your design fights physics, the factory will fight you on price. Luna: It makes me think about how software engineers can refactor code freely, but hardware engineers are locked into physical constraints once the tooling is cut. Is that the core tension? Lucas: That is part of it. The other huge area is tolerance stackup. Founders love to specify tight tolerances on everything because they want premium feel. But if you ask for plus or minus zero-point-zero-five millimeters on every single interface in a robotic assembly, the parts will never fit together on the first try. The variations add up. Luna: So the cumulative error across five or six parts exceeds the allowable gap, and now you have scrap piles instead of finished products. Lucas: Precisely. The smart move is to relax tolerances on non-critical surfaces. Let the housing be plus or minus zero-point-two millimeters, which is easy to mold, and keep the critical bearing surfaces tight. You save money on machining time and inspection labor while maintaining performance. It is a classic engineering trade-off. Luna: I remember reading about a consumer electronics startup that reduced their defect rate by half just by changing the order of assembly steps. How does process flow tie into DFM? Lucas: It ties in directly. If a designer places a screw deep inside a sealed enclosure where no tool can reach, the assembly line worker either spends four minutes wrestling with a flexible shaft driver or the design gets rejected. Modular designs that allow top-down assembly are almost always cheaper to produce than complex interlocking structures. Luna: So DFM isn't just about the part itself, it's about how that part interacts with the human hands putting it together. Lucas: Yes, and the material choice plays into that too. Some founders pick exotic alloys because they sound impressive in a pitch deck. But if that material requires specialized heat treatment or corrosion resistance that your end user doesn't actually need, you are burning cash. Standard stainless steel or aluminum often suffices and is easier to source. Luna: It reminds me of the old adage that simplicity is the ultimate sophistication. In hardware, simplicity is literally the path to profitability. Lucas: I agree. And we should mention fasteners. Using five different types of screws in a single device creates inventory nightmares and slows down assembly. Standardizing on one or two head types and thread sizes can reduce assembly time significantly. It seems minor, but in high-volume manufacturing, seconds add up to millions. Luna: That standardization argument is so strong. It applies to software libraries too, but hardware makers often overlook it until they are ordering custom packaging and bins for dozens of unique components. Lucas: Speaking of packaging, another DFM win is designing for flat-pack shipping. Bulky enclosures waste freight space. If a robot arm can disassemble into two pieces for shipping, you cut logistics costs by thirty percent. That savings can be passed to the customer or kept as margin. Luna: Logistics is such an invisible cost center. Most founders think about the box arriving, not the box taking up half a shipping container. Lucas: It is. And let's talk about testing access. If you seal a device completely, how do you test it before it ships? Many designs forget to leave test points accessible. Adding small access holes or modular covers allows for automated testing on the line, which catches defects early rather than after the customer unboxes it. Luna: Early detection is obviously cheaper than returns. But sometimes those access holes compromise the aesthetic or the IP rating. How do you balance that? Lucas: You use magnetic covers or gasketed ports that maintain the seal during operation but pop off for testing. It adds a tiny bit of complexity, but it saves thousands in warranty claims. It is better to spend a few cents on a magnet than hundreds on a replacement unit. Luna: That is a practical heuristic. Cost of prevention versus cost of failure. It seems like DFM is really just risk management disguised as engineering. Lucas: I think that is a fair way to put it. And it requires collaboration. Too many teams hand off CAD files to manufacturing partners and expect miracles. The best results come when designers and factory engineers sit together during the prototype phase to review every joint and surface. Luna: Collaboration is key. But getting a mechanical engineer to listen to a factory technician can be hard when the engineer thinks they know best. How do you bridge that culture gap? Lucas: By involving the manufacturer early. Bring them in during the design freeze, not after. Their feedback on tooling limitations can shape the design before any steel is ordered. It shifts the dynamic from adversarial to partnership. Luna: So the timeline extends slightly earlier, but the overall project velocity increases because you avoid re-spins. It is a net win. Lucas: Absolutely. And one final point on materials: consider the environmental impact. Regulations are tightening globally on restricted substances. Designing with compliant materials from day one avoids last-minute reformulation crises when you scale internationally. Luna: Sustainability is becoming a constraint, not just a nice to have. It forces smarter material choices that often align with durability and cost anyway. Lucas: They usually do. Recycled plastics are improving in quality and price. Using them can lower your carbon footprint and appeal to eco-conscious buyers, all while potentially reducing raw material costs. Luna: It seems like the most successful hardware startups treat manufacturing as a co-creator, not just a vendor. That mindset shift changes everything. Lucas: I believe that is the difference between a hobby project and a scalable business. When you respect the factory floor, it respects your bottom line. Luna: What is one actionable step a listener could take tomorrow to improve their own DFM score? Lucas: Review your current CAD models and look for any internal corners that aren't radiused. Add a two-millimeter fillet where possible and calculate the potential tooling savings. It is a low-hanging fruit that pays immediate dividends. Luna: Simple, concrete, and likely to save money. I like that. It proves that small details drive big outcomes. Lucas: Exactly. As we look ahead to next quarter, I wonder if we will see more startups adopting generative AI tools to automatically optimize their designs for manufacturability before they even hit the drawing board. Luna: That would be fascinating. If algorithms can predict factory friction, it might democratize some of this expertise for smaller teams who can't afford senior manufacturing consultants. Lucas: It certainly has the potential. For now though, the fundamentals of good design still rule the roost. Thanks for joining us on this deep dive into the mechanics of making things.