Latest / Hardware Startups with Fexingo: Devices, Robotics, and Manufacturing Tech Companies / How Hardware Startups Use PCB Depaneling for Faster Production
Transcript
- Lucas: You get a PCB prototype back from the fab, and it's a panel with ten copies of your board, all connected by little tabs or a scored line. You need to separate them — that's depaneling — and most hardware startups mess this up at least once. Luna: Mess up how? I mean, you just snap them apart, right? Lucas: That's exactly the instinct, and it works — until it doesn't. I talked to a sensor startup last month that was snapping their boards by hand and getting about a 15 percent defect rate. Cracked traces, chipped edges, components popping off. They switched to a simple router jig, cost them maybe two hundred dollars, and their defect rate dropped below one percent. Luna: Fifteen percent is brutal when you're only making a few hundred units. That's real money. Lucas: Exactly. And look, I know this sounds like a niche topic, but if you're building hardware, you're going to face depaneling. So let's talk about the two main methods — V-scoring and routing — and when each makes sense for a startup. Luna: And if today's worth a coffee to you, that's the link — buy me a coffee dot com slash fexingo. Seriously, these deep dives take a lot of research, and listener support keeps the show ad-free. Lucas: Yeah, absolutely. It's a small thing that adds up. So, V-scoring — the board fab cuts a V-shaped groove on both sides of the panel along the break line. You can snap the boards apart by bending them. It's fast, cheap, and zero tooling. But the edge you're left with is the V-groove itself, which is a bevel, not a straight edge. If your board needs to slide into a tight enclosure or a card guide, that bevel can cause problems. Luna: Right, so V-scoring is great for simple rectangular boards where the edge finish doesn't matter. But if you have a complex shape or components near the edge, routing gives you more control. Lucas: Exactly. Routing uses a small CNC bit to cut around each board, leaving them attached by little tabs — usually three to five tabs per board. You break the tabs after routing. The edge is clean and straight. The trade-off is you need a router machine, which can be a few thousand dollars for a benchtop model, or you can outsource it to an assembly house. Luna: But even with routing, there are gotchas. If the tabs aren't placed right, or if you snap them too aggressively, you can still crack the board. Lucas: Absolutely. Tab placement is a design for manufacturing issue. You want tabs near supported areas, not near fragile components. And the way you break them matters — you should score the tabs first with a V-cut or use a tab router that cuts them cleanly. The sensor startup I mentioned was using a hand-held rotary tool with a carbide bit, a cheap jig, and it worked beautifully. Luna: So what would you recommend for a startup that's just moving from prototype to low-volume production, say a few hundred boards? Lucas: I'd say start with V-scoring if your board shape allows it. It's the cheapest and fastest. If the edge quality matters, then go with routing. For a few hundred boards, you can buy a benchtop router from companies like LPKF or a cheaper Chinese CNC that costs under a thousand dollars. Or you can just send the panels to a local assembly shop — they'll depaneling for maybe ten to twenty cents per board. Luna: And if you're doing thousands of boards, then you're looking at dedicated depaneling machines — like a punch machine or a laser depaneler. Those are big capital investments. Lucas: Right. Laser depaneling is amazing for fragile boards with lots of components — no mechanical stress, no vibration. But you're talking fifty thousand dollars plus. For most startups, the router or outsourcing is the sweet spot. The key is not to ignore it. I've seen startups burn through a whole batch of assembled boards because they snapped them wrong and cracked a BGA — a ball grid array package — and then they had to rework or scrap the whole lot. Luna: That's painful. And it's such an avoidable mistake. So what's the single best piece of advice you'd give to a founder listening right now? Lucas: Design for depaneling from day one. Talk to your PCB fab about panelization — they can suggest tab locations and V-scoring angles. And build a simple test: take a scrap panel, separate a few boards, inspect the edges under a microscope, and measure any cracks or chipping. Do that before you run your first real assembly. It takes an afternoon and saves you weeks of headaches. Luna: And if you're prototyping, you can even use a pair of tin snips or a shear — but only for very simple boards. For anything with components, invest in that jig. Lucas: Yeah, or even a cheap tile cutter — I've seen startups use a manual tile snap cutter for V-scored boards. It works surprisingly well. The point is, depaneling is one of those steps that seems trivial until it costs you a batch. Treat it like a real process, and your production will be much smoother. Luna: Alright, so to recap: V-score for simple rectangular boards with loose tolerances, route for complex shapes or when edge finish matters, and always test your depaneling method before full production. Lucas: That's the summary. And if you're curious about panelization rules — like minimum tab width, or how to avoid copper near the V-score — we can do a follow-up episode specifically on DFM for depaneling. Let us know if that's useful. Luna: I'd listen to that. For now, buy me a coffee dot com slash fexingo if this helped — and we'll be back next week with another hardware deep dive.