Latest / Hardware Startups with Fexingo: Devices, Robotics, and Manufacturing Tech Companies / How Hardware Startups Use Stencil Design for Perfect Solder Paste Deposition
Transcript
- Lucas: If you've ever watched a pick and place machine run, you know the board whips through in seconds. But what really determines whether those components solder right or not happens milliseconds before the chip ever touches the board. It's the stencil. Luna: The stencil. The metal sheet with the holes. Most founders I talk to barely think about it until their first assembly run goes sideways. Lucas: Exactly. And it's one of the highest-leverage decisions in SMT assembly. A well-designed stencil costs maybe 200 dollars and can save you twenty thousand in rework. A bad one guarantees defects. Luna: So what are we talking about when we say 'stencil design'? Thickness, aperture size, wall finish? Lucas: All of that. Let's start with thickness. The standard for most boards is 0.005 inches — five thou. But if you're mixing huge QFNs with tiny zero-four-zero-two capacitors, that one thickness might not work for both. Luna: Right, because the larger part needs more paste volume to get a good joint, but the tiny cap will get too much paste and tombstone. Lucas: Tombstoning — where the part stands up on end because surface tension pulls unevenly — is a classic volume-mismatch problem. So you have to decide: do you compromise, or do you use a step stencil? Luna: Step stencils — where part of the stencil is etched thinner or thicker — add cost and complexity, but sometimes they're the only way. Lucas: Right. A step stencil might have a 0.006-inch region for the power components and a 0.004-inch region for the fine-pitch parts. The challenge is the step transition — if it's too abrupt, paste can smear. Luna: I've seen that. A startup I worked with tried a two-step stencil and the paste bridged across the step. They ended up going back to a single thickness and accepting a slightly higher defect rate on the small parts. Lucas: That trade-off is real. But there's a rule of thumb that helps: aperture area ratio. It's the area of the aperture opening divided by the area of the side walls. You want that ratio above 0.66 for good paste release. Luna: Below 0.66 and the paste sticks to the stencil walls instead of transferring to the board. Lucas: Exactly. So for a 0.005-inch stencil, a 0.010-inch-square aperture has an area ratio of about 0.5 — too low. You'd need at least a 0.015-inch-square opening to get above that threshold. That matters when you're placing 0201 components, which have a pad size around 0.010 by 0.012. Luna: So if you're using 0201s — and more startups are, to save space — you basically can't use a 0.005-inch stencil. You need to drop to 0.004 inch. Lucas: Or even 0.003 inch for 01005 parts. And that changes your whole print process. Thinner stencils flex more, they wear faster, and they're more sensitive to squeegee pressure. Luna: Let's talk about the aperture walls. You mentioned finish earlier. I've read that electropolishing the walls can improve release by reducing roughness. Lucas: Electropolishing is huge. When a laser cuts the stencil, it leaves a rough, slag-like edge inside the aperture. That roughness grabs paste particles. Electropolishing smooths the walls and rounds the corners slightly — it's like sanding a wooden mold so the concrete doesn't stick. Luna: Does that affect the aperture size? If you're polishing away material, doesn't the hole get bigger? Lucas: It does, slightly. Typically by 0.0002 to 0.0005 inches per wall. That's why you'd specify the aperture size after electropolishing, not before. And most good stencil houses will automatically compensate if you tell them you want electropolishing. Luna: What about nano-coatings? I've seen vendors offering a coating that makes the surface even more slippery for paste. Lucas: Nano-coatings are a newer option. They're typically a fluoropolymer or silane-based layer applied to the bottom of the stencil and the aperture walls. They reduce the coefficient of friction and improve paste release, especially for fine-pitch apertures. Luna: Do they last? I imagine the coating wears off after a few thousand prints. Lucas: Vendors claim around ten thousand prints, but in practice it's less, especially if you're using aggressive cleaning solvents. Still, for a pilot run of a few hundred boards, it can give you better first-pass yield. Luna: And yield is everything when you're a startup. One bad stencil can wipe out your whole production schedule. Lucas: Which brings me to something I really want founders to hear: before you commit to a stencil vendor, ask for a sample print report. A good vendor will print paste through your stencil onto a glass slide and measure the volume and height of each deposit with a 3D solder paste inspection machine. They'll share the CpK values for each aperture. Luna: CpK — process capability index. You want above 1.33 ideally. Lucas: Right. If your vendor can't or won't provide that, find another vendor. It's the cheapest quality insurance you'll ever buy. A stencil is 200 bucks. A production batch of boards can be tens of thousands. Luna: You know, this reminds me of something. If today's episode helps you avoid a stencil disaster, it's the kind of practical knowledge that keeps this show useful. And the way we keep it ad-free is listener support. If you've gotten value from these conversations, you can find us at buy me a coffee dot com slash fexingo. Just a way to keep the show going for folks who are building real hardware. Lucas: Yeah, it's a small gesture that lets us stay independent and focused on what matters for founders. And no pressure — we're just glad you're listening. Luna: Alright, back to stencils. So let's say you've got a design with a mix of large and small apertures. You've chosen your thickness, you've specified electropolishing. What about aperture shape? I've seen rectangular, homeplate, and even circular apertures. Lucas: Shape matters for release. For a given area, a square aperture releases paste better than a rectangle because the walls are more uniform in the direction of pull. For very fine pitch, you'll see homeplate-shaped apertures — a pentagon that tapers at the corners — which reduce the chance of bridging between adjacent pads. Luna: Bridging is when two deposits merge and short the pins. For something like a 0.4-millimeter-pitch QFN, a tiny shift in the stencil can cause that. Lucas: Exactly. And that's where stencil tension comes in. A stencil is stretched over a frame, and the tension needs to be high enough that it doesn't lift off the board during printing. If the tension is low, the stencil can 'walk' or sag, causing misregistration. Luna: What's the right tension? I've heard numbers like 35 Newtons per centimeter. Lucas: That's typical for a new stencil. But tension decays over time, especially if you're using a metal squeegee. Some stencil houses offer a tension measurement certificate with the stencil. That's another thing to ask for. Luna: So you get your stencil, you mount it, you set up your print. What's the first sign something's wrong? Lucas: Inconsistent paste height. If you're using SPI, you'll see some deposits tall, some short. That can be a stencil clog, a damaged aperture, or a pressure issue. But the most common root cause is a stencil that wasn't cleaned properly between prints. Luna: Cleaning frequency is a whole other topic. Some shops clean every print; others every five prints. What's right? Lucas: It depends on paste type and environment. A good rule is to clean every print for fine-pitch work. Under-wipe and vacuum cleaning can extend the interval. But the stencil is the bottleneck — if it's dirty, nothing else matters. Luna: I want to wrap by talking about something that catches a lot of startups: fiducials on the stencil. Lucas: Yes. The stencil needs fiducial marks that align with the board's fiducials. If the stencil fiducial is off by even 20 microns, your paste will be off, and the pick and place will also be off because it aligns to the board, not the paste. That's a recipe for misalignment that's very hard to diagnose. Luna: So when you order the stencil, you're also specifying the fiducial locations based on your board design. Lucas: Exactly. And you should verify them on the first stencil with an optical measurement. It takes 10 minutes and can save you hours of debug. Luna: Alright, I think we've given the audience a solid framework. Stencil thickness, aperture area ratio, wall finish, tension, fiducials. Anything else you'd add? Lucas: Just one thing: if you're designing your PCB, talk to your assembly house before you finalize the layout. They may have preferences on aperture shape or panelization that make your stencil life easier. The earlier you loop them in, the fewer surprises you'll have. Luna: Good advice. Next episode we'll talk about the squeegee itself — metal versus polyurethane, angle, pressure. It's all connected. Lucas: Looking forward to it. For now, if you're in the middle of a prototype run, check your stencil spec. It might be the cheapest fix you can make.