Latest / Hardware Startups with Fexingo: Devices, Robotics, and Manufacturing Tech Companies / How Hardware Startups Use Vibration Testing to Survive Shipping
Transcript
- Lucas: You spend months designing a prototype, get it assembled perfectly, ship it to your first beta customer, and it arrives dead. The most common cause isn't a design flaw in the electronics — it's shipping vibration. Luna: We've all seen those 'fragile' stickers. But you're saying the damage is real and measurable, not just a logistics excuse? Lucas: Very real. I was talking to a medical wearable startup last month — they lost 18 percent of their first production batch to units that failed out of the box. The culprit? A single pallet in transit hit acceleration spikes over 5 G's on a random vibration profile. It snapped off unsupported ceramic capacitors and cracked solder joints on a connector near the edge of the board. Luna: Eighteen percent — that's brutal for a startup on a first run. How do you even catch something like that before you ship? Lucas: You vibration-test the prototype. The standard method is a random vibration profile based on the expected shipping environment. For a parcel shipped via FedEx or UPS, there are industry-standard profiles — typical overall G-rms levels around 1.15 to 1.5, but the peak accelerations can spike much higher. The test runs on an electrodynamic shaker with a slip table, and it'll cost you maybe two thousand dollars for a day at a test lab. Luna: Okay, so two grand to avoid an 18-percent scrap rate. That math works. But what's the difference between random vibration and the sinusoidal vibration I hear about? Lucas: Sinusoidal is a single frequency sweep — think of a fan blade spinning at a constant RPM. It's good for finding resonances. But real-world shipping vibration is random — multiple frequencies at once, varying amplitudes, like driving a truck over a gravel road. The random profile better simulates the chaotic bumps and rattles your product will actually see. Luna: So you test the bare board? Or the whole assembly in its enclosure? Lucas: You test the fully assembled product, preferably in its shipping package. The package itself can amplify or dampen vibration. The test lab will strap your product to the slip table, attach accelerometers to the board at critical points, and run the profile for maybe an hour per axis. Three axes — X, Y, Z — so three hours. Then you inspect for failures. Luna: What are you looking for? Cracked solder joints, loose connectors, broken components? Lucas: All of the above. The most classic failure is a large ceramic capacitor — say a 10 microfarad, 1206 package — soldered near the edge of an unsupported board. Under vibration, the board flexes, the capacitor's brittle ceramic cracks. You can catch that with X-ray inspection after the test. Also, connectors with surface-mount tails often develop micro-cracks in the solder fillets. Luna: And if you find cracks, what's the fix? Thicker board? More mounting screws? Lucas: Several fixes. First, avoid cantilevered PCB mounting — where the board is only supported along one edge. Use four-point mounting with standoffs at all four corners. Second, for heavy components like large capacitors or transformers, apply a bead of conformal coating or a staking compound — room-temperature vulcanizing silicone — to lock them down. The silicone adds damping and prevents the component from wiggling. Luna: I've seen some startups use a full coating over the entire board. Is that overkill? Lucas: It can be, if your only concern is vibration. Conformal coating is primarily for moisture and contamination. For vibration, selective staking on the heaviest parts is usually enough. But if you're also dealing with humidity or dust, a full acrylic or parylene coating makes sense. Just remember, coating adds cost and can complicate rework. Luna: Got it. So the test is cheap relative to the scrap cost. But I'm guessing a lot of startups skip it because they think their design is 'rugged enough' or they don't know the test exists? Lucas: Exactly. I've seen hardware teams run thermal tests and EMC tests religiously, but vibration is an afterthought. And the funny thing is, a lot of the damage happens on the last mile — the local delivery truck bouncing over potholes. The same medical wearable startup I mentioned — once they added a vibration test to their qualification plan and designed in four-point mounting and staked the connector, their shipping failure rate dropped to under one percent. Luna: That's an incredible return on a two-thousand-dollar investment. A lot of these episodes remind me that most hardware failures are physics problems, not electronics problems. Lucas: Right. And physics problems are predictable — they follow the same equations every time. You just have to take the time to measure and mitigate. Which brings me to something else that's been on my mind. If these conversations are useful for what you're building or running, we keep the show ad-free and accessible because a small number of listeners chip in. Luna: Yeah, it's a simple model. A couple of dollars a month genuinely covers the hosting and research costs. It makes a real difference. Lucas: Buy me a coffee dot com slash fexingo — that's it. No tiers, no exclusive content, just knowing you're helping keep this going for everyone. Luna: And we appreciate it. So back to vibration — you mentioned the test profile is based on the shipping carrier. How do you get that data? Lucas: There are published standards: ASTM D4169 and ISTA 2A, for example. ISTA — the International Safe Transit Association — has profiles for various shipment types: parcel, truck, air freight. The test lab will have those programmed into their shaker controller. You just tell them 'I want ISTA 2A parcel' and they run it. Luna: So even a two-person garage startup can call a test lab and say 'run ISTA 2A on my prototype' and get meaningful data? Lucas: Absolutely. And if you don't have a local test lab, there are mail-in services. You ship them your product, they test it, ship it back with a report. Turnaround is often a week. The bigger hurdle is culture — you have to be willing to break your own prototypes to learn. Luna: Breaking things on purpose. That's a hard mindset shift for a lot of founders. Lucas: It is. But the alternative is breaking them in your customer's hands. And that's a much more expensive lesson.