Latest / Hardware Startups with Fexingo: Devices, Robotics, and Manufacturing Tech Companies / How Hardware Startups Use NASA-Qualified Vibration Shakers
Transcript
- Lucas: So you spend months designing a smart thermostat or a handheld medical sensor. You finalize the PCB, get the enclosure injection-molded, assemble a dozen units, and they all work perfectly on the bench. Then you ship twenty to early beta testers, and within a week three come back with dead screens or intermittent power failures. Luna: Classic field failure — and nine times out of ten it's a vibration issue that never showed up in your desk test. Lucas: Exactly. And the fix is something that a lot of hardware startups overlook because they think it's nasa level stuff — industrial vibration shakers. But the same type of system that qualifies a Mars rover payload is increasingly accessible to early-stage companies. Luna: Before we get into the hardware — Lucas, I know we talk a lot about practical tools on this show. If listeners find these deep dives useful for what they're building or running, a couple of dollars a month is genuinely what keeps these episodes ad-free and independent. Buy me a coffee dot com slash fexingo — really helps keep the lights on in the lab. Lucas: Yeah, and it makes a real difference. So — vibration shakers. A proper multi-axis electrodynamic shaker with a slip table and a thermal chamber can run anywhere from fifty thousand to a hundred and fifty thousand dollars new. That's prohibitive for a five-person startup. Luna: But you're not saying they need to buy one. Lucas: No way. There's a growing ecosystem of shared test labs — places like Element Materials Technology, NTS, even some university engineering departments — where you can rent time by the day. Typical rate is around five hundred dollars per eight-hour shift. For a startup that's a big number, but compared to the cost of a recall, it's nothing. Luna: Let's put a number on that. You mentioned a medical device startup — what happened? Lucas: There's a well-known case from about three years ago. A startup building a wearable insulin pump controller. They'd done all the functional testing, passed EMC, passed drop testing. They shipped a hundred units to a clinical pilot. Within two weeks, seven units had failed — the display would flicker and then go blank. Root cause turned out to be a solder joint on the main connector that was hitting its resonance frequency right around 200 Hertz during normal walking motion. Luna: So the user's own footsteps were shaking the device apart. Lucas: Exactly. The fix was a small conformal coating bead and a slight change in the board mounting — cost about twelve cents per unit. But the recall, the delayed pilot, the re-testing — that was estimated at over two hundred thousand dollars. A single day of random vibration testing before shipping would have caught it. Luna: And the shaker lab would have charged maybe a thousand dollars for that day including the fixture design. Lucas: Right. So the ROI is immediate. But the question is — what kind of vibration test should a startup run? There are basically two categories: sine sweep and random vibration. Luna: Sine sweep is the one where you slowly ramp through frequencies, looking for resonances. Lucas: Yes. You mount the device on the shaker, attach a small accelerometer to the PCB or the enclosure, and sweep from maybe 5 Hertz up to 2,000 Hertz. Wherever you see a big spike in the response, that's a resonant frequency — a frequency where the structure amplifies vibration. If that resonance coincides with a known input from the environment — like a car engine at idle, which is around 30 Hertz, or a washing machine spin cycle at maybe 300 Hertz — you've got a problem. Luna: Random vibration is more like a simulation of real-world abuse — you apply a broad spectrum of frequencies at once, with amplitudes based on standards like mil std-810 or IEC 60068. Lucas: Exactly. Random vibration is what you'd use to qualify a product for shipping. The standard profile for parcel delivery, for example, includes energy from about 2 Hertz up to 200 Hertz, with the peak around 10 to 20 Hertz — that's the resonant frequency of most cardboard boxes. Your product inside that box is going to experience that. Luna: And a lot of startups skip it because they think 'we're not military or aerospace.' Lucas: But your product is going to ride in a UPS truck, then a FedEx plane, then a delivery van — that's a pretty harsh vibration environment. I've seen data from a logistics company showing that packages experience sustained acceleration up to 2 Gs in the 5 to 20 Hertz range during transport. Luna: So what's the practical playbook for a startup? When do you do this test? Lucas: I'd say there's a simple rule of thumb: any product with a battery, a motor, or a screen — which is most consumer hardware — should see a vibration profile before the first production run. That's the absolute latest. Ideally, you do a sine sweep on your first functional prototype, even if it's a 3D-printed enclosure, because you want to know if your board has a bad resonance early. Luna: And if you find one, it's much cheaper to fix at the prototype stage than after you've ordered injection molds. Lucas: Exactly. You can move a mounting boss, change a screw location, add a stiffening rib — all of which cost almost nothing in CAD time. After the mold is cut, you're looking at thousands of dollars in modification fees. Luna: What about the test procedures themselves? Do you need a specialized engineer to run the shaker? Luna: And those standards are publicly available — about a hundred dollars to buy. Lucas: Yeah, and some are even summarized in application notes from shaker manufacturers like Unholtz-Dickie or LDS. So the barrier to entry is really just the rental cost and the time. I'd argue that for a startup raising a seed round, spending two thousand dollars on vibration testing across two or three design iterations is one of the highest roi things you can do. Luna: Because nothing kills a hardware startup's reputation faster than early units failing on customers. Lucas: Exactly. And there's an added benefit: investors love to see that you've done reliability testing. It signals that you understand manufacturing risk. I've seen pitch decks where a slide showing a product on a shaker with a 'passed mil std-810' label actually closed the round. Luna: Interesting. So it's not just engineering — it's also a fundraising signal. Lucas: Totally. Hardware investors are scarred by stories of products that died in the field. Showing that you've done the testing — and that your product survived — is a huge de-risker. Luna: Let's talk about one more specific scenario: what if your product is meant to be mounted on a drone or a robot? That's a whole different vibration environment. Lucas: Great point. For drones, you're looking at high-frequency vibration from the motors — typically 150 to 500 Hertz, with harmonics up to a couple thousand. And the amplitude can be significant. I worked with a startup building a LiDAR module for agricultural drones. They were using a COTS IMU that had a resonance at 800 Hertz. The drone motors were producing strong vibration at 780 Hertz. That's a near miss. They had to add a vibration isolator — a small rubber grommet — to decouple the IMU from the frame. Luna: And that's a cheap fix, but only if you know about it before you start flying. Lucas: Exactly. So the general principle — test early, test often, and test with realistic profiles — applies across all hardware. I think in the next five years we'll see desktop shakers become more common, like how desktop reflow ovens are now standard in hardware labs. Luna: Is there a product like that yet? Lucas: A couple of companies are working on it. There's one called Vibration Labs — they're developing a benchtop shaker that's about the size of a desktop 3D printer, targeting under $5,000. It's limited to maybe 5 Gs and 2,000 Hertz, which covers most consumer electronics needs. I haven't seen it ship yet, but the concept is promising. Luna: That would be a game-changer for early-stage hardware studios. You could keep a shaker on the bench and run a quick sine sweep every time you revise the board. Lucas: Exactly. And that kind of iteration speed is what separates hardware startups that ship reliably from ones that struggle with field failures. So to sum up: vibration testing is not just for aerospace. It's a practical, affordable tool that any hardware startup can access today, and the cost of not doing it — in recalls, reputation, and delayed time to market — far outweighs the rental fee. Luna: And the good news is you don't need to become a vibration expert. Just find a test lab, pick a standard, and run the profile. Lucas: Right. One day of testing could save you months of headache. I'm curious — for startups listening, what's the one component in your product that you're most worried about failing under vibration? For me, it's always the battery connector. Luna: Mine would be any large capacitor mounted parallel to the board. I've seen those snap right off. Lucas: Great example. So go shake your product before your customers do. That's all for this episode of Hardware Startups with Fexingo. We'll be back next week with another practical deep dive.