Latest / Hardware Startups with Fexingo: Devices, Robotics, and Manufacturing Tech Companies / How Hardware Startups Use Vibration Testing to Avoid Field Failures
Transcript
- Lucas: So you've got a prototype that works perfectly on your bench. The firmware's solid, the enclosure snaps together just right. But the moment it ships, it has to survive a journey through a parcel network, then sit on a truck dashboard, or maybe get bolted to a piece of industrial machinery that shakes all day long. Luna: And that's where vibration testing comes in. It's one of those things that sounds like overkill until you've had a product come back with cracked solder joints or a loose connector. Lucas: Exactly. And for hardware startups, vibration testing is often the difference between a product that works in the demo and a product that works in the field. Today we're going to talk about why vibration testing matters, how to approach it as a small team, and some of the common pitfalls to avoid. Luna: Let's dive in. So first, what are we actually testing for? I mean, we're not trying to simulate an earthquake, right? Lucas: No, not usually. The goal is to replicate the kinds of vibration a product will see in its real life. Think about a sensor bolted to a conveyor motor, or a handheld device that gets dropped into a bin. The frequencies and amplitudes are fairly predictable. But if you don't test, you're basically hoping your design can handle the unknown. Luna: And hoping is not a strategy. So how do we actually do this in a way that's practical for a startup? I imagine you can't just buy a shaker table and call it a day. Lucas: You can, actually, but it's not cheap. A decent electrodynamic shaker can cost anywhere from twenty thousand to a hundred thousand dollars. Plus you need a controller, accelerometers, fixtures. So most startups start by renting time at a certified test lab. Luna: That's what we did at my last company. We sent our prototype to a lab in the next state, and it came back with a list of failures we'd never seen on the bench. Lucas: That's the value. A lab will run a standard test profile, like a random vibration profile defined by something like the ISTA standard. But the real craft is in choosing the right profile and the right levels, because there's a fine line between under-testing and over-testing. Luna: What does over-testing look like? I mean, isn't more testing always better? Lucas: Not necessarily. If you exaggerate the vibration levels too much, you might break a part that would have been fine in the real world. Then you spend time and money fixing a problem that doesn't exist. That's the classic engineering trade-off. You want to test to a level that's representative of the worst case you expect to see, maybe with a safety factor, but not so high that you're designing for a torture test. Luna: So how do you figure out what the right level is? Is there a standard for that? Lucas: There are industry standards, like mil std-810 for military gear, but most consumer products use something like the ISTA 3A test for packaged goods. Those standards give you a general profile, but they're not specific to your product's environment. To get truly accurate data, you might do a field measurement with an accelerometer data logger on an actual truck or machine. That's advanced, but it's worth it if your product has to survive in a harsh environment. Luna: And what about acceleration levels? I've seen specs like '2 g rms' and wondered what that actually means in practice. Lucas: Great question. The 'g' is the acceleration due to gravity, and rms is root mean square, which is a way to average a random vibration signal. So a 2 g rms random vibration profile is fairly mild. A truck ride on a rough road might produce something like that. But if you're mounting something to a helicopter or a high-speed spindle, you could see 10 g or more. The key is matching the profile to your actual environment. Luna: Got it. So what are some common failure modes you see in vibration testing? It's not just solder joints, right? Lucas: Right, solder joints are a big one, especially with surface mount components. But you also see connector fretting, where a connector's contacts rub together and wear out. You see cracked enclosures, especially if there's a stress concentration. And you see fasteners backing out if they're not torqued properly or if there's no thread-locking compound. I remember a startup that made an automotive sensor, and they failed vibration testing because one of the mounting screws wasn't secured properly. It was a simple fix, but it would have been a costly recall if they hadn't caught it. Luna: That's a perfect example of why testing early is so important. And I guess that's the key message for startups: don't wait until you're in production to test. Lucas: Absolutely. And that brings up another point: vibration testing isn't just about catching failures. It's also about documenting that your product has been tested. That documentation is gold when you're talking to investors or customers, especially in industries like medical devices or automotive where compliance is expected. Luna: So you're saying a test report is a sales tool as much as a quality tool. Lucas: Exactly. It shows you're serious about reliability. And that can be a competitive advantage, especially for a startup trying to win enterprise customers. Luna: Speaking of serious about reliability, I know we're not running ads on this show, and that's a choice we make because we believe in what we're doing. If you find these episodes useful for what you're building or running, you can support the show and keep it ad-free by buying us a coffee at buy me a coffee dot com slash fexingo. It's a small way to keep this going, and we appreciate every bit. Lucas: Yeah, we don't have sponsors to answer to, so we can talk about the stuff that actually matters to operators and builders. Thanks for that, Luna. And speaking of reliability, let's get back to the test plan. Luna: So once you've decided to test, how do you actually set up a test? What's the process? Lucas: The first step is to define the test profile. You can start with an industry standard, but then you tailor it to your product's environment. You decide on the duration, the frequency range, and the intensity. Then you need a fixture to mount your product to the shaker table. The fixture has to be rigid enough that it doesn't resonate at the test frequencies, and it has to mimic how the product is actually mounted in the real world. Luna: I've seen people use a thick piece of aluminum with some threaded holes, is that enough? Lucas: For a simple test, yes. But you have to be careful that the fixture's natural frequency is well above the test frequencies, otherwise you'll be testing your fixture, not your product. If you're testing a heavy or oddly shaped product, you might need a custom fixture that matches the mounting points exactly. Luna: And what about orientation? Should you test in all axes? Lucas: Ideally, yes. Vibration can come from any direction, so testing in all three axes is the most thorough. But that triples your test time and cost. Many startups start with the axis that's most likely to cause problems, like the vertical axis if the product will sit on a vehicle. But if you can afford it, testing in all three axes is the gold standard. Luna: How long should a test run? I've heard of tests lasting hours, but also some that are just minutes. Lucas: It depends on the standard and the intended product life. A typical ISTA test might run one to three hours. But some industries have their own standards. For example, automotive electronic components often need to withstand a certain number of hours at a given level, like 8 hours per axis. The idea is to simulate the cumulative vibration the product will see over its lifetime, but you can't test for years, so you accelerate by increasing the intensity. Luna: There's that acceleration factor again. So you're saying a short but intense test can replicate years of use? Lucas: Right. There's a concept called fatigue damage equivalence. If you double the vibration level, you might be able to test for a quarter of the time. But you have to be careful not to introduce failure modes that wouldn't happen at lower levels, like fatigue in materials that are more sensitive to high stress. That's why it's important to work with a test engineer who knows what they're doing. Luna: That makes sense. So if a startup doesn't have a test engineer on staff, what should they do? Lucas: They should find a lab that offers consulting as part of their service. A good lab will help you define the test profile and interpret the results. They might even suggest design changes that make your product more robust. That's worth the money. Luna: Let's talk about cost. What should a startup budget for vibration testing? Lucas: Renting a shaker at a lab can run anywhere from a hundred to three hundred dollars an hour, depending on the lab and the complexity. A basic test with a standard profile might take a day, so you're looking at a few thousand dollars. If you need custom fixtures or multiple axes, it can easily go to ten or twenty thousand. For a startup, that's not trivial, but it's a fraction of the cost of a recall or a failed field deployment. Luna: And there are ways to save money, like testing early in the design process when failures are cheaper to fix, and maybe testing only the critical components rather than the whole system. Lucas: Absolutely. You can do a quick test on a prototype board to check for component stress, then do a full system test later when you have a production-intent prototype. That two-stage approach can save a lot of money. Luna: So what's the one thing you'd want a hardware founder to take away from this episode? Lucas: That vibration testing is not a luxury. It's a step that separates a product that works in a demo from a product that works in the real world. And doing it early, even with a simple test, can save you from a costly surprise down the road. Luna: Well said. And if you've got a story about a vibration test that caught something unexpected, we'd love to hear it. But for now, thanks for joining us.