Latest / Hardware Startups with Fexingo: Devices, Robotics, and Manufacturing Tech Companies / How Hardware Startups Use Environmental Stress Screening
Transcript
- Lucas: So we've talked a lot about testing strategies on this show, from thermal cycling to acoustic testing. But there's one approach that sits right at the intersection of manufacturing and reliability that we haven't really dug into: environmental stress screening, or ESS. Luna: ESS. That's the thing where you stress every single unit before it ships, right? Not just a sample, but the whole batch. Lucas: Exactly. The idea is to push each unit through a controlled dose of thermal, vibration, or electrical stress to expose latent defects that would otherwise show up in the customer's hands. It's not about testing the design; it's about catching the manufacturing flaws — the marginal solder joint, the cracked capacitor, the connector that's not quite seated. Luna: So it's like a final quality gate, but a smart one. And I remember you mentioning a medical device startup that used ESS to cut field failures by forty percent. Can you walk us through that? Lucas: Yeah, I got to see this firsthand. They were producing a portable diagnostic device, and they were shipping thousands of units a month. Their field failure rate was around two percent, which sounds low until you're talking about medical equipment. Two percent of ten thousand units is two hundred angry customers, and in healthcare, that's a big deal. Luna: Two percent does sound scary in that context. So what did they do? Lucas: They implemented a thermal cycling screen. Every single unit spent about four hours going through a cycle from zero to sixty degrees Celsius, with a dwell time at each extreme. Then they added a short vibration sweep at the end. The whole thing added maybe fifteen dollars per unit in cost, but they cut field failures from two percent down to one point two percent. Luna: That's a significant drop. But I'm curious — why did it work? What were they catching? Lucas: Mostly poor solder joints on a specific power connector, and a couple of components that were prone to early thermal fatigue. The thermal cycling stressed the joints exactly the way they'd be stressed in the field — maybe not as harshly, but enough to make a weak joint fail during screening rather than in the ICU. Luna: So it's the difference between testing a sample and testing every unit. And I guess that's the core of ESS — you're not looking for design flaws, you're looking for process variation. Lucas: Right. And that's why ESS is so valuable for startups that are scaling up. When you're moving from prototype to volume, your manufacturing process is still maturing, and you're more likely to have these intermittent defects. ESS catches them before they become a reputation problem. Luna: Now, I've heard of HALT — Highly Accelerated Life Testing. How is ESS different from HALT? Lucas: HALT is a design validation tool. You take a few units, you stress them to destruction, and you learn the limits of the design. ESS is a production screen — you apply a defined, milder stress to every unit to catch latent defects without damaging good ones. HALT tells you where the cliff is; ESS keeps you from driving off it. Luna: That's a useful way to put it. So for a startup, should you do HALT first, then set up ESS based on what you learn? Lucas: Ideally, yes. HALT can help you understand the margins of your design, and then you can set ESS profiles that are well within those margins but still effective at catching the defects that matter. But a lot of startups skip HALT and just start with ESS because it's simpler to implement. Luna: And is that a mistake? Lucas: It can be. If you don't know your design margins, you might set your ESS profile too harsh or too mild. Too harsh and you're over-stressing good units, potentially shortening their life. Too mild and you're not catching anything, so it's wasted money. There's a sweet spot, and it usually comes from data. Luna: You mentioned the cost — fifteen dollars per unit. That adds up. How do startups decide if ESS is worth it? Lucas: It comes down to the cost of failure. For that medical device, a field failure meant a service visit, possibly a replacement, and potential regulatory scrutiny — easily a few hundred dollars per incident. So fifteen dollars per unit to avoid even a fraction of those failures was a no-brainer. For a consumer gadget with a lower failure cost, you might not need the same level of screening. Luna: So it's an economic decision as much as an engineering one. Lucas: Exactly. And that's something we talk about a lot on this show — the business of hardware. Speaking of which, if these conversations are useful for what you're building or running, a lot of listeners have been asking how to support the show. The easiest way is to buy us a coffee — it's buy me a coffee dot com slash fexingo. It genuinely helps us keep this going, and we appreciate anyone who does. Luna: Yeah, and it's totally optional — but it's nice to know people find this useful. Alright, back to ESS. So what are the most common stress types? Lucas: Thermal cycling is the big one, because temperature changes cause expansion and contraction, which stresses solder joints and interconnects. Vibration is second — it catches loose components and poorly secured connectors. And then there's electrical stress, like power cycling, which can reveal shorts or weak insulation. Luna: Do most startups use all three, or is one typically enough? Lucas: It depends on the product and the failure modes you're worried about. For a stationary device, thermal might be enough. For something that goes into a car or a drone, vibration becomes critical. I've seen startups that do a combined thermal and vibration screen, which is more realistic for many applications. Luna: And when in the manufacturing process should ESS happen? Is it at final assembly, or earlier? Lucas: That's a great question. The general rule is to screen early enough that you catch defects before they get covered up by later assembly steps. For example, if you screen the main PCB before it goes into the housing, you can rework it easily. If you wait until the final product, you're dealing with a more complex teardown. Luna: Makes sense. But doesn't screening at the component level defeat the purpose? I mean, you're not stressing the entire assembly. Lucas: Right, that's the trade-off. You want to screen at a level where defects are detectable but also where you haven't added too much cost. Some startups do a quick thermal screen on the bare board, then a more comprehensive screen on the final assembly. It's about finding the sweet spot. Luna: What about startups that are low-volume, maybe a few hundred units a month? Is ESS even feasible for them? Lucas: It's more feasible than you might think. You don't need a huge chamber. There are benchtop thermal chambers that can handle a few boards at a time. And you can start with a simpler screen, like power cycling with a temperature monitor. The key is to be consistent and track the results. Luna: What kind of data should startups be tracking? Lucas: You want to track the failure rate during screening, the types of failures, and where in the process they occur. That data will tell you if the screen is effective and if your manufacturing process is improving. If you see the failure rate decreasing over time, that's a good sign. If it stays constant, maybe you have a deeper issue. Luna: So ESS can also be a metric for your manufacturing quality. Lucas: Exactly. It's a health check for your production line. And when you do catch something, you've got a perfect sample for failure analysis — you can open it up, find the root cause, and fix the process. That's where the long-term value is. Luna: I've also heard that ESS can be used to 'burn in' a product, like running it for a few hours to catch early failures. Is that the same thing? Lucas: Burn-in is a type of ESS, but it usually refers to running the device at elevated temperature for an extended period, sometimes with functional testing. It's common in electronics to weed out infant mortality failures. But it's not the same as thermal cycling — burn-in is more about time and load, while thermal cycling is about stress from temperature changes. Luna: And which is more common for hardware startups? Lucas: I'd say thermal cycling is more common for catching latent defects, but burn-in is still used, especially for products that are expected to run continuously, like servers or industrial controllers. It really depends on the application. Luna: Going back to the medical device example — that startup cut field failures to one point two percent. But is there a point of diminishing returns? Lucas: Definitely. You can't catch everything, and over-screening can reduce your throughput and increase cost. The goal is to get your field failure rate to a level that's acceptable for your industry and your customer expectations. For medical, that might be under one percent. For a toy, maybe two percent is fine. Luna: So it's about being strategic about where you apply ESS. Lucas: Right. And as your manufacturing process matures, you might find that you can reduce the level of screening or even eliminate it for certain components. It's a dynamic process. Luna: That makes sense. So for a startup just starting out, what's the first step to implement ESS? Lucas: Start by looking at your field failure data. If you're already shipping, what are the most common failure modes? If you're pre-launch, look at similar products and their warranty claims. That will tell you what to screen for. Then start small — maybe a simple thermal cycle on a sample of units to see if it catches anything. If it does, you can ramp up gradually. Luna: And if it catches nothing? Is that a bad sign? Lucas: Not necessarily. It might mean your manufacturing process is already good, or your stress profile is too mild. You need to calibrate. But sometimes catching nothing is actually a good sign — it means you're shipping quality products. Luna: I guess there's a bit of a mindset shift here. Instead of just hoping for the best, you're actively looking for problems before your customers do. Lucas: Exactly. And that's the essence of ESS — it's proactive quality. For hardware startups, where a field failure can tank your reputation and your funding prospects, it's worth taking seriously. Luna: So what's the one thing you'd want our listeners to remember from this episode? Lucas: That ESS is a business decision as much as an engineering one. It's about balancing the cost of screening against the cost of failure. And if you're scaling up, it might be one of the best investments you can make in your product's reliability.