Latest / Hardware Startups with Fexingo: Devices, Robotics, and Manufacturing Tech Companies / How Hardware Startups Use Potting Compounds to Survive Harsh Environments
Transcript
- Lucas: So we are in mid-2026, and I keep seeing hardware startups that ship a beautiful prototype, it works great in the demo, and then three months later they are drowning in field failures because a sensor got damp or a connector vibrated loose. Luna: Right — the classic lab versus real world gap. And the fix is usually some kind of protective coating. Lucas: Exactly. But there is a specific technique that is less talked about than conformal coating but often more effective for harsh environments: potting. Encapsulating the entire circuit board in a solid compound — epoxy, urethane, silicone — so it becomes a single block of material. Luna: I have seen videos of people pouring what looks like clear epoxy over a PCB. But is that something a five-person startup can do without a factory? Lucas: That is the part that surprised me. I talked to a founder at a 2025 agtech startup — they were building soil sensors that go into irrigation fields. Mud, water, temperature swings. They started potting in a garage with a two-hundred-dollar vacuum chamber from a used lab-equipment auction. Luna: Wait — a vacuum chamber? To get rid of bubbles? Lucas: Exactly. When you pour epoxy over a board, air gets trapped under components. Those bubbles become pathways for moisture. So you put the board and the compound in a vacuum chamber before it cures, pull the air out, and the epoxy flows into every crevice. Luna: And that gave them IP68 rating? Lucas: Yes. Their field failure rate dropped from eighteen percent to under two percent. That is the difference between a startup that gets reorders and one that burns through its seed round replacing units. Luna: Okay, so potting works. But what about repairability? If you encase the whole board in epoxy, you cannot replace a single component when something fails. Lucas: That is the big trade-off. Once potted, the board is essentially a disposable module. So you design for that — you pot the cheap part and keep the expensive part separable. For example, pot the sensor head but leave the main processor board accessible. Luna: That makes sense for a twenty-dollar sensor, not a thousand-dollar compute module. Lucas: Exactly right. And there are different materials for different trade-offs. Urethane is softer, better for vibration, easier to cut open if you absolutely need to repair. Epoxy is harder, better chemical resistance, but impossible to remove without destroying the board. Luna: Silicone is the third option, right? I have heard that for high-temperature applications. Lucas: Yes. Silicone stays flexible up to maybe two hundred degrees Celsius, but it is more expensive and tends to have lower adhesion. So you need a primer or a mechanical lock — like holes in the board that the silicone flows through and holds onto. Luna: Let me ask about the process itself. How do you even do this at low volume without making a mess? Lucas: The agtech founder I mentioned used a two-part epoxy in a syringe mixer — you know, those dual-barrel cartridges with a static mixing nozzle. He would dispense just enough to cover the board, then put it in the vacuum chamber for maybe five minutes. Luna: No cleanroom? No nitrogen purge? Lucas: Garage. He said the key was degassing the epoxy before pouring — put the mixed compound in the vacuum chamber first, let it bubble, then pour it over the board and do a second vacuum cycle. That gave them bubble-free results. Luna: If these conversations are useful for what you are building or running, there is a simple way to keep them going ad-free. Lucas: A couple of dollars a month is genuinely what keeps these going — buy me a coffee dot com slash fexingo, if you have gotten something out of them. Luna: It makes a real difference, and it is appreciated. Lucas: So back to the potting — one thing that surprised me was the cost. That epoxy they used was about thirty dollars for a quart kit. That potted maybe a hundred sensor boards. So thirty cents per unit in material. Luna: That is negligible compared to the cost of a field failure. Lucas: Exactly. But there is a hidden cost: labor. Hand-potting is slow. If you are doing a thousand units a month, you need jigs, automated dispensing, maybe a conveyor oven for curing. That is where startups trip up — they scale the design but not the process. Luna: So what is the volume threshold where you switch from hand-pouring to automation? Lucas: Based on what I have heard from contract manufacturers, around ten thousand units per year. Below that, manual potting with a vacuum chamber is fine. Above that, you want a dispense robot and a heated cure station. Luna: And what about the new bio-based potting compounds? I saw a startup last year that made a soy-based epoxy. Lucas: That is real. There are a few companies now offering plant-derived polyols for urethane potting. They claim lower carbon footprint and similar performance. But the catch is shelf life — bio-based compounds tend to degrade faster in the cartridge. Luna: So for a startup that pivots or delays production, that could be a problem. Lucas: Right. You do not want to be throwing away fifty-dollar cartridges because the material expired. So for now, most startups stick with conventional epoxies and urethanes. Luna: One more thing: thermal management. If you encase a hot component in epoxy, does it overheat? Lucas: It can. Epoxy is a thermal insulator — about 0.2 watts per meter-kelvin. Compare that to aluminum at 200. So you have to be careful. Some compounds are filled with ceramic or alumina powder to improve thermal conductivity, but that makes them more expensive and harder to dispense. Luna: So you would not pot a power supply, but you might pot a low-power sensor. Lucas: Exactly. Or you pot around hot components, leaving a cavity filled with thermally conductive grease. There are ways to work around it, but you have to design for potting from the start. Luna: That seems like the recurring theme: design for the process, not the other way around. Lucas: It is. And potting is one of those techniques that, if you plan for it, can save you from a nightmare of field returns. But if you add it as an afterthought, you will end up with a mess of bubbles and unrepairable boards. Luna: Have you seen any examples where potting backfired? Lucas: Yes. A drone startup potted their flight controller to protect from crash impact. But they used an epoxy with too high a coefficient of thermal expansion. When the board heated up, the epoxy expanded more than the components, and it cracked solder joints. Luna: So material selection is everything. Lucas: It is. That is why you see experienced startups building a small test matrix: five boards with epoxy A, five with epoxy B, five with urethane. Subject them to thermal cycling and vibration before committing to a hundred units. Luna: That sounds like a week of testing that saves months of field headaches. Lucas: Exactly. And the beauty is that the test setup is cheap. You can build a thermal cycling chamber from a Styrofoam box and a hair dryer. Not precise, but good enough to catch a bad material match. Luna: So for a startup listening right now: if you have a device that will see moisture, vibration, or chemicals, potting might be your cheapest insurance. Lucas: And do not be afraid of the mess. A vacuum chamber, a dual-cartridge gun, and some off-the-shelf epoxy can get you from prototype to field-ready for under five hundred dollars in tooling. Luna: That is a fraction of the cost of even a single recall. Lucas: Exactly. And as the material science improves — especially with bio-based and thermally conductive options — potting will become even more accessible. Luna: It is one of those unsung manufacturing hacks that can make or break a hardware startup.