Latest / Hardware Startups with Fexingo: Devices, Robotics, and Manufacturing Tech Companies / Why Hardware Startups Should Use Helium Leak Testing
Transcript
- Lucas: So you spend months designing a sleek, sealed enclosure for your product. It's IP67 rated on paper. Then you ship a hundred units, and four come back with moisture inside. You lost money, you lost trust, and you have no idea where the leak is. Luna: And that's the moment most hardware founders discover they should have done helium leak testing in production. Lucas: Exactly. Helium leak testing is the gold standard for finding microscopic leaks in sealed assemblies. It's used everywhere from aerospace to medical devices, but most hardware startups don't think about it until it's too late. Luna: I mean, I know the basic idea — you fill the part with helium, then sniff for it escaping. But what makes it so much better than just dunking the thing in water? Lucas: For starters, helium atoms are tiny — about 30 times smaller than a water molecule. So a leak that helium can find might be completely invisible to a bubble test. Also, helium is inert and non-toxic, and the background concentration in air is very low, so you get a clean signal. Luna: Right, so if you detect helium, you know it came from your part. That's clever. Lucas: The typical setup involves a mass spectrometer tuned to helium's atomic mass. You either put the part in a vacuum chamber and spray helium on the outside, or you fill the part with helium and put it in a vacuum chamber to sniff the escaping gas. The sensitivity is down to 10 to the minus 9 cubic centimeters per second — that's a leak so small it would take years to lose a gram of gas. Luna: Okay, but for a startup, that sounds expensive. What are we talking — five figures for a machine? Lucas: A decent benchtop helium leak detector starts around twenty thousand dollars used, up to fifty thousand new. That's real money. But the cost per test is low — maybe a few cents in helium per part, plus labor. The real question is when the investment pays off. Luna: So for a startup making fifty units a month, probably not worth buying the machine. But for a few hundred a month, maybe it starts to pencil out. Lucas: Exactly. Most early-stage startups should use a test lab. There are dozens of contract labs that will run helium leak tests for you — typical pricing is fifty to a hundred dollars per hour, and they can batch test small parts. You can validate your design before you commit to buying equipment. Luna: Have you seen a startup where this actually saved them from a disaster? Lucas: Yeah, I talked to a founder of a small drone company earlier this year. They had a sealed flight controller enclosure that they thought was solid. They sent ten prototypes to a lab for helium testing, and five of them showed a microleak around the USB port gasket. The leak was so small that a bubble test wouldn't have caught it. They redesigned the gasket and moved on. If they had shipped those units, they'd have had field failures within months. Luna: And that's a classic hardware startup story — a minor design detail that looks fine in CAD but fails in real-world sealing. Lucas: The thing is, sealing is incredibly sensitive to surface finish, compression set of the gasket material, and even the torque on the screws. Helium testing gives you objective data on whether your assembly process is consistent from one unit to the next. Luna: So it's not just about finding leaks — it's about process control. Lucas: Exactly. And the data from helium testing can feed back into your design. If every leak is at the same location, you know you have a design issue. If leaks are random, it might be a process issue — inconsistent torque, or a burr on the sealing surface. Luna: And if this kind of detailed engineering talk is useful to what you're building or running, honestly, if today was worth a coffee to you, that's the link — buy me a coffee dot com slash fexingo. Lucas: It's the smallest ask we can make, and it keeps the show independent and ad-free. Really appreciate those who chip in. Luna: So back to helium testing — you mentioned mass spectrometers. Are there different methods depending on the part geometry? Lucas: Yes. For small parts like sensor housings or connectors, the most common method is the vacuum chamber test. You put the part in a chamber, pump the chamber down, then spray helium around the outside of the part. If helium gets in, the spectrometer sees it immediately. For larger assemblies, you might use a sniffer probe — you fill the part with helium, then scan the seams with a handheld probe connected to the detector. Luna: The sniffer method sounds slower — more manual. Lucas: It is. But for a startup doing low volumes, it's practical. You can build a simple fixture to pressurize the part with helium, then have a technician run the probe along the gaskets and welds. The key is to have a known leak standard — a part with a calibrated leak — to verify the system is working. Luna: Are there any gotchas with helium itself? Like, does it get trapped in materials and cause false readings? Lucas: Great question. Helium can permeate through some plastics and elastomers, especially if they're thin. So if you're testing a part with a rubber gasket, you might see a background signal that slowly rises as helium diffuses through the gasket. You have to distinguish that from a real leak. The way to handle it is to establish a baseline — test a known-good part and see what the permeation rate is. Then you set your reject threshold above that. Luna: So it's not a pass-fail binary. You have to calibrate. Lucas: Right. But once you set the threshold, the test is very repeatable. And you can even measure the leak rate quantitatively, which lets you grade parts — this one is tight, this one meets spec, this one is marginal. That data is gold for your quality team. Luna: What about cost of the helium itself? I've heard helium is getting scarce. Lucas: It is a non-renewable resource, and prices have been volatile. But the amount used per test is tiny — we're talking milligrams per part. The bigger issue is availability. If you're using a lab, they handle the supply chain. If you buy your own detector, you'll need a bottle of helium. For a startup, the cost is negligible — maybe a hundred dollars for a bottle that lasts months. Luna: Let's talk about alternatives. There's also pressure decay testing, right? You pressurize the part with air and measure the pressure drop over time. Lucas: Pressure decay is much cheaper — you just need a pressure regulator and a transducer. But it's far less sensitive. You can detect a leak of about 10 to the minus 3 cubic centimeters per second with pressure decay, while helium can go to 10 to the minus 9. So if your product needs to stay dry for years, helium is the way to go. If it only needs to survive a splash, pressure decay might be enough. Luna: So the choice depends on the severity of failure. A waterproof smartwatch? Probably needs helium. A weatherproof outdoor camera? Maybe pressure decay is fine. Lucas: Exactly. And you can combine them. Use pressure decay as a screening test on every part, then use helium on a sample for verification. That gives you cost efficiency with high confidence. Luna: What about fda regulated products? Medical devices? Lucas: For medical devices, helium testing is often mandatory. If you're making an implantable device or a sterile fluid path, regulators expect a quantitative leak test. Helium is the accepted standard. The equipment has to be validated, and you need to document the test method and results. But the bar is high — and helium helps you clear it. Luna: So for a startup in medtech, buying a helium detector might be a necessary cost of doing business. Lucas: Or at least budgeting for lab testing in the validation phase. I've seen startups get through FDA clearance using a contract lab, then buy their own detector once they hit production volume. Luna: That phased approach makes sense — don't buy capital equipment until you know your product is right. Lucas: Right. And speaking of capital, there's another option: some equipment suppliers offer helium leak detectors on lease. You can pay per month with a service contract, which keeps your cash burn lower. Luna: That could be a good middle ground for a startup that's doing maybe a thousand units a month. Lucas: Exactly. And the nice thing is, helium testing scales. Once you have the detector, you can automate the test with a pneumatic fixture and a PLC. Then you're testing every part in seconds, with data logged to your quality system. Luna: I want to circle back to the drone founder story. What was the cost of the helium testing for those prototypes? Lucas: He told me it was about four hundred dollars to test ten units at the lab. That included fixture setup and a written report. So forty bucks per prototype. Compare that to the cost of a field failure — warranty replacement, shipping, damaged reputation — forty bucks is nothing. Luna: That's a great ROI story. And it's easy to pitch to a skeptical CTO. Lucas: The interesting thing is, the founder told me that after fixing the gasket design, he went back and tested the revised prototypes. Zero leaks. That gave him confidence to start production without a 100% helium test — he only does it on a sample basis now. Luna: So the upfront testing actually saved him recurring costs. He doesn't need to test every unit. Lucas: Exactly. And that's the pattern I see with smart hardware startups. They use helium testing early to validate the design and the process, then dial back the testing frequency once they have statistical confidence. Luna: But you have to be careful not to dial back too much. If you change a supplier or a material, you should revalidate. Lucas: Absolutely. Any change in the sealing interface — different gasket durometer, different screw torque spec, different surface finish on the mating parts — all of that can affect leak rates. So you should have a change control process that triggers a new helium test. Luna: Are there any standards for helium leak testing that startups should know about? Lucas: Yes. ASTM E493 is the standard test method for helium leak detection using a mass spectrometer. It covers both the vacuum and sniffer methods. Also, ISO 20485 covers leak testing methods in general. Most labs follow those, and if you're in a regulated industry, you'll probably need to cite them. Luna: Good to know. So if a startup is looking for a lab, they should ask if they follow ASTM E493. Lucas: Right. And also ask about calibration. The lab should have a calibrated leak standard that they use to verify the system before each test session. Otherwise, you don't know if the results are accurate. Luna: I feel like we've covered the main points — when to use helium, how it works, costs, alternatives, and a real-world case. Any final takeaways? Lucas: I'd say: don't wait for field failures to discover you have a leak problem. Helium testing is accessible and affordable if you use a lab. And the data it gives you is actionable. If you're designing a sealed product, put helium testing in your development plan. It's one of those things that feels like overkill until it saves your product. Luna: And if you want to hear more about the drone founder's story, we might get him on the show sometime. Lucas: I'll reach out. For now, if you're building hardware and you're not sure about your sealing, find a lab, run a helium test. It's a small investment for a lot of peace of mind. Luna: Thanks for listening, and we'll catch you next time on Hardware Startups with Fexingo.