Pop Goes the Hinge
Pop Goes the Hinge
Diagnosing early iPhone Duo failures using engineering RCA frameworks, the gaps in IP68 testing for hinged devices, design as deterministic creativity, and more.
Topics worth your time: viewing foldable reliability from factory yield to the bathtub curve, deterministic design catalyzed by bursts of creativity, what an IP rating can (not) tell you about a hinge, and another big week of funding for autonomy, manipulation, and grid storage.
P.S. Interested in learning about medical devices? Join us live to talk through unique considerations in medical device prototyping on 11/10. Register here!
Interesting Links
Between reported production yield problems and customer reports of early-life device failures, both of which appear to be hinge-related, iFixit’s teardown is worth reading to understand the design of Apple’s newest foldable and the category at large. A useful way to think about foldables is to start with the enabling technology: flexible OLEDs, which first shipped as fixed, curved panels in the Samsung Galaxy Round & LG GFlex. A screen that bends enables a phone that folds, but getting there requires considerable mechanical complexity, compounded by manufacturing and assembling a hinge to strict tolerances across millions of units. To some degree, entering the market years later can reduce that risk by learning from the failure modes of earlier devices, but a new design still has to prove itself in production. Any mix of mismatched tolerance stack-ups, process drift, and assembly variation can translate into problems at scale. The first iPhone Duo units reached customers this weekend, and a number of owners are reporting a similar failure sequence: the hinge binding partway through opening, an audible ‘pop,’ and a line of failed pixels across the inner display. iFixit’s Lead Teardown Technician walks through a possible failure mechanism (starting at 14:30), and a separate investigation by Korean YouTuber CuriousHagi reveals a hinge link fractured at its notch, with damage to the inner display directly above it. It’s too early to know how widespread these failures are, but they follow reports of final-assembly yields just above 60% at Foxconn, with analysts pointing to the hinge as a key production bottleneck.
Note: yield is measured at the factory, so 60% yield ≠ 40% failure rate. It may suggest a process running with little margin.
These early failures are reminiscent of the original Samsung Galaxy Fold, which also ran into display reliability issues in 2019. The suspected mechanisms differ, but the failure modes make for useful case studies for understanding product reliability. A useful reliability framework from issue #121:
1) Reliability from development to production. Early prototypes often fail due to design flaws or manufacturing uncertainties. Over time these issues are resolved, but reliability can dip again when scaling during development phases or production ramp.
2) The “Bathtub Curve”: a classic model that maps failure rates across a product’s lifespan: early “infant mortality” failures, a long period of stable operation, and eventual “wear-out” due to aging and stress.
This chapter from NASA provides an excellent overview of reliability engineering, covering failure types and statistical methods used to predict system performance.
Frontier models are quite capable in specific engineering domains where the result is verifiable. Ballista is showcasing automated tube routing for fluid systems, a task that would normally require a large manifold and a lot of guess-and-check by a team. In fluids engineering terms, it’s a packaging problem; creating compact tubing is difficult, but not necessarily difficult to verify for software or for a human (though it is perhaps quite time-intensive). One catch: in Ballista’s example, the tool also moved parts of the assembly that aren’t tubing, so the scope of edits (what is in/out of bounds) appears to be untested. For the fundamentals of what goes into a rocket’s fluid system, MIT Rocket Team’sintro to feed system design is a great place to start.
Alex Slocum’sFUNdaMENTALs of Design is a wacky, creative, and highly useful textbook for the principles of design engineering, written for MIT’s robot-building design course. Topic 1 covers the design process, and it reads far differently (dare I say, in a more engaging fashion) than your standard textbook. His lesson is that design can be both creative and deterministic: ideas funnel from broad strategies down to individual components, every risk gets a countermeasure, and a FRDPARRC table (pronounced “fred park”, short for functional requirements, design parameters, analysis, references, risks, and countermeasures) keeps track of it all. Design is a passionate process “to be carried out in a careful, systematic, deterministic manner catalyzed with random hyper adrenalin driven bursts of super creativity.” Watch the full lecture for Topic 1 here.
IP68 means dust-tight (IP6X) and protected against continuous immersion beyond 1 m (IPX8, depth and duration set by the manufacturer) under IEC 60529, which raises a question: at what point does a standards body need to revisit its test methods? For IP6X, the device sits in a chamber of circulating talc powder, sieved to under 75 µm, for up to 8 hours while a slight vacuum is drawn on the enclosure to pull dust through any gap. Any dust found inside is considered a failure. It’s a reasonable proxy for a standard smartphone, but the test doesn’t specify that moving parts need to be actuated during or after exposure. For devices built around an exposed hinge, the rating may say less than it appears to.
And one fun link to round out the week:
Sponsored: Viam andMongoDB are hosting a Robot Hackathon in SF on Oct 29–30. Teams get a robot arm, a depth camera, and time with industry mentors. The top three win SO-101 kits and up to $10K in credits!
Presented by DRCY: AllSpice’s AI Design Review Agent
Meet DRCY, AllSpice’s AI-powered agent built specifically for hardware design reviews. DRCY runs first-pass checks on schematics and datasheets to surface critical errors early, such as config errors, voltage mismatches, and much more. Hardware teams can spend less time on rechecks, spend more time innovating, and ship with confidence.
Give DRCY a try here.
Startup News
In out-of-left-field news, Warhound emerged from stealth with $200M in seed funding to clone and genetically engineer working dogs for the military and emergency responders. The startup comes from Palmer Luckey and Colossal Biosciences CEO Ben Lamm. It plans to biobank DNA from top-performing dogs, clone them, and use gene edits to screen out inherited diseases. Investors include Sequoia, In-Q-Tel, and ARK Invest.
Bedrock Robotics raised a $600M Series C to expand deployments of its autonomy kit, which retrofits existing excavators and other construction equipment so they can operate on their own. The round was co-led by CapitalG and Emergence Capital.
Xona Space Systems raised a $250M Series D to scale production of its low Earth orbit navigation constellation. The company says Pulsar’s signals are up to 100x stronger than GPS with centimeter-level precision. Four of the six satellites it recently launched will support beta service starting in early 2027. The round was led by Future Ventures.
Eka Robotics raised a $150M Series A to expand its vision-force-action model for robotic manipulation. The model combines camera input with force readings from Eka’s gripper, so robots can detect slippage and adjust their grip mid-task. It’s trained largely in simulation with reinforcement learning, and Eka says the system handles about 1,500 picks per hour. The round was co-led by a16z and F-Prime.
Peak Energy raised a $108M Series B to scale manufacturing of its sodium-ion grid storage systems and fill more than $1B in existing orders. The company’s batteries are passively cooled and don’t require the active cooling systems most grid batteries rely on. The company’s 4 GWh factory in Sacramento is set to start production in Q1 2027. The round was led by GM Ventures.
Open Jobs
Describe your ideal role on our job board, and we’ll find matches. If you’re hiring, promote your open role here.
Early Career:
Jabil is looking for a Mechanical System Architect Intern (Summer 2027) in Austin, TX
Panasonic Avionics is looking for a Member of Technical Staff (Hardware) in Irvine, CA
Mid-Level:
Revel is looking for a Forward Deployed Test Engineer in Los Angeles, CA
Aeva is looking for a Systems Engineer, LIDAR Characterization in Mountain View, CA
Senior to Staff:
OpenAI is looking for an Electrical Engineer, PCBA Test Infrastructure, Robotics in San Francisco, CA ($318–485K)
AnySignal is looking for a VP of Engineering, Ground Hardware in Los Angeles, CA ($240–325K)
Tools From Our Sponsors
Design & Simulation
Duro — AI-native PLM that centralizes part data, manages change orders, and connects the whole stack so your team can focus on execution, not administration.
Onshape for Startups – Free access to Onshape Professional or Onshape Government (compliant with ITAR/EAR) for hardware startups and entrepreneurs.
Viam – Infrastructure for physical AI. Develop, deploy, and operate robotics applications with any agent, any model, and any hardware.
Manufacturing
Loombotic – Instant-quote custom wire harnesses. Upload drawings or design online, fast turnaround.
Xometry — Instant quoting on custom parts, from R&D prototypes to high-volume production runs.
Fictiv – On-demand custom manufacturing from prototype to production.
Access Every Link We’ve Ever Shared 🎞️
Refer three friends to Hardware FYI and unlock our full archive.
Missed an issue? Hunting for that one link? Every tool, guide, and article we’ve ever shared is in here.
P.S. If you find something interesting—an article, tool, funding round, project, really anything—send it our way.
P.P.S. 20,000+ engineers, founders, and operators read this newsletter. If you want to reach them, get in touch.
© 2026 Benji Chia
548 Market Street PMB 72296, San Francisco, CA 94104