Introduction
There’s a unique excitement in designing and building Internet of Things (IoT) devices. From selecting the perfect microcontroller to writing the firmware that brings it to life, the process is a rewarding journey of creation. But beneath the surface of software logic and component selection lie powerful, non-obvious physical and electrical principles that can quietly sabotage a product, turning a brilliant concept into an unreliable failure with abysmal battery life.
These silent killers aren’t typically covered in introductory tutorials; they’re found buried deep within application notes, high-reliability design guides, and manufacturing handbooks. They represent the hard-won knowledge that separates a hobbyist project from a robust, mass-market product.
This article distills five of the most surprising and impactful lessons from these deep-dive technical documents. We’ll explore the hidden forces related to power management, mechanical stress, electromagnetic interference, and manufacturing that every IoT and embedded systems designer should master.
1. Your Device Isn’t “On” or “Off”—It’s Mostly Sleeping, and That Changes Everything
The difference between a naive “always on” design and a sophisticated low-power architecture isn’t a minor optimization—it’s the difference between a product that works for a week and one that can last for a decade on a single battery. The impact of aggressive low-power design is staggering and often underestimated.
A stark example from a STMicroelectronics application note (AN6044) illustrates this. A simple MCU implementation that remains in “run mode” drains its battery in just 5 days. The exact same function, when implemented with a low-power strategy, could last for over 8 years. This isn’t a 10% or 20% improvement; it’s a monumental shift in a product’s viability.
The core strategy involves leveraging the microcontroller’s various power states, such as Run, Stop, and Standby. Instead of continuously running, the device spends the vast majority of its time in the lowest possible power state. It wakes only for the briefest moment to perform a critical task—like taking a sensor reading or transmitting data—and then immediately goes back to sleep. This requires a fundamental change in firmware architecture, moving from a simple while(1) loop to an event-driven model where the MCU only responds to interrupts or wakeup events. Adopting this “sleep-first” mentality is the single most important factor for achieving multi-year autonomy in battery-powered IoT devices.
“Sleep if you can”.
2. Your PCB Isn’t Just an Electrical Circuit; It’s a Mechanical System Tearing Itself Apart
Engineers are trained to think of printed circuit boards (PCBs) in terms of voltages, currents, and signal paths. We often forget that a PCB is a physical, mechanical object subject to stress, strain, and fatigue—especially when exposed to the temperature swings of a harsh environment.
One of the most insidious failure modes stems from the Coefficient of Thermal Expansion (CTE). The common FR-4 substrate of a PCB is a composite of woven fiberglass and epoxy resin. This structure is anisotropic, meaning its physical properties are not the same in all directions. While the fiberglass weave effectively minimizes expansion along the X and Y axes, it does little to prevent expansion in the Z-axis (the thickness of the board). The CTE of copper, used for plating the barrels of vias, is much lower than the Z-axis CTE of the FR-4.
The consequence of this mismatch is severe. As the device heats up and cools down, the board expands and contracts more in thickness than the copper via running through it. This “differential z-axis expansion leads to shear stress between the via wall and the dielectric layer.” Over hundreds or thousands of thermal cycles, this repeated stress can physically crack the copper barrel of the via, causing intermittent connections or a complete, maddeningly difficult-to-diagnose open circuit. This failure is counter-intuitive because it isn’t a component that has failed; it is the very structural integrity of the board that has given way.
3. You’re Building Accidental Antennas That Are Screaming Noise
Every trace on your PCB has the potential to become an antenna, radiating electromagnetic interference (EMI) that can disrupt nearby electronics or cause your own device to fail mandatory EMC compliance tests. This often happens completely by accident, born from a misunderstanding of how current actually flows in a circuit.
The culprit is “loop area.” A signal doesn’t just travel down a trace; it must also return to its source, typically through a ground plane. The physical space enclosed by the signal trace and its return path forms a current loop. A larger loop is a more efficient antenna, meaning it is better at radiating—or receiving—electromagnetic noise. A high-speed digital signal routed far from its return path creates a large loop that can “scream” EMI, derailing your project during certification.
Furthermore, the ground plane itself is not just a simple “zero-volt” reference. The entire device, especially its ground plane, acts as an extension of the antenna system. Its size, shape, and integrity are critical parts of the RF design that dictate wireless performance and EMI characteristics. The key takeaway is to always consider the return path for every critical signal. By keeping signal traces and their return paths as close together as possible, you minimize the loop area and prevent the creation of these “accidental antennas.”
4. Those Convenient Holes in Your Board are Secretly Stealing Your Solder
Placing an open via—a simple plated-through hole—very close to a component pad seems like a harmless and convenient way to route a signal. However, during manufacturing, this innocuous hole can become a thief, directly causing a critical assembly defect known as “solder thieving.”
The mechanism is simple physics. During the solder reflow process, the solder paste on the component pads liquifies. If an open via is nearby, the molten solder is wicked down into the via hole by capillary action. This leaves an insufficient amount of solder on the pad to form a robust, reliable connection with the component’s lead. This problem is especially pronounced for vias placed in the large thermal pads under many modern ICs, where they can also cause outgassing issues that physically push solder away from the intended area.
Fortunately, the solutions are well-established. The cheapest, though less reliable, option is “tenting” the via by covering it with solder mask. A more robust solution is to have the vias filled with LPI solder-resist. For the highest reliability, especially for vias in pads (like under a BGA), the best option is to specify epoxy-filled and capped vias, which are plugged and then plated over with copper to create a perfectly flat, solderable surface. It’s a striking example of how a seemingly trivial design choice can have a direct and costly impact on manufacturing yield and product reliability.
5. A Product’s Life Begins on the Shelf, Not in the User’s Hands
As a designer, it’s natural to focus on the active life of a product—how it will perform when the end-user powers it on. But for any high-volume product, there is a long and often overlooked period between the factory and the first use. A device can sit in a warehouse or on a retail shelf for months, or even years, before it’s activated. If the product isn’t explicitly designed for this dormant period, its battery could be significantly—or completely—drained before the customer ever opens the box.
This is why a dedicated “shelf mode” is a critical design requirement. There are two main approaches. The simplest is a physical battery disconnect, like the clear plastic tabs you pull out of a new toy. But for sealed, modern electronic devices, the solution must be electronic: a true ultra-low power state that keeps the system alive with minimal energy consumption.
A prime example is the startup circuitry on the STEVAL-ASTRA1B asset tracker, which is designed for this exact purpose. In its shelf mode, the entire startup circuit consumes “approximately 150 nA, or about one milliamp per year.” This incredibly low draw ensures the battery’s capacity is preserved during storage. This forces a crucial shift in the designer’s perspective: you are not just designing for the active life of the product, but for its inactive life as well.
Conclusion
Moving a design from a proof-of-concept to a reliable, mass-market product requires a mastery of challenges that live at the intersection of electrical engineering, physics, and manufacturing science. The most dangerous problems are often the ones you don’t see coming: the mechanical stress from thermal expansion, the unintentional antennas created by poor layout, or the slow drain of a battery in a warehouse. By understanding these hidden forces, we can design more robust, reliable, and successful products.
Now that you’ve seen these hidden forces at play, which one might be silently shaping the success—or failure—of your current project?