How can poor power sequencing cause embedded electronics to fail unpredictably?

Some embedded products fail in ways that seem almost designed to be awkward. They work during one test, reset during another, behave differently after a cold start, or only fail when a radio, sensor, display or actuator wakes at the same time as the processor.

The issue is not always a faulty component or a bad PCB. Sometimes the problem is timing. Embedded electronics power sequencing controls how different parts of a system turn on, stabilise and begin communicating. If that order is wrong, the product may start in an unreliable state before the main application has even begun.

Why start-up behaviour matters in embedded electronics

Start-up is one of the most important moments in an embedded system. It is when power rails rise, clocks begin, processors boot, firmware initialises, sensors settle, memory is checked and communication interfaces come online. A lot happens very quickly, and each part of the system may depend on another part being ready first.

This is not just a theoretical concern. NXP’s 2026 i.MX 95 applications processor data sheet defines an 11-stage power-up sequence, covering supplies for the processor, analogue functions, DDR, PHY, PLL, I/O and reset behaviour. It also warns that deviating from the required power-up and power-down sequences can cause excessive current during power-up, prevent the device from booting or, in the worst case, cause irreversible damage.

Not every embedded product uses a processor that complex, but the example shows the principle clearly. Power-up is not always a single event. In many systems, it is a controlled chain of dependencies. One part of the product may need to be stable before another is allowed to start. A reset line may need to stay asserted until multiple supplies are valid. Memory may need to be powered before the processor can initialise properly. I/O supplies may need to follow a defined order to avoid unpredictable behaviour.

On a simple board, these details may be fairly easy to manage. But as products become more connected and feature-rich, start-up becomes more complicated. A single product may include a microcontroller, voltage regulators, analogue sensors, wireless modules, displays, motor drivers, memory devices and external interfaces. Each of those parts has its own power requirements, timing behaviour and ready state.

If the processor starts reading a sensor before the sensor has stabilised, the firmware may record a false value. If a radio module draws a current burst before the supply rail is ready, the board may dip or reset. If memory is accessed before its voltage or clock conditions are valid, the system may fail to boot properly. If peripheral devices come online in an unexpected order, communication can become inconsistent.

Good embedded electronics power sequencing makes start-up predictable. It ensures the right rails are available at the right time, the processor does not assume devices are ready too early, and firmware has a controlled way to check the system before normal operation begins.

Rails, sensors, radios and processors that wake in the wrong order

Power sequencing problems often appear when different parts of the system have different priorities. The processor may need a stable core voltage before it can boot. Analogue circuits may need a clean reference before measurements are trusted. Radios may need short bursts of high current. Sensors may need warm-up time. Displays or memory devices may need specific reset timing.

Voltage rails are usually the starting point. Many embedded systems use several rails: perhaps 12V input, 5V for peripherals, 3.3V for logic and a lower voltage for the processor core. If one rail rises too slowly, collapses under load or appears before another rail it depends on, the system can enter an undefined state.

Current demand also matters. A cellular module may be physically small, but its power requirements can still be significant. Quectel, for example, specifies that the power supply for one of its LTE modules should be able to provide current up to 3.0A. The same hardware design guide recommends allowing the supply to stabilise for at least 30 milliseconds before pulling down the module’s power key, and specifies a power-key pulse of at least 500 milliseconds.

That kind of detail matters because average current does not tell the whole story. A product may look safe from a steady-state power budget, but still fail when a radio, modem, relay, motor or display wakes at exactly the wrong time. If the power supply cannot handle the burst, the voltage may dip, the processor may reset, or another part of the board may start in an invalid state.

Sensors add their own problems. Some sensors need time to stabilise before their readings can be trusted. Others may return default, invalid or noisy values during start-up. If firmware treats those early readings as real data, the product may make poor decisions immediately after booting.

The processor itself should not be treated as immune to these conditions. Brownout thresholds, reset circuits, watchdog timers and bootloader behaviour all influence whether the system recovers cleanly or gets stuck. In one Texas Instruments microcontroller erratum, an abrupt supply ramp faster than 1V per 100 microseconds can cause a brownout condition to be incorrectly detected, even though the voltage does not fall below the brownout threshold. The result is an unexpected reset.

That is a useful reminder that faster is not always better. Start-up reliability depends on the shape of the power event as well as the final voltage. A multimeter may show that the rail is correct, but it will not show whether the ramp, dip, delay or transient caused the processor to make the wrong decision in the first few milliseconds.

This is why embedded electronics power sequencing is both a hardware and firmware concern. Hardware needs to provide stable power, suitable reset behaviour and enough margin for real loads. Firmware needs to wait for devices to become ready, check status properly and recover gracefully when something does not respond as expected.

Testing power sequencing under real operating conditions

Power sequencing should not only be checked on a clean bench supply. That environment is useful, but it can hide the conditions that create real start-up faults. Products need to be tested in ways that reflect how they will actually be powered, installed and used.

Vehicle-powered electronics are a good example. Analog Devices describes a severe cold-crank condition where a nominal vehicle battery voltage drops to 3.2V. In its example, a traditional protection diode causes the input to a downstream regulator to fall to around 2.7V, triggering undervoltage shutdown and causing the 1.8V output rail to collapse.

On a bench supply, the same embedded product may start perfectly every time. In a vehicle, it may be trying to boot while the input voltage is collapsing, recovering and sharing power with other loads. If the power architecture, reset behaviour and firmware do not account for that condition, the result can be an intermittent reset that is extremely difficult to reproduce later.

For battery-powered devices, testing should include different states of charge, including low battery and recovery after deep discharge. For industrial products, it may mean supply interruptions, long cables, shared power rails and noisy equipment nearby. For connected devices, it may mean waking the radio, sensors and processor in different combinations to see whether the system remains stable.

Testing should also include repeated start-up cycles. A product that boots once is not necessarily reliable. Cycling power many times, across temperature and supply variations, can reveal timing issues that only appear occasionally. This is particularly useful for products that may be switched on and off frequently or left unattended for long periods.

The sequence itself should be measured, not guessed. Engineers may need to look at rail rise times, reset lines, current peaks, communication attempts and firmware logs during start-up. If the product fails unpredictably, capturing what happens in the first few seconds can be more useful than testing the system once it is already running.

Firmware testing matters too. The software should handle missing devices, slow responses, invalid readings and failed initialisation without leaving the product unusable. A sensor that is not ready should not crash the system. A radio that fails to connect should not cause endless resets. A peripheral that takes longer to wake than expected should be handled as a normal condition, not a surprise.

Good embedded electronics power sequencing turns start-up into a controlled process. The system powers up, checks itself, waits for what needs to settle, handles exceptions and only then moves into normal operation.

At TAD Electronics, we design embedded hardware and firmware with real operating conditions in mind. That includes power architecture, sequencing, brownout behaviour, low-power modes, communication timing and recovery from unusual states. If your product is suffering from resets, inconsistent start-up or difficult-to-repeat faults, our risk-free design scoping process can help identify whether power sequencing is part of the problem.

Click here to get in touch or here to read more.

‘Engineering Design, Imagine what could exist’

Got a web design question or mobile application need? Our in-house design agency, Bluebrick Studios, has you covered. Check out their site to find out how they can help you achieve your mission.

FAQ

What is power sequencing in embedded electronics?

Power sequencing in embedded electronics is the order and timing in which different voltage rails, processors, sensors, communication modules and peripherals turn on and become ready during start-up.

Why do electronics fail during start-up?

Electronics can fail during start-up if power rails are unstable, devices wake in the wrong order, current demand causes voltage dips, sensors are read too early or firmware does not handle incomplete initialisation properly.

How can poor power sequencing cause intermittent faults?

Poor power sequencing can cause intermittent faults because start-up timing may change with temperature, battery state, component tolerance, supply behaviour and load conditions. The system may work in one condition but fail when devices wake more slowly, draw more current or respond in a different order.