A product can draw only 2 A during normal operation and still reset its adapter every time it is plugged in. The problem is usually inrush current: the short burst required to charge input capacitors, start a converter, initialize a motor, or wake a processor. A power adapter rated for the steady-state load is not automatically rated for that startup event.
The first capacitor is often the largest load
When a product is connected, its input capacitors initially look like a short circuit. The current is limited by the source impedance, cable resistance, connector resistance, protection device, and any intentional soft-start element. A rough estimate is:
Iinrush ≈ C × dV/dt
If the voltage rises quickly across a large capacitance, the current pulse can be much higher than the operating current. Cable inductance and adapter control-loop behavior can then create an input dip or an overvoltage ring.
Measure the voltage at the product connector, not only at the bench supply. A long cable can make a stable source look unstable at the load.
Separate the different startup events
Inrush is not always one clean pulse. A board may first charge its input bulk capacitor, then enable a buck converter, then start a processor or motor. Each stage can create a separate current step. If the protection circuit reacts to the first event but not the second, the symptom may appear as a delayed reset rather than an obvious overcurrent fault.
Create a startup timeline with input voltage, converter enable, power-good, and load current. This makes it easier to decide whether the solution is a slower ramp, a current-limited switch, a larger source, or a change in firmware sequencing.
Choose the protection function deliberately
An NTC can reduce a cold-start surge, but its resistance falls as it heats and it may be ineffective during a quick restart. A resistor limits current but wastes power. A hot-swap controller or eFuse can provide controlled slew rate, current limiting, fault reporting, and reverse-current blocking. A load switch may be enough for a smaller branch.
The power-management IC category at MOZ Electronics is a useful starting point when comparing eFuses, load switches, hot-swap controllers, and related power-path functions. The final choice must match the source voltage, load capacitance, fault energy, and required restart behavior.
Check the capacitor discharge path
After unplugging the product, the input capacitor may remain charged. That affects the next plug-in event, connector arcing, service safety, and the behavior of a power-path controller. Define the required discharge time and verify it under the actual load and protection state.
A first-pass capacitor-discharge calculation can estimate the resistor or active discharge requirement. Use the component’s pulse rating, voltage rating, and temperature rise for the final design.
Layout can add an invisible current limiter
The protection component may be rated for the current while the PCB path is not. Inspect connector pins, vias, fuses, trace neck-downs, and copper pours between the source and the protected load. A narrow path can create a voltage dip that looks like an adapter failure.
For production boards, connect the power-path review to footprint, stencil, polarity, and inspection decisions. MOZPCB’s PCB assembly guidance is useful when a protection device or power switch depends on a reliable exposed pad or a particular assembly orientation.
Validate the worst restart
Test a cold plug-in, a warm plug-in, a fast unplug-and-replug, minimum input voltage, maximum load capacitance, and the highest expected ambient temperature. Capture the connector voltage and current with enough bandwidth to see the first edge. Repeat with the adapter and cable that customers will actually use.
The component derating calculator can help identify weak voltage, current, or power margins around the startup path. For application context, Octatronics’ power and energy resources help connect the protection choice to the final product environment.
The practical lesson is that a power adapter is part of a startup system. Steady-state wattage is only one requirement; the product must also control how it asks for energy during the first few milliseconds.
Top comments (0)