A switched-mode power supply (SMPS) converts electrical power by rapidly switching a transistor on and off, typically thousands of times per second, to transform voltage levels with far less wasted energy than older designs. Where a traditional linear power supply might operate at around 60% efficiency, an SMPS typically achieves 80% or higher by avoiding the constant heat dissipation that linear regulators produce. This core trick of switching, rather than continuously regulating, is what makes SMPS the dominant power supply design in laptops, phones, TVs, and virtually every modern electronic device.
Linear vs. Switching: Why Switching Wins
To understand why SMPS exists, it helps to see what it replaced. A linear power supply takes incoming AC voltage, steps it down through a large transformer, converts it to DC, and then uses a regulator that essentially burns off excess voltage as heat. Think of it like controlling water flow by partially blocking a pipe: the water you block still has to go somewhere, and in this case it becomes waste heat. That works fine for small, low-power devices, but it’s bulky, heavy, and inefficient at higher power levels.
An SMPS flips the process. Instead of stepping down the voltage first with a big transformer, it rectifies the incoming AC to DC right away, then chops that DC into a high-frequency AC signal using a fast-switching transistor. Because this new AC signal oscillates at tens or hundreds of thousands of cycles per second (compared to the 50 or 60 cycles per second from your wall outlet), the transformer needed to step it down can be dramatically smaller and lighter. After the transformer, the signal is rectified back to DC and filtered smooth. The result is the same stable DC output, but with a fraction of the size, weight, and heat.
The Switching Cycle Step by Step
At the heart of every SMPS is a power transistor, most commonly a MOSFET. This transistor acts as an ultra-fast electronic switch. When it’s fully on, current flows through with very little resistance and almost no energy is lost. When it’s fully off, no current flows at all. The only moment energy is wasted as heat is during the brief transition between on and off, which is why faster switching and cleaner transitions matter so much to efficiency.
Here’s what happens in each cycle:
- Switch ON: The transistor closes the circuit, and current flows through an inductor (a coil of wire wound around a magnetic core). Energy builds up in the inductor’s magnetic field as the core material becomes magnetized.
- Switch OFF: The transistor opens the circuit, cutting off the current. The magnetic field in the inductor collapses, and as it does, it releases the stored energy. This collapsing field generates a voltage with the opposite polarity to what was originally applied, pushing current forward to the output. This energy exchange between input and output is a direct application of a physics principle called Lenz’s law.
- Filtering: Capacitors on the output side smooth out the rapid pulses into a steady DC voltage that your device can use.
This cycle repeats thousands of times per second. Typical switching frequencies for consumer power supplies range from about 20 kHz to several hundred kHz, well above the range of human hearing. Higher frequencies allow even smaller transformers and inductors, though they also create new engineering challenges around heat and electrical noise.
How the Output Stays Stable
The key to delivering a rock-steady voltage despite changing loads (your laptop drawing more power when gaming versus idle, for example) is a feedback loop using pulse width modulation, or PWM. The controller chip continuously monitors the output voltage and compares it to a fixed reference voltage. If the output dips below the target, the controller keeps the transistor switched on for a longer fraction of each cycle, delivering more energy. If the output climbs too high, the on-time shrinks.
This ratio of on-time to total cycle time is called the duty cycle. A 50% duty cycle means the switch is on half the time and off half the time. By adjusting the duty cycle dynamically, the controller compensates for fluctuations in the input voltage, changes in load current, or even temperature shifts inside the supply. The adjustment happens so quickly (within a few switching cycles) that the output voltage appears perfectly constant to whatever device it’s powering.
Common Circuit Configurations
Not every SMPS is wired the same way. The arrangement of the switch, inductor, and diode determines what kind of voltage conversion the supply can do. Engineers call these arrangements “topologies,” and each one suits a different job.
- Buck converter: Produces an output voltage lower than the input. This is the most common topology, found in voltage regulators on computer motherboards that step 12V down to the 1V or so a processor needs.
- Boost converter: Produces an output voltage higher than the input. Used in battery-powered devices where a single cell’s 1.5V or 3.7V needs to be raised to 5V or 12V.
- Buck-boost converter: Can produce an output that’s either higher or lower than the input, and in some configurations inverts the polarity. Useful when the input voltage swings above and below the desired output, like a battery draining over time.
- Flyback converter: Uses a transformer to provide electrical isolation between input and output, meaning there’s no direct electrical connection between the wall outlet and your device. This is the topology inside most phone chargers and small AC adapters.
For higher-power applications like server power supplies or industrial equipment, more complex topologies come into play: forward converters, half-bridge, full-bridge, and phase-shifted full-bridge designs. These handle hundreds or thousands of watts while maintaining high efficiency and manageable heat levels.
Electrical Isolation and Safety
In any power supply that plugs into a wall outlet, keeping dangerous mains voltage completely separated from the low-voltage output is critical for safety. SMPS designs that use a transformer (like the flyback) achieve this naturally, since the input and output sides of the transformer are magnetically coupled but electrically separate.
But isolation creates a problem for the feedback loop. The controller needs to know what’s happening on the output side, yet it sits on the input side, separated by that isolation barrier. The standard solution is an optocoupler: a tiny component that contains an LED and a light sensor sealed together. The output side drives the LED brighter or dimmer depending on how far the output voltage strays from its target. The light sensor on the input side picks up that signal and passes it to the PWM controller, all without any electrical connection crossing the barrier. The output voltage is monitored, compared to a precision reference, and the resulting error signal modulates the optocoupler’s LED current to close the feedback loop across the isolation boundary.
Dealing With Electrical Noise
The one major tradeoff of switching power supplies is electromagnetic interference (EMI). Every time the transistor snaps on or off at high speed, it generates sharp electrical transients that can radiate as radio-frequency noise or travel back along the power cord. This is why cheap phone chargers sometimes cause static on a nearby AM radio, or why poorly designed SMPS units can interfere with sensitive audio equipment.
To keep this noise under control, SMPS designs include input EMI filters, typically a combination of inductors and capacitors arranged as a low-pass filter. The inductor blocks high-frequency noise from escaping back to the power line, while the capacitor shunts it to ground. Additional damping networks (small resistor-capacitor pairs) tame resonant peaks that the filter itself can introduce. Regulatory standards require all commercial power supplies to meet conducted and radiated emission limits, so any properly certified SMPS has already been tested and filtered to keep its noise within acceptable bounds.
Why SMPS Dominates Modern Electronics
The combination of high efficiency, small size, and light weight makes SMPS the only practical choice for most modern electronics. A linear supply delivering 200 watts would need a transformer weighing several pounds and a large heatsink to handle the 80 or more watts lost as heat. An SMPS doing the same job might weigh a fraction of that and lose only 40 watts or less. That efficiency gap only widens at higher power levels, which is why linear supplies are now mostly reserved for specialty applications like sensitive analog audio circuits or laboratory instruments where extremely low electrical noise matters more than size or efficiency.
The next time you pick up a laptop charger and notice how light it is compared to the heavy “wall wart” adapters of decades past, that’s the direct result of switching at high frequencies. The physics hasn’t changed, but the ability to switch transistors cleanly at hundreds of thousands of cycles per second has made it possible to shrink what was once a brick-sized power supply into something that fits in your pocket.