A resistor that regulates itself
A linear regulator holds an output voltage steady by placing a controllable resistive element — the pass transistor — in series between an unregulated input rail and the load, and continuously adjusting its resistance to absorb exactly the excess voltage. It is, functionally, an automatically-adjusting resistor: as the input voltage rises or the load draws more current, the pass element's effective resistance changes to keep the output exactly where a feedback loop wants it.
The unavoidable heat
Every watt not delivered to the load as regulated output power gets dissipated as heat in the pass transistor, with a simple and unforgiving formula: P_dissipated = (V_in − V_out) × I_load. Drop 5V at 500mA and the transistor burns 2.5W regardless of how efficient the load itself is — there's no way around it, because a linear regulator has no mechanism (no switching, no energy storage in an inductor) for converting excess voltage into anything but heat. This is the single biggest disadvantage against a switching (buck) converter, which can achieve the same voltage step with 85-95% efficiency instead of what can easily be under 50%.
P_pass = (V_in - V_out) * I_load // pure heat, no way around it efficiency = P_out / P_in = V_out / V_in // independent of I_load! example: 5V -> 3.3V at 1A P_pass = (5 - 3.3) * 1 = 1.7 W wasted efficiency = 3.3/5 = 66% (before quiescent current losses)
Dropout voltage: how "low" is low-dropout?
The pass transistor needs some minimum voltage across it to keep operating as a controllable element rather than a fully-saturated switch — this minimum input-output margin is the dropout voltage. Older regulator topologies (built around an NPN bipolar emitter follower) typically needed 2V or more of headroom; an LDO (low-dropout) regulator instead uses a PNP or PMOS pass device operated near saturation, cutting the required headroom to as little as 100-300 mV. Below that margin, the regulator loses regulation and V_out sags along with V_in — a critical constraint in battery-powered designs where the input voltage slowly droops as the battery discharges.
PSRR: rejecting noise on the input
Power supply rejection ratio (PSRR) measures how well the regulator's feedback loop attenuates ripple or noise arriving on the input rail before it reaches the output — expressed in dB, with a high PSRR meaning input noise is heavily suppressed. This is precisely why LDOs remain popular for powering sensitive analog and RF circuits even though switching converters are far more efficient: a switcher's own switching-frequency ripple would need extensive additional filtering to reach the same noise floor an LDO delivers with no active switching at all. It's common to see a switching converter handle the bulk voltage conversion efficiently, followed by a small LDO as a final "clean-up" stage feeding a noise-sensitive block.
The thermal ceiling on output current
Because dissipated power scales directly with load current, an LDO's maximum safe current isn't just a function of the transistor's current rating — it's bounded by how much heat the package and any heatsinking can carry away while keeping the junction temperature below its rated maximum (typically 125-150°C for silicon). A regulator perfectly capable of delivering 3A at a 200mV drop might only be safely rated for a few hundred milliamps at a 5V drop, because the thermal budget, not the silicon, becomes the limiting factor.
Frequently asked questions
Why use an inefficient linear regulator instead of a switching converter?
Because linear regulators produce essentially no switching noise, respond to load changes far faster, and are cheap and simple with no inductor or complex control loop. For low-noise analog and RF supply rails, or for small voltage drops where the power loss is tolerable, an LDO beats a switcher on nearly every axis except efficiency.
What does "dropout voltage" mean?
The minimum voltage the input must exceed the output by for the regulator to still hold regulation. Below that margin the pass transistor can no longer provide enough headroom to maintain the set output voltage, and V_out starts sagging along with V_in — this is what distinguishes a "low-dropout" (LDO) design from an older regulator needing 2V or more of headroom.
What sets the maximum current an LDO can safely deliver?
Thermal dissipation in the pass transistor, P = (V_in−V_out)·I_load, which must stay within what the package and any heatsinking can carry away without the junction temperature exceeding its rated maximum. A bigger input-output voltage difference or higher ambient temperature both reduce the safe current headroom well before the transistor's absolute current rating is reached.
Try it live
Everything above runs in your browser — open Linear Voltage Regulator and change the parameters while it is running. Nothing is installed, nothing is uploaded, the whole model lives in one tab.
▶ Open Linear Voltage Regulator simulation