LDO vs Buck Converter: Choosing a Power Rail
Haiben Editorial Team · 22 October 2026
The two ways to make a lower rail
Almost every board has to turn one voltage into another. A 12 V input has to feed a 3.3 V sensor; a 5 V rail has to feed a 1.8 V core; a lithium cell has to feed a 2.5 V reference. Two regulator topologies do most of that work, and engineers spend a surprising amount of time choosing between them: the linear regulator and the switching regulator. This article compares the low-dropout linear regulator against the step-down (buck) switching converter on the terms that actually decide a design, which is why the shorthand ldo vs buck converter comes up in almost every schematic review.
A linear regulator, including the low-dropout type, uses a pass transistor operating in its linear region. A control loop adjusts the transistor's gate or base voltage so the output stays at the target, and the excess input voltage is dropped across the transistor. A LDO regulator differs from an older standard linear regulator in one important respect: its dropout voltage can be small, so it can regulate with only a few hundred millivolts between input and output.
A buck converter chops the input with a switch, stores energy in an inductor, and delivers it to the output through a diode or synchronous rectifier, with a control loop setting the duty cycle. Because the inductor carries the current rather than a resistive element, the conversion is largely lossless in the ideal case. That structural difference, energy burned as heat versus energy moved by an inductor, is the root of every trade-off below.
Efficiency: the dropout equation vs switching losses
Start with the linear regulator because the maths is almost embarrassingly simple. To a first approximation, an LDO's efficiency equals the output voltage divided by the input voltage. The power it dissipates is the voltage across the pass transistor multiplied by the load current:
Pdiss = (Vin − Vout) × Iload
And η ≈ Vout / Vin.
Two consequences follow immediately. First, efficiency is fixed by the voltage ratio and nothing else: a 3.3 V output from a 5 V input is about 66% efficient at any load current, whether that current is 10 mA or 1 A. Second, the heat generated scales with load current, so the same conversion that is trivial at low current becomes a thermal problem at high current.
A buck converter does not obey that formula. Its losses come from conduction (I²R in the switch, inductor and traces) and from switching (gate charge, output capacitance and crossover losses that scale with frequency). The ratio of input to output voltage influences the duty cycle and therefore the loss balance, but it does not impose a hard ceiling the way the linear equation does. As a result, a buck's efficiency tends to stay comparatively flat across conversion ratio: a 12 V to 3.3 V buck and a 5 V to 3.3 V buck may both land in a similar band, while an LDO's efficiency would fall from roughly 66% to roughly 27% over that same change.
This is the first decision axis in any ldo vs buck converter comparison. Large step-down ratios and high load currents push you toward switching. Small ratios or small currents leave the linear option viable.
Noise and PSRR: where the LDO wins
Linear regulators are quiet because nothing in the pass path switches. The output ripple is set by the reference noise, the error amplifier, and any feedthrough from the input, and it is broadband and low-level. There are no switching harmonics at the output and no inductor to radiate. For an analog front end, a precision reference, a PLL, a low-phase-noise clock or an RF chain, this matters more than a few percent of efficiency.
The other linear advantage is power-supply rejection ratio. An LDO actively attenuates ripple already present on its input, and its PSRR is typically strong at low frequency and falls as frequency rises. In a signal chain where a switching regulator upstream creates a 500 kHz ripple, a downstream LDO can remove most of it. The same reasoning applies in reverse: a buck's output ripple is intrinsic to its operation and cannot be eliminated, only filtered or pushed to a higher frequency where it is easier to filter.
The practical pattern, which many mixed-signal boards follow, is a switching converter for the bulk conversion followed by a linear regulator on the sensitive rail. That is the hybrid answer described at the end, and it is the closest thing to a free lunch in power design. Note also that an LDO's noise performance is only as good as its input bypassing and layout; a poorly bypassed linear regulator can pass input noise straight through at high frequency, where its loop gain has fallen away.
Headroom, heat and board area
The LDO has one hard requirement: the input must stay above the output by at least the dropout voltage. A true low-dropout regulator might need only a few hundred millivolts even at useful currents, but that margin must hold at minimum input, over temperature, and at maximum load. If the input sags close to the output, the pass transistor leaves regulation and the output follows the input down. A buck converter has no equivalent constraint in the same sense; it can step down from any input above its own operating requirement, and it can even regulate through a transient where the input dips, provided the controller can maintain duty cycle.
Heat is the second constraint. Because dissipated power is (Vin − Vout) × Iload, an LDO in a package with a modest thermal resistance quickly runs out of headroom as current rises. A 1 A load across a 2 V difference means 2 W in the pass device, which will need careful copper area, thermal vias, or a larger package. That heat is not merely wasted energy; it is a reliability and enclosure-temperature issue.
Board area often favours the linear part. An LDO needs an input and output capacitor and little else. A buck needs an inductor, input and output capacitors, and attention to loop layout, and its switch node is an EMI source that has to be contained. At low currents and small step-down ratios the linear regulator can be the smaller and simpler solution overall. At higher currents the inductor becomes a fixed cost that is easy to justify.
| Property | LDO | Buck converter |
|---|---|---|
| Efficiency | ≈ Vout/Vin; fixed by the voltage ratio | Set by conduction and switching losses; comparatively flat across conversion ratio |
| Dissipated power | (Vin − Vout) × Iload | I²R conduction plus frequency-dependent switching losses |
| Output noise and ripple | Low, broadband, no switching harmonics | Switching ripple at the switching frequency and its harmonics |
| Input rejection | Actively attenuates input ripple (PSRR) | Ripple passes through by design of the conversion |
| Headroom requirement | Input must exceed output by the dropout voltage | No equivalent dropout constraint |
| External components | Input and output capacitors | Inductor, input and output capacitors, loop layout |
| EMI | Essentially none from the regulator itself | Switch node is a deliberate EMI source |
The linear voltage regulators category covers the wider family, including standard and low-dropout parts, for reference when the linear route is the right one.
Decision framework and the hybrid answer
A workable decision sequence for the ldo vs buck converter question looks like this. First, compute the linear option's efficiency and dissipation from the two formulas above at maximum load and worst-case input. Second, check dropout: does the minimum input stay above the output plus dropout at full current and over temperature? Third, ask whether the load is noise-sensitive. If it is analog, RF or precision, add a linear stage regardless of what supplies it. Fourth, check the thermal budget: can the package and board area remove the dissipated power without pushing junction temperature too high? Fifth, count the bill of materials and board area, including the inductor and its keep-out for EMI.
In practice the answers cluster. Low current with a small step-down ratio usually lands on an LDO. High current or a large step-down ratio lands on a buck. Noise-sensitive rails land on a buck followed by an LDO. The hybrid is worth spelling out because it resolves the apparent conflict rather than compromising on it: the buck does the efficient bulk conversion, the LDO cleans the rail and provides PSRR, and the LDO's dissipation stays small because its input is already close to its output.
Two engineering details improve whichever topology you choose. Keep the LDO's input and output capacitors close to the pins, and do not share a long trace with a switching node. On the buck side, keep the switch node small and the input capacitor loop tight; that does more for EMI than most filter additions. And remember that an LDO's dropout voltage is a specification at a stated current and temperature, not a constant, so leave margin rather than designing to the typical number.
For a next step, review the LDO regulators range alongside the buck converters category and match the part to the load current, input range and noise budget the design actually needs.