Electronics Guide

Linear voltage regulators are fundamental building blocks of power supply design, providing stable, low-noise DC output voltages from higher and potentially variable input sources. Unlike switching regulators that use pulse-width modulation and energy storage elements, linear regulators operate by continuously dissipating excess power as heat through a series pass element, trading efficiency for simplicity, low noise, and excellent transient response.

From discrete series pass transistor circuits to highly integrated three-terminal regulators, linear regulation technology spans a wide range of complexity and capability. This article explores the principles, topologies, and practical considerations essential for designing and applying linear voltage regulators in electronic systems where clean, stable power is paramount.

Fundamentals of Linear Regulation

Linear voltage regulators maintain a constant output voltage by varying the resistance of a control element in series or parallel with the load. The regulator continuously compares the output voltage against a reference and adjusts the control element to compensate for changes in input voltage, load current, or temperature.

Basic Operating Principle

A linear regulator functions as a variable resistor controlled by a feedback loop. When the output voltage tends to rise (due to decreased load or increased input), the control element increases its resistance, dropping more voltage and maintaining the output. Conversely, when output tends to fall, the control element decreases its resistance. This continuous analog adjustment provides inherently clean output with no switching noise.

Key Performance Parameters

Several parameters characterize linear regulator performance:

• Line Regulation: The change in output voltage for a given change in input voltage, expressed as millivolts per volt or as a percentage. Good regulators achieve line regulation below 0.01% per volt
• Load Regulation: The change in output voltage from no load to full load, expressed as millivolts or as a percentage. Values below 0.1% indicate excellent load regulation
• Dropout Voltage: The minimum difference between input and output voltage required to maintain regulation. A conventional NPN-pass regulator such as the 7805 needs roughly 2 V to 3 V; a low-dropout part regulates on a few hundred millivolts, and the low-dropout label is applied by convention below about 1 V
• Quiescent Current: The current the regulator consumes for its own operation, exclusive of load current. It enters at the input pin and leaves at ground, which matters for both battery life and dissipation
• Output Noise: The AC component superimposed on the DC output, specified as microvolts RMS over a stated bandwidth. The bandwidth is part of the specification: 10 Hz to 100 kHz is the common one, and a figure quoted without it cannot be compared
• Ripple Rejection: The ability to attenuate AC components on the input, measured in decibels at a specified frequency. The frequency is equally part of the specification, because rejection falls as the control loop runs out of gain

Efficiency Considerations

Linear regulator efficiency is fundamentally limited by the input-output voltage differential. Neglecting the current the regulator consumes for itself:

Efficiency = V_out / V_in

For a 5 V output from a 12 V input the ratio is 5/12, so maximum efficiency is approximately 42%. The remaining 58% of the input power dissipates as heat in the pass element. Quiescent current lowers the figure further: negligibly at heavy load, but decisively at light load, where a regulator drawing 50 µA for itself while delivering 100 µA to the load spends a third of the input power on its own operation whatever the voltage differential. This dissipation limits linear regulators to applications where the voltage differential is small or the load current is modest. The simplicity, low noise, and fast transient response often justify the efficiency penalty.

Series Pass Regulator Design

Series pass regulators place the control element in series between the input supply and load, modulating the voltage drop to maintain constant output. This topology handles the majority of linear regulator applications and forms the basis for integrated regulator ICs.

Basic Series Pass Circuit

The fundamental series pass regulator comprises four essential blocks:

• Reference Voltage: A stable voltage source, typically a Zener diode, bandgap reference, or precision reference IC, against which the output is compared
• Error Amplifier: Compares a fraction of the output voltage to the reference and generates a control signal proportional to the error
• Pass Element: A transistor (BJT or MOSFET) that acts as a variable resistance, controlled by the error amplifier output
• Feedback Network: Usually a resistive voltage divider that scales the output voltage for comparison with the reference

NPN Pass Transistor Configurations

Traditional series regulators use an NPN transistor as the pass element, typically in common-collector (emitter follower) configuration. The error amplifier drives the base, and the emitter delivers current to the load. This arrangement offers several advantages:

• Natural current gain: The transistor beta multiplies the error amplifier drive current, allowing a modest amplifier to control substantial load currents
• Inherent stability: The emitter follower provides unity voltage gain with current gain, simplifying feedback compensation
• Wide bandwidth: NPN emitter followers exhibit excellent high-frequency response

The primary limitation is dropout voltage, and the reason lies in the base drive rather than in the pass transistor itself. An emitter follower holds its base roughly one base-emitter voltage above the output, and the stage that supplies that base current works from the input rail, so it needs headroom of its own. Dropout is therefore the base-emitter voltage plus the driver's saturation voltage plus any drop across a current-sense resistor, which lands at 2 V to 3 V in practical implementations. The pass transistor's own collector-emitter saturation voltage, a few tenths of a volt, is the smallest term in that sum, which is why specifying a better NPN does not lower the dropout of this topology. Only rearranging the pass element does, which is the subject of the low-dropout section below.

Darlington Pass Configurations

For higher current applications where the error amplifier cannot supply sufficient base current, a Darlington pair provides increased current gain. The two-transistor combination multiplies the effective current gain (approximately the product of the two individual betas), reducing drive requirements dramatically. However, the Darlington configuration adds approximately 0.6 V to the dropout voltage and can exhibit thermal runaway concerns if not properly compensated.

PNP and P-Channel Pass Elements

Using a PNP transistor or a P-channel MOSFET as the pass element reverses the connection: the input feeds the emitter or source and the load hangs on the collector or drain. The error amplifier can then drive the device hard toward saturation, and the output approaches the input to within the saturation voltage or the channel resistance drop rather than stopping a base-emitter voltage short of it. That is the whole of the low-dropout idea, and the section on LDO regulators below develops what it costs.

Shunt Regulator Configurations

Shunt regulators place the control element in parallel with the load rather than in series. The pass element diverts excess current away from the load to maintain constant voltage. While less common than series regulators for primary power supplies, shunt configurations excel in specific applications.

Basic Shunt Regulator Operation

In a shunt regulator, a series resistor limits current from the input supply. The shunt element (typically a transistor or Zener diode) maintains the output voltage by conducting whatever current the load does not require. When load current increases, shunt current decreases proportionally. When load current decreases, shunt current increases to absorb the excess.

Zener Diode Shunt Regulators

The simplest shunt regulator uses a Zener diode as both the reference and control element. The Zener maintains approximately constant voltage across its terminals as current varies within its operating range. Design considerations include:

• Series resistor selection: Must provide adequate Zener current at minimum load while not exceeding maximum Zener current at no load
• Power dissipation: The Zener dissipates maximum power at minimum load current
• Dynamic impedance: Lower Zener impedance provides better regulation but requires higher current
• Temperature coefficient: Zener voltage drifts with temperature; the coefficient is smallest for devices near 5 V to 6 V, and temperature-compensated references cancel it by pairing an avalanche diode with a forward-biased junction near 6.2 V

Active Shunt Regulators

Replacing the Zener with an active circuit improves regulation dramatically, because an error amplifier and a pass transistor present a far lower dynamic impedance than a diode. The TL431 adjustable precision shunt regulator exemplifies the approach: an internal 2.5 V reference and error amplifier drive an open-collector output, so a two-resistor divider programs the cathode voltage anywhere from the reference up to 36 V. Active shunt regulators find extensive use in:

• Reference voltage generation: Creating precision voltages for ADCs, DACs, and measurement circuits
• Feedback networks: Providing the reference and error amplifier functions in switching power supply feedback loops
• Voltage clamping: Protecting circuits from overvoltage conditions
• LED drivers: Maintaining constant current through LED strings

Shunt Regulator Limitations

Shunt regulators suffer from poor efficiency, especially at light loads where most input power dissipates in the shunt element. They also require careful matching between the series resistance and load current range. These limitations restrict shunt regulators to low-power applications or situations where their specific characteristics, such as simplicity, bidirectional operation, or overvoltage clamping, provide unique benefits.

Three-Terminal Regulator Applications

Three-terminal fixed-voltage regulators revolutionized power supply design by integrating reference, error amplifier, pass transistor, and protection circuits into a single package. The classic 78xx (positive) and 79xx (negative) series, introduced in the 1970s, remain ubiquitous despite their age, while modern variants offer improved performance.

Standard 78xx/79xx Regulators

The 78xx series provides fixed positive output voltages of 5 V, 6 V, 8 V, 9 V, 10 V, 12 V, 15 V, 18 V, and 24 V, with a current capability of 1 A or more depending on the package. Key characteristics include:

• Input, output, and ground terminals: The connection requires nothing beyond bypass capacitors
• Dropout voltage: Texas Instruments specifies 2 V for the LM7805 at 1 A and a junction temperature of 25 degrees Celsius, so the input must stand at least that far above the rated output at full load, and further at temperature extremes
• Output tolerance by grade: The ordinary 7805 grades hold 4.8 V to 5.2 V, or ±4%, at 25 degrees Celsius. An A grade such as the LM340A-5.0 halves that to 4.9 V to 5.1 V. The grade suffix, not the family, decides whether a part meets a tolerance requirement
• Line and load regulation: A few tens of millivolts across the rated input and load ranges, on the order of 0.1% to 1% of a 5 V output
• Built-in protection: Thermal shutdown, current limiting, and safe operating area protection, present in every grade

The 79xx series provides corresponding negative voltages with similar specifications, though internal topological differences result in slightly different performance characteristics.

Capacitor Requirements

Proper capacitor selection ensures stability and performance:

• Input capacitor: Manufacturers recommend 0.22 µF to 0.33 µF at the regulator input, placed physically close to the device. It is strictly necessary only when the regulator sits far from the supply filter capacitor, where lead inductance raises the local input impedance enough to provoke oscillation, but it costs little to fit in every case
• Output capacitor: The 78xx family does not need an output capacitor to remain stable, a property Texas Instruments lists among the LM340 family features, because the NPN emitter-follower output is a low-impedance node. A 0.1 µF capacitor lowers the high-frequency output impedance and 10 µF or more improves transient response, but neither is a stability condition. Do not carry this permission over to a low-dropout regulator, where the output capacitor forms part of the compensation
• Capacitor type: Ceramic capacitors work well for the small values; aluminum electrolytics suit the larger ones, with attention to equivalent series resistance and to how far that resistance rises at low temperature

Enhanced Performance Configurations

Simple circuit additions extend three-terminal regulator capabilities:

• Output voltage adjustment: Adding a resistor between the ground pin and ground raises the output voltage by lifting the reference point. The adjustment resistor forms a divider with the regulator's quiescent current path
• Current boosting: An external pass transistor increases current capability beyond the regulator's rating while the IC provides regulation and protection
• Foldback current limiting: External resistors can modify the current limit characteristic to reduce pass element dissipation during overload
• Tracking regulators: Summing positive and negative outputs creates tracking dual supplies for operational amplifier circuits

Adjustable Voltage Regulators

Adjustable regulators offer flexibility to generate any output voltage within a specified range using external resistors to set the output. The LM317 (positive) and LM337 (negative) represent the most widely used adjustable linear regulators, though numerous alternatives exist with varying current capabilities and features.

LM317/LM337 Operating Principles

The LM317 maintains a constant 1.25 V reference between its output and adjustment (ADJ) terminals. A resistive divider between output and ground, with its midpoint connected to ADJ, sets the output voltage according to:

V_out = V_ref(1 + R₂/R₁) + I_adjR₂

where V_ref is 1.25 V, R₁ connects from output to ADJ, R₂ connects from ADJ to ground, and I_adj is the adjustment pin current. Texas Instruments specifies I_adj as 50 µA typical and 100 µA maximum over the full operating temperature range, and designs the part to hold it constant against line and load. The second term is therefore small and predictable rather than negligible in general. With R₁ = 240 Ω it contributes 12 mV at a 2.5 V output and 50 mV at 6.5 V, and it grows with R₂ to roughly 0.9% of the output at the top of the adjustment range. Where that error is acceptable, the expression simplifies to:

V_out = (1.25 V)(1 + R₂/R₁)

R₁ carries a fixed program current of 1.25 V divided by its resistance, and that current is also the only load the regulator sees when nothing else draws from the output. This is why practical designs use R₁ values between 120 Ω and 240 Ω, which deliver 10.4 mA and 5.2 mA respectively. The LM317 does not regulate below a minimum load current, specified as 3.5 mA typical and 10 mA maximum at a 40 V input-to-output differential; left unloaded, its output drifts upward. The 240 Ω value covers the typical figure, while a design that must guarantee regulation against the 10 mA worst case at a large differential needs the 120 Ω value or a separate bleed resistor. This minimum-load requirement is the specification that most often surprises a first-time LM317 user.

Output Voltage Range and Accuracy

The LM317 output voltage adjusts from 1.25 V, with R₂ at zero, up to 37 V. The ceiling is the input-to-output differential the part tolerates, 40 V absolute maximum, rather than anything in the divider. Texas Instruments specifies the reference between 1.2 V and 1.3 V across the rated load range, a spread of ±4% about the nominal 1.25 V, and that spread dominates the output voltage uncertainty because the divider ratio multiplies it. Resistor tolerance adds to it in proportion. Precision applications require individual calibration, a tighter-tolerance variant, or a separate reference and error amplifier in place of the internal ones.

Protection and Bypassing

Recommended protection elements include:

• Input protection diode: A diode from output to input, cathode at the input, conducts if the input collapses while the output capacitor remains charged, shunting that discharge around the regulator die. Texas Instruments recommends it whenever an output capacitor is fitted
• ADJ pin bypass capacitor: A 10 µF capacitor from ADJ to ground filters reference noise and raises ripple rejection by roughly 15 dB, from about 65 dB to about 80 dB typical, measured at 120 Hz with a 10 V output
• ADJ protection diode: A second diode from the ADJ pin to the output, cathode at the output, gives the ADJ bypass capacitor a low-impedance discharge path if the output is pulled down or shorted, so that the capacitor does not discharge back through the internal circuitry of the adjustment pin. It belongs in the circuit whenever the ADJ bypass capacitor does. Note that this diode runs between ADJ and the output, not between ground and the output

Current Regulator Applications

The LM317 also functions as an adjustable current regulator. A single sense resistor between the output and ADJ terminals carries the whole load current, and the load returns from ADJ to ground, so the regulator holds its 1.25 V reference across the sense resistor:

I_out = 1.25 V / R_sense

A 12 Ω resistor therefore sets 1.25/12, or 104 mA; 12.4 Ω, the nearest one-percent value, gives 101 mA. The adjustment pin current adds to the total, but at 50 µA it stays well under a tenth of a percent here. The resistor must be rated for the power it dissipates, which is the square of 1.25 V divided by R_sense, or 130 mW at 12 Ω. This configuration suits LED drivers, battery chargers, and constant-current loads, provided the load leaves the LM317 its 3 V of headroom above the reference.

Current Limiting and Foldback Protection

Current limiting protects both the regulator pass element and the load from damage during fault conditions. Various current limiting approaches offer different trade-offs between simplicity, protection effectiveness, and power dissipation.

Simple Current Limiting

Basic current limiting uses a sense resistor in series with the load to monitor current flow. When the voltage across this resistor reaches roughly 0.6 V, it turns on a transistor that diverts drive away from the pass element and prevents any further increase. The current limit threshold is:

I_limit = 0.6 V / R_sense

For a 0.6 Ω sense resistor, the limit is 1 A. While simple, this approach causes maximum power dissipation in the pass element during a short circuit, when the pass element sustains the full input voltage at the limited current. A regulator running from 12 V and limiting at 1 A dissipates 12 W into a dead short, several times its worst-case dissipation in normal service, and that is the condition the heat sink must survive or the thermal shutdown must interrupt.

Foldback Current Limiting

Foldback limiting reduces the current limit as the output voltage decreases, which sharply reduces pass element dissipation during a short. As the output approaches zero volts, the limit folds back to a fraction of its normal value, commonly 20% to 30%. In the example above, folding back to 25% turns 12 W of short-circuit dissipation into 3 W, which a modest heat sink handles indefinitely. The characteristic creates a safe operating envelope that keeps the pass element within its power rating even during a sustained fault.

Foldback limiting uses a resistive divider from the output voltage to modify the sense transistor bias point. As output voltage drops, the transistor turns on at progressively lower load currents. While excellent for protection, foldback limiting can prevent startup into capacitive loads that draw high initial charging current.

Constant Power Limiting

Some sophisticated regulators implement constant power limiting, which reduces current inversely with the input-output differential to maintain constant pass element dissipation regardless of operating point. This provides optimal utilization of the pass transistor's safe operating area across all conditions.

Safe Operating Area Considerations

Pass transistor destruction can result from:

• Excessive junction temperature: From sustained high power dissipation
• Secondary breakdown: A bipolar transistor failure mode occurring at high voltage and moderate current where localized heating creates current-hogging hot spots
• Excessive current: Beyond the maximum current rating regardless of voltage

Effective protection must address all three failure mechanisms by limiting current, implementing thermal shutdown, and potentially incorporating secondary breakdown protection for high-voltage applications.

Thermal Protection Circuits

Thermal protection prevents regulator destruction from overheating, whether caused by excessive load current, inadequate heat sinking, high ambient temperature, or sustained fault conditions. All modern integrated regulators include thermal shutdown, and discrete designs should incorporate similar protection.

Thermal Shutdown Operation

Thermal shutdown circuits use a temperature-sensing element, usually a forward-biased junction whose voltage falls at roughly 2 millivolts per degree, placed on the die beside the pass transistor. When the junction temperature exceeds a threshold, commonly between 150 degrees Celsius and 175 degrees Celsius, the shutdown circuit reduces or removes drive to the pass element and lets the device cool.

Most implementations include hysteresis to prevent rapid cycling between the on and off states. The regulator shuts down at an upper threshold and does not resume until the temperature falls to a lower one, typically 10 degrees to 20 degrees below the shutdown point. The result under a sustained overload is a regulator that pulses its output on and off at a few hertz rather than one that fails, which is a useful diagnostic signature: an output that collapses and recovers on a slow, regular cycle points at dissipation rather than at the load.

Discrete Thermal Protection

For discrete regulator designs, thermal protection can be implemented using:

• Thermistors: NTC thermistors mounted on the heat sink or pass transistor case sense temperature. A voltage divider with the thermistor drives a comparator that controls regulator operation
• Thermal cutoff devices: One-shot thermal fuses or resettable thermal switches provide backup protection
• Temperature-sensing ICs: Devices like the LM35 or DS18B20 provide accurate temperature measurement for sophisticated protection schemes
• Integrated thermal shutdown ICs: Purpose-built protection ICs combine temperature sensing with latching or non-latching shutdown logic

Heat Sink Design Considerations

Proper thermal management prevents thermal protection from activating during normal operation:

• Power dissipation calculation: P_d = (V_in − V_out)I_load + V_inI_q. The second term is easy to omit and easy to misplace. The quiescent, or ground-pin, current enters at the input pin and leaves at ground rather than passing through the load, so the whole input voltage stands across it, not the input-to-output differential. At heavy load the term is a rounding error; at light load in a modern low-quiescent part it dominates, because the differential term shrinks with the load while the quiescent term does not
• Thermal resistance budget: The total resistance from junction to ambient must not exceed (T_J(max) − T_A)/P_d at the worst-case ambient temperature, and that budget is spent along the whole path, not on the heat sink alone
• Heat sink selection: θ_SA = (T_J(max) − T_A)/P_d − θ_JC − θ_CS, where θ_JC is junction-to-case and θ_CS is the case-to-sink resistance of the mounting interface. An insulating washer is not free: it commonly adds one to two degrees Celsius per watt, a large share of the budget in a tight design, and a design that forgets it specifies a sink it cannot afford
• Data-sheet junction-to-ambient figures: A junction-to-ambient resistance quoted in a thermal table is measured on a standardized JEDEC test board with a defined copper area, not on the board being designed. Texas Instruments lists 23.9 degrees Celsius per watt for the TO-220 LM340 under those conditions, and a part soldered to a small copper pour in still air performs far worse. Build the budget from the individual resistances and the actual copper, or measure the finished board
• Airflow effects: Forced airflow substantially reduces heat sink thermal resistance; natural convection requires larger heat sinks

Low-Dropout (LDO) Regulators

Low-dropout regulators minimize the voltage difference required between input and output to maintain regulation, enabling efficient operation from low-headroom sources such as batteries approaching discharge or post-regulator applications where minimizing power loss is critical.

LDO Architecture

LDO regulators achieve low dropout by using a PNP bipolar transistor or a P-channel MOSFET as the pass element, with the input at the emitter or source and the output at the collector or drain. Driving the device toward saturation removes the base-drive headroom that sets the dropout of an NPN follower. That rearrangement is not free, however, and the four common pass elements differ on three counts that matter in practice: how little dropout they achieve, how much current they consume for their own drive, and how they constrain the output capacitor.

The table explains why the PNP dominated the first generation of low-dropout parts and why the P-channel MOSFET displaced it. A PNP LDO delivering 500 mA with a current gain of 50 sends 10 mA to ground as base current, wasting more power than the load draws in a standby system, and that figure worsens exactly when the part is closest to dropout. The MOSFET gate draws nothing in the static condition, so quiescent current stays in the microamps across the whole load range. LDO Regulator Fundamentals develops these choices and their consequences in detail.

Stability Considerations

LDO regulators present challenging stability requirements due to the high-impedance output node and the PNP/PMOS pass element's current-source characteristic. Key stability factors include:

• The equivalent series resistance window: Many LDOs, and nearly all the older PNP-pass parts, specify a band of acceptable equivalent series resistance rather than a minimum capacitance, because the zero that this resistance forms with the output capacitor supplies the phase margin. Data sheets plot the band as a stability region, informally the ESR tunnel. Fall below its floor and the regulator oscillates, at a frequency set by the loop and often in the tens or hundreds of kilohertz; rise above its ceiling and transient response degrades as the capacitor stops behaving like a capacitor
• The ceramic trap: A modern X5R or X7R ceramic has an equivalent series resistance of a few milliohms, far below the floor of an older LDO stability region. Substituting one for the tantalum the data sheet assumed is the most common way to turn a working linear supply into an oscillator, and the symptom is deceptive: a rail that reads clean on a handheld meter while carrying tens of millivolts of ringing on an oscilloscope. Parts designed after ceramics became the default are compensated for them and specify a capacitance range instead. The data sheet states which kind a part is, and the distinction does not transfer between families or even between members of one family
• Capacitance under bias: A class II ceramic loses capacitance as DC voltage is applied across it, and a 10 µF X5R in a small case can deliver half its rated value or less at its working voltage. A design that satisfies a minimum-capacitance requirement on the bill of materials can violate it on the board
• Minimum load current: Some LDOs require a minimum load for stability, because the output pole moves with load current. A bleed resistor solves it, at the cost of the standby current the LDO was chosen to save

LDO Performance Trade-offs

LDO design involves balancing competing requirements:

• Quiescent current vs. transient response: Lower quiescent current reduces efficiency loss at light loads but slows the error amplifier, degrading transient response
• Dropout vs. drive: A larger pass device reduces dropout, but the cost differs by device type. A bigger bipolar draws proportionally more base current, which goes straight to ground; a bigger MOSFET draws no more static gate current but presents more gate capacitance, which slows the loop and worsens transient recovery
• Noise vs. complexity: Achieving very low output noise requires sophisticated error amplifier design and careful reference implementation
• PSRR vs. bandwidth: Power supply rejection ratio typically degrades at higher frequencies as the control loop gain decreases

LDO Applications

LDO regulators excel in applications including:

• Battery-powered devices: Maximizing usable battery life by operating to lower cell voltages
• Post-switching regulator filtering: Creating clean supply rails from switching converter outputs with minimal additional voltage drop
• RF and analog circuits: Providing ultra-low-noise supplies for sensitive amplifiers, PLLs, and data converters
• Automotive applications: Operating from cold-cranking voltage dips while maintaining regulation

Negative Voltage Regulators

Negative voltage regulators provide regulated negative output voltages referenced to ground, essential for bipolar analog circuits, operational amplifiers, and other applications requiring both positive and negative supplies.

Negative Regulator Topologies

Negative regulators use circuit topologies complementary to their positive counterparts:

• Complementary low-dropout pass element: A negative regulator mirrors the positive case. Where a positive LDO uses a PNP or P-channel device so that the output can approach the input, a negative LDO uses the complementary NPN or N-channel arrangement, and low dropout follows for the same reason. Negative parts reach the performance of their positive counterparts: Analog Devices specifies the LT3094 at 235 mV of dropout, 0.8 µV RMS of output noise from 10 Hz to 100 kHz, and 74 dB of power supply rejection at 1 MHz, across a −1.8 V to −20 V input range at up to 500 mA
• PNP-based designs: Mirror the NPN positive regulators, with higher dropout voltage but simpler drive requirements
• Integrated solutions: The 79xx series and LM337 provide negative equivalents to positive three-terminal and adjustable regulators

79xx Series Characteristics

The 79xx series provides fixed negative voltages with specifications similar to, but not identical to, their positive counterparts:

• Pinout differences: Pin assignments differ from 78xx; ground and input are swapped. Care must be taken when replacing positive with negative regulators in a circuit
• Minimum load requirements: Many 79xx regulators require a few milliamps of load current before they regulate properly, where the positive parts do not. Check the figure in the data sheet and fit a bleed resistor if the load can fall below it
• Different transient response: Internal topology differences result in different dynamic behavior compared to positive regulators
• Capacitor requirements do not carry over: The 78xx family tolerates an absent output capacitor, but negative three-terminal parts commonly specify a required output capacitor and sometimes a capacitor type, so read the negative part's own data sheet rather than assuming symmetry with its positive counterpart

Design Considerations for Negative Rails

Designing negative supply circuits requires attention to:

• Ground reference: All voltages are negative with respect to ground; measurement equipment must be properly configured
• Capacitor polarity: Electrolytic capacitor polarity reverses from positive rail designs; the positive terminal connects to ground
• Protection diodes: Diode orientation reverses compared to positive regulator circuits
• Current direction: Current flows from ground toward the negative supply; meters may show negative current readings

Dual-Tracking Supplies

Dual-tracking power supplies maintain matched positive and negative output voltages that track together during startup, shutdown, and load variations. This tracking prevents potentially destructive latch-up conditions in CMOS circuits and ensures proper biasing of operational amplifiers and other symmetric circuits.

Tracking Requirements

Different applications have varying tracking requirements:

• Startup/shutdown tracking: The positive and negative rails should rise and fall together to prevent either rail from being present alone, which can forward-bias protection diodes or cause latch-up
• Static matching: The magnitude of positive and negative voltages should be equal under steady-state conditions
• Dynamic tracking: Transient load variations on one rail should not excessively affect the other
• Tracking accuracy: Specifications range from 5% for general analog circuits to 0.1% for precision applications

Implementation Approaches

Several techniques achieve tracking between positive and negative regulators:

• Master-slave configuration: One regulator serves as master with a fixed reference. The other regulator (slave) uses the master output as its reference, tracking any variations in the master
• Common reference: Both regulators derive their references from a common source, with matched gain in each feedback path
• Integrated dual regulators: Purpose-built ICs incorporate matched reference and error amplifiers for inherent tracking
• Active tracking circuits: Separate tracking circuitry monitors both outputs and adjusts one to match the other

Practical Tracking Circuit

A practical master-slave tracker uses the positive regulator as the master with standard resistor-set output voltage. The negative regulator receives a reference voltage derived from the positive output through an inverting amplifier with unity gain magnitude. This configuration ensures the negative rail matches the positive rail magnitude.

Startup tracking requires additional circuitry to ensure neither rail leads or lags significantly. Common approaches include:

• Soft-start capacitors: Matched RC time constants on both regulators' reference inputs provide similar ramp rates
• Sequential enable: Logic circuitry enables the second regulator only after the first reaches regulation
• Active sequencing ICs: Purpose-built sequencing controllers manage supply rail ordering

Noise and PSRR Optimization

Linear regulators' primary advantage over switching types lies in their inherently low output noise and high power supply rejection ratio (PSRR). Understanding and optimizing these characteristics is essential for noise-sensitive applications.

Noise Sources in Linear Regulators

Output noise originates from several sources within the regulator:

• Reference noise: The voltage reference contributes wideband noise that appears at the output multiplied by the regulator's gain. Bandgap references typically exhibit lower noise than Zener-based references
• Error amplifier noise: Input-referred noise from the error amplifier is amplified by the closed-loop gain
• Pass element noise: The pass transistor contributes 1/f noise and thermal noise, though these are typically attenuated by the feedback loop
• Resistor noise: Feedback network resistors generate thermal noise proportional to their resistance and temperature

Techniques for Low-Noise Operation

Minimizing output noise involves:

• Reference filtering: Bypassing the reference or adjustment pin with a capacitor filters reference noise
• Low-noise reference selection: Buried Zener and precision bandgap references offer lower noise than standard references
• Low feedback resistor values: Reducing feedback network resistance decreases thermal noise contribution
• Output filtering: Additional LC or RC filtering on the output attenuates remaining noise
• Selecting low-noise LDOs: Some LDOs are specifically designed for low noise, with optimized internal circuitry

Power Supply Rejection

PSRR quantifies the regulator's ability to attenuate input voltage variations, including ripple from rectified AC, switching converter noise, and other supply disturbances. PSRR is expressed in decibels:

PSRR (dB) = 20 log₁₀(ΔV_in / ΔV_out)

The input perturbation sits above the output perturbation inside the logarithm, so a well-rejecting regulator yields a large positive number. A PSRR figure quoted without a frequency is close to meaningless, and that omission is the standard trap in supply specifications. Rejection follows the loop gain: it is highest at DC, holds while the error amplifier still has gain in reserve, then falls at roughly the rate the loop gain falls once the loop bandwidth is passed, until above the unity-gain frequency the regulator rejects nothing at all and the output capacitor and any series impedance do the whole job. Texas Instruments specifies the LM340A at 68 dB minimum and 80 dB typical at 120 Hz, the ripple frequency of a full-wave rectifier on 60 Hz mains, and that figure says nothing about the same part at 1 MHz, which is where a switching converter places its ripple. Parts intended to post-regulate a switching converter therefore quote rejection at megahertz frequencies: Analog Devices specifies 74 dB at 1 MHz for the LT3094. Matching the rejection figure to the frequency of the disturbance actually present is the whole of the selection problem.

Improving PSRR

Strategies for enhancing power supply rejection include:

• Pre-regulation: An RC filter or pre-regulator stage ahead of the main regulator provides additional input filtering
• Cascaded regulation: Two regulators in series multiply their individual rejection ratios, so in decibels the figures add. The gain is real but bounded in practice, because coupling through shared ground and through the board sets a floor that no amount of cascading gets below, and because each stage costs its own headroom and dissipation
• Feed-forward compensation: Some regulators inject a compensating signal that cancels input variations, improving high-frequency PSRR
• Higher bandwidth error amplifiers: Extending the control loop bandwidth maintains PSRR to higher frequencies

Practical Design Example

A complete design example illustrates the application of linear regulator principles. Consider a 5 V, 500 mA regulated supply drawn from a 9 V to 12 V input, powering sensitive analog circuitry.

Design Requirements

• Output voltage: 5.0 V ±2%
• Output current: 0 to 500 mA
• Input voltage: 9 V to 12 V DC
• Output noise: less than 50 µV RMS, 10 Hz to 100 kHz
• Load regulation: better than 0.5%
• Thermal protection required

Component Selection

The headroom available at the low end of the input range is 4 V, far more than a low-dropout part needs, so an LDO would add cost and a stricter output-capacitor constraint without buying anything in return. A three-terminal fixed regulator in the 7805 family is the right choice. The grade, however, is decided by the tolerance requirement rather than by the family:

• Regulator: LM340A-5.0 in a TO-220 package. Texas Instruments specifies the A grade at 4.9 V to 5.1 V at a junction temperature of 25 degrees Celsius, which is the ±2% the requirement asks for. The plain LM7805 and LM340 grades are specified at 4.8 V to 5.2 V, widening to 4.75 V to 5.25 V across the rated dissipation; they meet every other requirement here but miss the tolerance by a factor of two, so the ordinary LM7805CT is not a valid substitution in this design. Both grades carry the same internal current limiting, thermal shutdown, and safe-area protection
• Input capacitor: 10 µF aluminum electrolytic for bulk energy storage, with 100 nF of ceramic close to the input pin for high-frequency bypassing
• Output capacitor: 10 µF for transient response, with 100 nF of ceramic to hold the high-frequency output impedance down. Ceramic is acceptable for both, because this family does not depend on the output capacitor for stability; the same substitution in an older LDO would not be safe

Thermal Design

Maximum dissipation occurs at maximum input voltage and full load, and both terms of the dissipation expression belong in it. The LM340A draws at most 6 mA of quiescent current:

P_d = (12 V − 5 V)(0.5 A) + (12 V)(6 mA) = 3.5 W + 0.07 W = 3.57 W

The quiescent term adds two percent here, which the heat sink margin absorbs. It is kept in the expression because the same arithmetic applied to a battery-powered LDO at a hundred microamps of load gives the opposite answer, with the quiescent term carrying most of the dissipation. Dropping it, as worked examples often do, teaches the wrong habit.

With a maximum junction temperature of 125 degrees Celsius and a 40 degrees Celsius ambient, the total thermal resistance from junction to ambient must not exceed:

(125 − 40) / 3.57 = 23.8 degrees Celsius per watt

That budget covers the whole path. Texas Instruments gives 1.7 degrees Celsius per watt from junction to the case bottom for this TO-220 package, the tab being the thermal path, and a greased mica or elastomer insulating washer adds roughly 1.5 degrees Celsius per watt. What remains for the heat sink is:

23.8 − 1.7 − 1.5 = 20.6 degrees Celsius per watt

A small heat sink rated below 20 degrees Celsius per watt in natural convection satisfies this with a little margin in hand. Leaving the insulating washer out of the budget, a common slip, would have suggested that 22 degrees Celsius per watt suffices and would have spent the margin before construction began. The data-sheet junction-to-ambient figure of 23.9 degrees Celsius per watt for this package is likewise no substitute for the calculation: it is measured on a standardized test board, and taking it at face value places the junction at 40 degrees Celsius plus 3.57 watts times 23.9 degrees Celsius per watt, which is 125 degrees Celsius, exactly at the limit with no margin and no heat sink at all. In an enclosed space with limited airflow, a larger sink is necessary.

Noise Reduction

Texas Instruments specifies the LM340A at 40 µV RMS of output noise from 10 Hz to 100 kHz, which is the band the requirement uses. The part therefore meets the 50 µV limit, but only as a typical figure with no guaranteed maximum behind it, so the design has margin on paper and none that can be relied on. Additional measures are warranted:

• LC filter: A 10 µH inductor followed by a 100 µF capacitor places a low-pass corner at f = 1/(2π√(LC)), which works out to 5.0 kHz. Note carefully what this does and does not achieve. It attenuates the decade above 10 kHz well and leaves the 10 Hz to 5 kHz portion of the specified band untouched, and that lower portion is where a bandgap-referenced regulator concentrates its noise. An LC filter improves a wideband measurement far less than its corner frequency suggests
• A quieter regulator: Where the low-frequency noise must genuinely come down, the answer is a low-noise LDO post-regulator rather than a filter. Parts designed for the purpose reach below one microvolt RMS over this same band, roughly fifty times better than a 7805, at the cost of the headroom the LDO needs and the extra part
• Ferrite bead: A ferrite bead in series with the output adds impedance at the frequencies where the filter inductor has begun to behave capacitively
• Attention to layout: Keep high-current paths away from sensitive signal traces; use ground planes and decouple at the load rather than only at the regulator

Conclusion

Linear voltage regulators remain indispensable in modern electronics despite the proliferation of switching converters. Their simplicity, low noise, excellent transient response, and freedom from electromagnetic interference make them ideal for powering noise-sensitive analog circuits, serving as post-regulators after switching converters, and providing clean power in applications where efficiency is less critical than power quality.

Understanding the fundamentals of series and shunt regulation, three-terminal and adjustable regulators, current limiting and thermal protection, LDO operation, and noise optimization enables engineers to select appropriate devices and design effective linear power supplies. While switching regulators dominate applications where efficiency is paramount, linear regulators continue to excel in their niche, providing the clean, stable power that sensitive circuits demand.

As integrated circuit technology advances, linear regulators continue to evolve with lower dropout voltages, reduced quiescent currents, improved PSRR, and sophisticated protection features, ensuring their continued relevance in the ever-expanding landscape of electronic power conversion.

Related Topics