Electronics Guide

What Isolation Provides

An isolation barrier blocks the flow of direct and low-frequency current between two circuits while allowing signals or power to cross by another means. The two sides may therefore sit at widely different potentials without a damaging current flowing between them. If a fault places a hazardous voltage on one side, the barrier prevents that voltage from appearing on the other side, protecting both users and equipment. The quality of an isolation barrier is characterized by the voltage it can withstand, both continuously and as a transient, and by its ability to maintain separation over the product's lifetime despite environmental stress.

Isolation also defines a separation of reference potentials. Each side of the barrier has its own ground or common, and the two grounds need not be related. This independence is what allows isolation to break ground loops and to permit a low-voltage controller to interact safely with high-voltage power circuitry. The signal or power that crosses the barrier does so through a coupling mechanism designed to transfer information or energy without compromising the separation.

Functional Versus Safety Isolation

Functional isolation provides the separation needed for a circuit to operate correctly but is not relied upon for protection against electric shock. It might break a ground loop or shift signal levels between subsystems that are not at hazardous potentials. Safety isolation, by contrast, is part of the protection that keeps users safe from electric shock and must meet stringent requirements for withstand voltage, creepage, clearance, and reliability. A single physical barrier may serve only a functional role in one design and a critical safety role in another, but the requirements imposed on it differ greatly between the two cases.

This distinction governs how an isolation barrier is specified and verified. A barrier serving a safety function must satisfy the insulation coordination requirements of the applicable safety standard, including type testing of its withstand capability and adequate spacing to survive pollution and overvoltage. A purely functional barrier need only meet the performance the circuit requires. Recognizing which role a barrier plays is the first step in selecting an appropriate isolation device.

Common-Mode Transients and Isolation

In many systems the two isolated sides experience rapidly changing voltage differences, especially in power conversion where a switching node swings by hundreds of volts in nanoseconds. An isolation device must transmit its intended signal faithfully while rejecting these common-mode transients, a capability quantified as common-mode transient immunity and expressed in kilovolts per microsecond. Insufficient immunity allows fast voltage swings across the barrier to corrupt the transmitted data or inject noise, so high common-mode transient immunity is a key figure of merit for isolators used in switching power and motor-drive applications.

The mechanism is displacement current. Every practical barrier has some capacitance between its two sides, on the order of a picofarad for an integrated isolator or an optocoupler and rising to several picofarads or more for a transformer, where the windings face each other over a large area. A voltage slewing at several tens of kilovolts per microsecond drives a current through that capacitance in proportion to the slew rate, and that current must not be able to flip the receiver's decision threshold. Designers therefore minimize barrier capacitance, use differential or balanced receivers, and add redundancy in the transmitted code. Wide-bandgap devices have raised the stakes: silicon carbide and gallium nitride transistors switch far faster than the silicon insulated-gate bipolar transistors they replace, so isolators intended for those power stages are specified with correspondingly higher immunity. Modern digital isolators commonly specify 100 kV/µs or more, while ordinary phototransistor-output optocouplers are usually an order of magnitude lower.

Optocouplers

An optocoupler, also called an optoisolator, transfers a signal across an isolation barrier using light. A light-emitting diode on the input side converts an electrical signal into light, which crosses a transparent insulating gap and strikes a photodetector, typically a phototransistor or photodiode, on the output side that converts the light back into an electrical signal. Because the only coupling is optical, no conductive path crosses the barrier, and the input and output may operate at very different potentials. Optocouplers have long been the standard means of isolating digital signals, feedback in power supplies, and gate-drive paths.

Optocouplers are valued for their proven safety pedigree, wide availability, and high withstand voltages. Their limitations include relatively modest speed, a current-transfer ratio that varies with temperature and ages over time, and significant input current to drive the light-emitting diode. Current-transfer ratio, the ratio of output current to input current, is the central parameter: a general-purpose phototransistor optocoupler is typically binned over a wide range, and the guaranteed minimum may be only a fraction of the typical value, so the drive circuit must be designed for the worst case. The light output of the diode also degrades with accumulated forward-current hours and with temperature, so the current-transfer ratio falls over the product's life. Conservative designs derate the initial value substantially and drive the diode at the lowest current that meets the timing requirement, since aging accelerates with drive current.

Speed varies widely by construction. Phototransistor-output devices, whose bandwidth is limited by the large base-collector capacitance of the photodetector, are suited to signals in the range of a few kilobits to a few hundred kilobits per second. Devices with an integrated photodiode and a transimpedance amplifier, sold as logic-output or high-speed optocouplers, reach tens of megabits per second at the cost of a supply on the output side. Optocouplers also draw milliamperes of input current, which matters in battery-powered and multichannel designs, and each channel requires its own package in most product families. Despite these limitations, optocouplers remain widely used, especially where their long record of safety certification is advantageous, where a simple current-driven input without an isolated supply is convenient, and in analog feedback roles where a linear optocoupler with a matched servo photodiode provides a stable transfer characteristic.

Digital Isolators

Digital isolators are integrated circuits that transmit digital signals across an isolation barrier using a capacitive or magnetic coupling element fabricated within the chip. They have largely superseded optocouplers in new designs that require high speed, because they offer far greater data rates, lower power consumption, tighter timing, and excellent stability over temperature and time. A digital isolator encodes the input logic state, transmits it across the integrated barrier, and reconstructs it on the output side, often integrating several channels in a single package.

Capacitively coupled isolators transmit signals across a high-voltage capacitor formed from an insulating dielectric layer between two electrodes inside the package. The transmitter modulates a high-frequency carrier that passes through the capacitor as a displacement current, while the low-frequency and direct-current components are blocked, preserving galvanic isolation. The receiver demodulates the carrier to recover the original logic signal. Silicon dioxide and similar dielectrics provide robust, stable insulation, and the small, well-controlled barrier supports high data rates and strong common-mode transient immunity.

Magnetically coupled isolators instead transmit signals across a pair of on-chip microtransformer coils separated by an insulating layer, typically a thick polyimide film. The transmitter drives a current pulse through the primary coil, inducing a pulse in the secondary coil that the receiver interprets to reconstruct the signal. The transformer transfers energy through a changing magnetic field, blocking any direct conductive path and thereby maintaining isolation. Like capacitive isolators, magnetic isolators achieve high speed, low power, and good stability, and both approaches are manufactured with standard semiconductor processes that integrate the barrier alongside the signal-conditioning circuitry.

Because a pulse-based scheme conveys only transitions, a signal that remains static carries no information across the barrier. Practical isolators solve this either by refreshing the state periodically, so the receiver reacquires the correct level within a defined interval after power-up or after a disturbance, or by using an on-off keyed carrier whose presence encodes the logic level directly. Both schemes define a failsafe output state that the receiver adopts when the input side is unpowered or the link is lost, which is a safety-relevant choice: a gate driver should default to holding a power transistor off, not on.

Typical commercial parts offer one to six channels in a single small-outline package, with channels available in either direction, data rates from tens to a few hundred megabits per second, propagation delays in the range of roughly ten to fifty nanoseconds, and channel-to-channel skew of a few nanoseconds. Supply current is often well under a milliampere per channel at low data rates and rises roughly in proportion to switching frequency, which is a marked advantage over an optocoupler's constant diode drive current. Timing is also far more stable: propagation delay varies only slightly over temperature and supply, whereas an optocoupler's delay tracks its aging current-transfer ratio. These properties are why digital isolators dominate new designs for isolated buses, gate drives, and multichannel interfaces, and why the topic is treated in more component-level detail in the article on galvanic isolation devices.

Isolation Transformers

An isolation transformer transfers power or signals across a barrier through magnetic coupling between separate primary and secondary windings that share no electrical connection. Because the windings are galvanically isolated, the secondary side floats relative to the primary, and a transformer with appropriate insulation between windings provides safety isolation between, for example, the mains and a low-voltage output. Transformers are the principal means of isolating power in switch-mode supplies, where the energy itself must cross the barrier, and in power line-frequency applications where a one-to-one isolation transformer separates equipment from the supply.

The isolation quality of a transformer depends on the insulation between its windings, including the wire insulation, any interwinding insulating tape or barrier, and the creepage and clearance maintained at the terminations. Transformers intended for safety isolation are constructed with reinforced or double insulation between primary and secondary and are tested to withstand high voltages between the windings. Two construction methods dominate. Margin-wound transformers leave an insulating margin tape at each end of the bobbin so that the windings are physically set back from one another, satisfying creepage along the bobbin surface at the cost of usable winding width. Triple-insulated wire instead places three independently applied insulating layers on the conductor itself, so the wire alone provides reinforced insulation and the windings may be wound directly on top of one another; this yields tighter coupling and lower leakage inductance in the same core size, which suits high-frequency switch-mode designs. Multiple layers of interwinding tape are a third approach, generally used in combination with the others.

Transformer isolation is not perfect at high frequency. Interwinding capacitance, typically a few picofarads to tens of picofarads, couples common-mode displacement current from the switching primary to the secondary, which is a principal source of conducted emissions in isolated converters. An electrostatic shield, a grounded foil or single-layer winding placed between primary and secondary, intercepts that current and returns it to the primary side, and is a standard remedy in designs sensitive to electromagnetic compatibility limits. Beyond power transfer, small signal transformers isolate data interfaces such as Ethernet, where they also provide impedance matching and common-mode rejection; Ethernet magnetics are conventionally specified for a 1500 V root-mean-square withstand between the cable side and the equipment side. Line-frequency isolation transformers, by contrast, are large and heavy but simple, and a one-to-one unit is still the standard way to float a test bench or a piece of legacy equipment from the supply. Transformer construction is treated further in the article on transformers.

Isolated Power and Specialized Devices

Many isolated systems must deliver power as well as signals across the barrier, since the circuitry on the isolated side needs a supply. Isolated direct-current converters use a small transformer to carry energy across the barrier while maintaining isolation, and some integrated devices combine a digital isolator with an integrated isolated power converter in a single package. Isolated gate drivers for power transistors integrate the isolation barrier with the drive circuitry, transferring both the control signal and, in some products, the gate-drive energy across the barrier. Isolated amplifiers and isolated analog-to-digital converters carry analog measurements across the barrier with defined accuracy, enabling sensing of high-voltage quantities from a low-voltage reference.

Creepage and Clearance

Clearance is the shortest distance through air between two conductive parts, and creepage is the shortest distance along the surface of solid insulation between them. Clearance governs the voltage at which the air gap breaks down and arcs over, and it is set by the peak working voltage and the transient overvoltages the barrier must withstand. Creepage governs the tendency for a conductive track to form along an insulating surface over time, a process driven by surface contamination and humidity, and it is set by the working voltage, the pollution degree of the environment, and the surface resistance of the insulating material.

Required creepage and clearance distances increase with working voltage and with the severity of the environment. IEC 60664-1 is the base standard that supplies these coordination rules, and product standards draw their tables from it. Three environmental classifications drive the numbers. Pollution degree describes the contamination expected at the barrier, running from degree 1, a sealed or hermetic environment with no conductive pollution, through degree 2, the usual case for household and office equipment where only non-conductive pollution occurs and occasional condensation is possible, to degrees 3 and 4 for industrial and outdoor environments where conductive pollution or persistent conductivity from rain must be assumed. Overvoltage category, running from I to IV, describes the transient the barrier may see from the supply, with category II covering equipment plugged into a fixed installation and categories III and IV covering fixed installations and the origin of the installation, where surges are progressively larger.

Material matters as well. The comparative tracking index measures how well an insulating surface resists the formation of a conductive carbonized track under a contaminated electric field, and standards sort materials into groups by that index: group I at 600 volts and above, group II from 400 to below 600, group IIIa from 175 to below 400, and group IIIb from 100 to below 175. A better material group permits shorter creepage at the same working voltage, which is why the choice of laminate and of connector plastics can decide whether a compact layout is certifiable. Clearance carries its own environmental dependence: air is a weaker insulator at reduced pressure, so clearances must be increased for equipment rated above 2,000 meters of altitude.

Designers obtain the required distances from the applicable safety standard and ensure the layout of an isolation device, its package, and the surrounding circuit board all maintain at least those distances across the barrier. The governing figure is the smallest distance anywhere along the barrier, not the nominal package pin-to-pin spacing, so a via, a test point, a silkscreen-invisible copper pour, or a mounting screw placed carelessly can defeat an otherwise correct design. Where board area is scarce, a slot milled through the printed circuit board along the barrier lengthens the creepage path without lengthening the clearance, and potting or conformal coating can permit a lower assumed pollution degree if applied in a qualified manner. Inadequate creepage or clearance is a common cause of insulation failure and certification rejection, and it is one of the most expensive faults to discover late, since correcting it usually means respinning the board.

Insulation Grades: Basic, Supplementary, and Reinforced

Safety standards classify insulation by the level of protection it provides. Basic insulation provides a single level of protection against electric shock under normal conditions. Supplementary insulation is an independent second layer applied in addition to basic insulation, so that protection remains if the basic insulation fails. Double insulation comprises basic plus supplementary insulation together. Reinforced insulation is a single insulation system that provides protection equivalent to double insulation, offering two levels of protection within one barrier that is built and tested to a correspondingly higher standard.

The grade of insulation required depends on the role the barrier plays in the safety scheme. A barrier that separates a user-accessible circuit from a hazardous voltage typically must provide double or reinforced insulation, so that no single insulation failure exposes the user to the hazard. Functional or basic isolation may suffice where additional protective measures exist elsewhere. Isolation components are specified and certified for a particular insulation grade and a maximum working voltage, and selecting a component of the correct grade is essential to a valid safety design.

Withstand Voltage and Working Voltage

An isolation barrier is characterized by several voltage ratings, and confusing them is a frequent source of design error. The maximum working voltage is the continuous voltage the barrier is designed to sustain across its life. The withstand voltage, verified by a high-potential test, is a much higher voltage the barrier must survive briefly without breakdown, demonstrating margin above normal operation. Surge or impulse ratings define the transient the barrier can survive, tested with a standardized 1.2/50 microsecond waveform.

Two rating systems coexist on most datasheets. Under UL 1577, the isolation voltage is the root-mean-square value the barrier withstands for sixty seconds in a type test, with values of 2,500, 3,750, and 5,000 volts root-mean-square being common; every production unit is then screened for one second at 120 percent of that rating. This single headline number is widely quoted but says nothing about continuous operation. The IEC component standards instead publish a family of ratings: a maximum repetitive peak isolation voltage for continuous operation, a transient isolation voltage for brief excursions, a surge isolation voltage for impulses, and, where applicable, a maximum working insulation voltage expressed as a root-mean-square value.

The practical consequence is that a part advertised as a 5,000-volt isolator may be qualified for only a few hundred volts of continuous working voltage. The sixty-second test voltage is a proof of construction, not a permission to operate there. Designers must identify the actual working voltage their application imposes across the barrier, including any direct-current offset and the peak of any superimposed ripple, and then confirm that it falls within the device's continuous rating rather than within its headline withstand number.

Barrier Lifetime and Wear-Out

Solid insulation does not fail only from a single overvoltage; it also wears out. Under sustained electric stress, microscopic defects in a dielectric propagate until a conductive path forms, a process known as time-dependent dielectric breakdown. The rate depends steeply on field strength and on temperature, so a barrier operated well below its breakdown voltage lasts far longer than one operated near it, and the relationship is strongly nonlinear rather than proportional.

Modern component standards address this explicitly. Manufacturers of integrated isolators subject samples to accelerated voltage stress until failure, fit the resulting distribution to a lifetime model, and extrapolate to determine the working voltage at which the barrier meets a specified service life, commonly stated as a target on the order of decades at a defined failure rate. The published continuous working-voltage rating is the result of that extrapolation, which is why it is so much lower than the one-minute withstand voltage. Alternating-current stress is generally more damaging than direct-current stress at the same magnitude, so datasheets often list a lower rating for alternating working voltage than for direct.

A repetitive partial-discharge test complements lifetime testing. Partial discharge is a localized breakdown within a void or at an interface that does not yet bridge the whole insulation but progressively erodes it. Detecting discharge activity below a specified threshold at a voltage above the working voltage screens out barriers with latent voids, which is why the test is applied to every unit in production for reinforced-insulation parts rather than only to type-test samples.

Product Safety Standards and Insulation Coordination

Product safety standards such as IEC 62368-1, the hazard-based standard for audio, video, information, and communication technology equipment that superseded the earlier IEC 60950-1 and IEC 60065, establish the framework for insulation coordination. They define how working voltages, pollution degrees, and overvoltage categories translate into required creepage, clearance, and insulation grade, classifying energy sources by hazard and prescribing safeguards accordingly. These standards specify the withstand tests an isolation barrier must pass and the conditions under which basic, supplementary, double, or reinforced insulation is required. The broader landscape of such standards is treated in the discussion of electrical safety standards, of which isolation requirements are an integral part.

Component Standards for Isolators

Isolation components are evaluated against dedicated safety standards that certify their suitability for safety isolation. IEC 60747-5-5 has long governed the safety ratings of optocouplers, defining the maximum surge and working voltages and the testing required to qualify them for reinforced or basic insulation. IEC 60747-17, the first international component standard written specifically for magnetic and capacitive digital isolators, extends comparable safety qualification to the integrated barriers used in modern isolator integrated circuits; its harmonized European and German edition is designated DIN EN IEC 60747-17 (VDE 0884-17), which superseded the earlier preliminary standard VDE V 0884-11. Components certified to these standards carry defined insulation ratings that a system designer can rely upon in the product safety case.

These standards qualify a component for a specific isolation class, and the test severity scales with the class. A device rated for reinforced insulation must survive a higher impulse than one rated for basic insulation only: under IEC 60747-17, the surge withstand test is applied at a peak of 1.6 times the rated surge voltage for reinforced insulation against 1.3 times for basic insulation, reflecting the larger margin a reinforced barrier must demonstrate. A repetitive partial-discharge test further screens for latent insulation weaknesses, and the qualified maximum repetitive and surge isolation voltages appear directly on the datasheet.

A parallel North American scheme, UL 1577, recognizes optocouplers and isolators by the sixty-second withstand test described above. The two schemes are complementary rather than interchangeable: UL 1577 certifies a withstand capability, while the IEC component standards additionally qualify a continuous working voltage, an insulation class, and a projected lifetime. Most isolators sold for safety-critical use carry certification under both, along with agency approvals such as CSA and CQC, and the datasheet lists each with its own set of ratings.

Using a certified isolation component simplifies system certification, because the component's qualified ratings can be cited directly as evidence that the barrier meets the required insulation grade and withstand voltage. The designer must still ensure that the surrounding layout preserves the necessary creepage and clearance and that the device is operated within its rated working voltage, but the component certification establishes the fundamental capability of the barrier itself. Certification also constrains later changes: substituting a pin-compatible part from another vendor does not automatically carry the safety case across, because the two devices may hold different insulation classes and working-voltage ratings despite identical footprints.

Application-Specific Standards

Particular product categories impose their own isolation requirements reflecting the hazards of the application. Medical electrical equipment standards, such as IEC 60601-1, define means of patient protection and means of operator protection with specific isolation requirements, recognizing that patients may be directly connected to equipment and especially vulnerable. A means of patient protection is more demanding than a means of operator protection at the same working voltage: for a mains working voltage of 250 volts, one means of operator protection follows the ordinary information-technology equipment spacings, whereas two means of patient protection call for on the order of 8 millimeters of creepage, 5 millimeters of clearance, and a dielectric strength test at 4,000 volts alternating current. This is why medical-grade power supplies and isolators are specified separately from their industrial equivalents even when their electrical performance is identical.

Other sectors set their own rules. IEC 61010-1 governs measurement, control, and laboratory equipment, where an instrument's input terminals may be connected directly to an energized installation and the measurement category assigned to those terminals determines the impulse withstand required. IEC 61800-5-1 covers adjustable-speed electrical power drive systems, addressing the isolation between control electronics and motor circuits. Energy-metering and grid-connected equipment standards impose requirements suited to their exposure to the distribution network, where overvoltage categories III and IV apply. Designers identify the standards applicable to their product category and ensure the isolation scheme satisfies the most demanding applicable requirement, since a product sold into several markets must meet the union of their obligations, not the average.

Board Layout Across the Barrier

The barrier should appear on the layout as an explicit, documented corridor running the full width of the board, with nothing crossing it except the isolation components themselves. No trace, no copper pour, no ground or power plane, no via, and no silkscreen conductor may bridge that corridor. Planes deserve particular attention, because an internal plane that extends under the isolator package shortens the effective clearance in a way that is invisible on the top-layer view. Component bodies, heat sinks, mounting hardware, and enclosure features must also respect the corridor.

Where the required creepage exceeds the available width, a slot milled through the board along the barrier is the standard remedy, since it forces the surface path around the slot edges while leaving the through-air clearance governed by the package. Connectors and test points on either side must sit outside the corridor, and any conformal coating or potting compound applied to reduce the effective pollution degree must be qualified for that purpose rather than assumed adequate. Documenting the corridor in the fabrication drawing and in the design review checklist is what keeps it intact through later revisions, when an engineer unfamiliar with the safety case may otherwise route a signal across it.

Powering the Isolated Side

Circuitry on the far side of a barrier needs a supply that does not compromise the barrier. The usual options are an auxiliary winding on the main power transformer, a small isolated direct-current converter, a transformer driver with a discrete transformer, or an isolator with an integrated power stage in the same package. The integrated approach is the most compact and simplifies the safety case, since one certified component covers both signal and power, but it delivers limited power, typically in the tens to low hundreds of milliwatts, and its high-frequency switching is a deliberate emissions source that requires attention.

Every isolated supply couples common-mode current across its transformer's interwinding capacitance, and that current returns through whatever path the system offers, often the very cable shields and chassis connections the designer hoped to keep quiet. Minimizing interwinding capacitance, adding an electrostatic shield, spreading the switching spectrum, and providing a deliberate low-impedance return through a high-voltage capacitor rated for the application are the standard countermeasures. A capacitor bridging the two grounds must be selected as a safety-rated component, because it sits across the barrier and its failure mode determines whether the barrier survives.

Choosing an Isolation Approach

The selection turns on a small number of questions. What insulation grade does the safety case demand, basic or reinforced? What is the true continuous working voltage across the barrier, and what surge must it survive? How fast must the signal cross, and does the design tolerate the drift and aging of an optocoupler's transfer characteristic? How severe are the common-mode transients, and how many channels are needed? Does the isolated side require power, and how much?

Optocouplers remain attractive where a current-driven input avoids the need for an isolated supply, where the signal is slow, where the required insulation is well within a mature and inexpensive part's capability, or where an existing certification is being carried forward. Digital isolators win where speed, channel count, timing stability, power consumption, or common-mode immunity is the binding constraint, which describes most new industrial and automotive designs. Transformers are the answer whenever the barrier must carry appreciable power rather than only information. Many systems combine all three, and the essential discipline is to treat each crossing as part of a single coherent barrier rather than as an isolated component choice.

Power Supplies and Converters

Isolated power supplies separate their output from the mains so that the output and anything connected to it remain at a safe potential even though the input is connected to hazardous line voltage. In switch-mode power supplies the main isolation transformer carries power across the barrier, while an optocoupler or digital isolator carries the regulation feedback from the output side back to the controller on the primary side. This arrangement lets the supply regulate its output precisely while maintaining a certified safety barrier between the user-accessible output and the dangerous input.

Two variations are worth noting. Primary-side regulation dispenses with the feedback isolator entirely, inferring the output voltage from the reflected voltage on an auxiliary winding, which saves cost and an aging component at the price of looser regulation. Where tight regulation is required without an optocoupler, a digital isolator or a dedicated isolated-feedback part carries the error signal with stable timing and no current-transfer-ratio drift. In either case the isolation transformer remains the safety barrier, and the feedback path must itself meet the same insulation grade, since it too bridges the primary and secondary domains.

Motor Drives and Power Electronics

In motor drives and inverters, isolated gate drivers separate the low-voltage control electronics from the high-voltage, fast-switching power stage, protecting the controller and the operator. The isolation barrier withstands the large, rapidly changing voltages of the switching nodes, and high common-mode transient immunity prevents those transients from corrupting the drive signals. Isolated current and voltage sensing carries measurements from the high-voltage power circuit to the low-voltage controller, enabling closed-loop control without a conductive connection between the two domains.

Isolated gate drivers also carry protection functions across the barrier. Desaturation detection senses an abnormal collector or drain voltage while the transistor is commanded on, indicating a short circuit, and the driver responds with a controlled soft turn-off to limit the induced voltage spike, then reports the fault back across the barrier to the controller. Miller clamping holds the gate firmly at its off level to prevent a fast rising voltage on the switching node from turning the device on parasitically. Traction inverters in electric vehicles push these requirements further, operating at 400 or 800 volts on the high-voltage side with reinforced insulation to the vehicle chassis, and their isolation monitoring is itself a safety function within the vehicle's functional safety scheme.

Communication Interfaces

Isolated communication interfaces prevent ground loops and protect equipment where data links connect systems with different ground potentials or span electrically noisy environments. Industrial buses, controller area networks, and serial links frequently employ digital isolators so that a ground potential difference between two nodes does not drive damaging current through the signal lines. A long RS-485 or CAN run across a factory floor can easily see several volts of ground potential difference between its ends under normal conditions and far more during a fault, which is why isolated transceivers, integrating the barrier, the line driver, and often an isolated supply in one package, are the default choice for such buses. Ethernet interfaces use isolation transformers that separate the equipment from the cable, blocking ground loops and providing a measure of protection against surges entering on the cable. The relationship between isolation and ground-potential differences is developed further in the article on grounding and bonding, and the surge threat on external cabling in the article on overvoltage protection.

Instrumentation and Measurement

Instrumentation often must measure voltages referenced to potentials far from the instrument's own ground, and isolation makes such measurements safe and accurate. Isolated amplifiers and isolated analog-to-digital converters carry an analog measurement across a barrier, allowing a low-voltage data-acquisition system to sense a high-voltage signal without a conductive connection. This isolation protects the measurement electronics and the operator, breaks ground loops that would otherwise corrupt sensitive readings, and permits accurate sensing of quantities such as high-side currents and floating voltages.

Medical Equipment

Medical electrical equipment relies on isolation to protect patients, who may be directly connected to a device and unable to tolerate even small leakage currents. Isolation barriers limit the current that can flow through a patient connection under both normal and fault conditions, implementing the means of patient protection that medical standards require. Applied parts that contact the patient are isolated from the mains and from other circuits to a degree determined by the type of contact. The classification runs from type B parts, which have no conductive patient connection, through type BF parts, which are floating, to type CF parts intended for direct cardiac application, where the leakage limits are strictest because current delivered near the heart is dangerous at levels far below the threshold of sensation at the skin. Ensuring that a single fault cannot subject the patient to a hazardous current is the organizing principle behind the two-means-of-protection requirement.