Signal Isolator

What is a Signal Isolator

A signal isolator is an interface device that conveys an analog process signal from an input circuit to an output circuit while breaking every direct electrical connection between them. The two sides share no common conductor and no common ground. Instead, the input stage measures the incoming signal, the measurement crosses an insulating barrier through light, a magnetic field, or an electric field, and the output stage re-creates a faithful copy of the original signal referenced to its own, independent ground. In most industrial units a third barrier also isolates the power supply, giving what manufacturers call 3-way or 3-port isolation: input, output, and power are mutually isolated.

The reason isolators exist is the ground loop. In a real plant, the field sensor, the cable shield, the marshalling cabinet, and the control system are earthed at physically separate points. Those earth points are rarely at exactly the same potential: motor drives, welding sets, frequency converters, and long cable runs can create a difference of a few millivolts to several volts between two grounds. When a continuous metallic path links them, that potential difference drives an unwanted current through the signal wires, adding error and noise to a 4-20 mA reading or, at worst, damaging an analog input card. By removing the metallic path, the isolator lets each ground float at its own potential and leaves only the intended signal to cross.

A signal isolator belongs to the wider family of signal conditioners, but the two terms are not synonyms. A signal conditioner is any module that amplifies, filters, linearizes, converts, or splits a signal; isolation is one optional function among many. An isolator is the subset whose first purpose is the galvanic barrier, and whose datasheet therefore specifies a port-to-port test voltage, typically 1.5 kV, 2.5 kV, or 3 kV AC applied for one minute under IEC 61010-1. A plain range converter that turns 0-10 V into 4-20 mA without that barrier is a conditioner but not an isolator. In practice most modern isolators do both jobs, isolating and converting in one DIN-rail module, which is why the categories overlap on supplier websites.

Functionally the device is closely related to its neighbors in a control cabinet. A signal conditioner performs the conversion and filtering, a safety barrier limits energy into hazardous areas, and a surge protector diverts transient overvoltage to earth. The isolator overlaps all three: many models add range conversion like a conditioner, the intrinsically safe variant is itself a certified barrier, and the isolation barrier inherently provides a degree of transient blocking. Understanding where one function ends and another begins is the first step in correct selection.

Four engineering attributes determine an isolator's quality and price: the isolation rating (test and working voltage), the transfer error it adds to the signal, the temperature drift across the cabinet operating range, and whether it draws power from the loop or from an external supply. These attributes set the total cost of ownership. A cheap non-isolated converter looks attractive until the first ground loop corrupts a flow reading or a transient destroys a PLC input card, at which point the retrofit cost of proper isolation far exceeds the saving. Specifying isolation correctly at design time is almost always cheaper than diagnosing it in the field.

Isolator Types and Architectures

Industrial signal isolators divide first by how they are powered and how many independent outputs they provide. The power architecture sets the wiring complexity and what the isolator can and cannot do. Choosing the wrong architecture is the most common selection mistake: a loop-powered isolator dropped into a loop with insufficient supply headroom will simply fail to reach 20 mA, while an unnecessary 4-wire unit adds a 24 VDC feed and cabinet space that the application never needed. The table below summarizes the four mainstream architectures.

Loop-powered (passive) isolators draw all their operating energy from the 4-20 mA loop they sit in, so they need no separate supply and reduce to two terminals in and two terminals out. This is the simplest possible wiring and the obvious choice when a 24 VDC feed is inconvenient. The cost is burden voltage: the isolator consumes a slice of the loop compliance, typically dropping in the region of 2.5 to 5 V or more, which the loop power supply must cover on top of the receiver burden and cable resistance. A loop-powered unit also drives only passive loads and cannot source power to a field transmitter. Examples include the Phoenix Contact MINI MCR-SL-1CP-I-I passive separator and the Datexel DAT511 loop-powered isolator.

3-wire and 4-wire externally powered isolators take a separate supply, commonly 24 VDC for 3-wire units or a wide 5-36 VDC plus 90-250 VAC range for 4-wire units. Because they have their own energy budget, they can convert ranges freely, for example 4-20 mA in to 0-10 V out, source loop power to a 2-wire transmitter, and present a low input burden so that voltage drop stays under about 1.5 V at full scale. The Weidmuller ACT20M family and Acromag 4-wire isolators are typical: the ACT20M-AI-AO accepts 0/4-20 mA or 0/2-10 V and isolates power from input and output as a true 3-port device. Externally powered units are the default when conversion, voltage output, or transmitter excitation is required.

Splitters and duplicators solve the multi-receiver problem. Wiring one current source through a PLC, a recorder, and a panel meter in series stacks their burdens and ties their grounds together, defeating the purpose of isolation. A splitter takes one isolated input and reproduces it as two or more independent, mutually isolated outputs, each driving its own receiver against its own ground. The Weidmuller ACT20M-CI-2CO and Datexel DAT4530 are representative 2-output devices. Loop-powered 2-output splitters exist for budget-constrained cases, but the combined output burden must be checked against the supply.

A fourth axis is channel count and packaging. Single-channel modules run from roughly 6 mm wide on a DIN rail (the Phoenix MINI MCR-SL is 6.2 mm), enabling dense marshalling. Multi-channel isolators pack two, four, or eight channels in one housing to save space and often share a power feed through a DIN-rail bus connector, trading per-channel replaceability for density. High-channel-count cabinets increasingly use bus-powered modules to avoid wiring an individual supply to every isolator.

Coupling Principles Across the Barrier

Every isolator must move information across an insulating gap without a wire. Three physical principles do this in industrial and component-level isolators: optical coupling (light), magnetic coupling (a transformer or coupled coils), and capacitive coupling (an electric field across a dielectric). Each principle trades isolation voltage, bandwidth, common-mode rejection, and cost differently, so the right choice depends on the application. The table below compares the three at the level that matters for selection.

Optical coupling uses an LED on the input side whose light crosses a transparent gap to a photodiode or phototransistor on the output side. Because nothing but photons crosses the barrier, optical isolation can reach high isolation voltages and provides excellent common-mode noise rejection, which is why optocouplers have been the workhorse of isolation for decades. The trade-off is speed and drift: the LED-to-photodiode transfer is comparatively slow and its efficiency ages and varies with temperature, so high-accuracy analog isolators that use the optical principle must add feedback to cancel that drift. For digital signals and moderate-bandwidth analog loops, optical coupling remains a reliable, well-understood choice.

Magnetic, or transformer, coupling passes the signal as a modulated magnetic field between two coils separated by an insulating gap. Transformers handle higher isolation voltages than optical or capacitive couplers of similar size, making magnetic coupling the preferred principle when the barrier must withstand large potential differences, and the technique is mechanically robust. The signal is usually chopped into a high-frequency carrier, sent across the transformer, then demodulated, which also lets the same transformer carry power across the barrier to feed the isolated stage. Many DIN-rail isolators use a transformer both to transfer the signal and to power the far side, which is how a 4-wire unit isolates its supply.

Capacitive coupling transfers the signal as a changing electric field through a small capacitor whose plates sit either side of a thin dielectric. Capacitive isolators are physically tiny, draw little power, switch fast (digital isolators of this type support loop bandwidths up to roughly 50 kHz), and are stable across temperature because a capacitor's behavior is far more predictable than an aging LED. Their isolation voltage sits between optical and magnetic for a given size, and a strong external electric field can in principle couple in, so practical parts use differential structures to reject it. In modern integrated digital isolators capacitive and magnetic techniques dominate, while board-level and DIN-rail analog isolators mix all three depending on the accuracy and isolation target.

One point often confused at selection time: the coupling principle is largely invisible on a DIN-rail isolator's datasheet, which quotes the resulting test voltage, working voltage, transfer error, and bandwidth rather than the internal physics. Engineers should select on those output specifications. The principle matters when it constrains an outcome, for example when only transformer coupling can meet a required multi-kilovolt barrier, or when only a capacitive or magnetic part can meet a high-bandwidth requirement that an optocoupler cannot reach.

Signal Interfaces and Standards

An isolator sits between a field signal and a control system, so its input and output interfaces must match both ends. The dominant industrial interface is the 4-20 mA current loop, prized because current is immune to cable voltage drop over long runs and because a reading of 0 mA flags a broken wire. But isolators also handle voltage signals, sensor inputs such as thermocouple and RTD, frequency and pulse signals, and the HART digital protocol overlaid on the current loop. The table below lists the common interfaces an isolator must support.

Most general-purpose isolators are current-to-current or current-to-voltage. A universal-input isolator widens this to accept thermocouple, RTD, millivolt, resistance, current, and voltage on one configurable input, with the output set by software or DIP switch to a standard 4-20 mA or 0-10 V. Moore Industries, PR electronics, and Acromag all offer universal modules of this kind. The advantage is spares rationalization: one part number covers many tag types, and the engineer configures rather than restocks. The cost is a slightly higher transfer error and price than a fixed single-function isolator, so dedicated high-accuracy loops still favor a purpose-built module.

HART transparency is a frequent trap. HART superimposes a low-amplitude frequency-shift-keyed digital signal on the 4-20 mA current. An isolator with insufficient bandwidth or an aggressive low-pass filter will strip the HART tones, leaving the analog value intact but killing remote configuration and diagnostics. When the loop carries HART, specify an isolator explicitly rated HART-transparent. The intrinsically safe isolated barriers in the Pepperl+Fuchs KFD2-STC family, for instance, pass HART and other SMART protocols bidirectionally precisely so that a handheld or asset-management system can reach a field transmitter through the barrier.

The governing standards an isolator must respect are international and explicit. IEC 61010-1 defines the one-minute dielectric withstand (test voltage) procedure and the safety requirements for measurement and control equipment. IEC 60664-1 sets insulation coordination: it derives the required clearance and creepage distances from the working voltage, the overvoltage category of the installation, and the pollution degree, and it distinguishes basic, supplementary, and reinforced insulation, with reinforced insulation dimensioned to withstand 160 percent of the basic-insulation withstand voltage. IEC 61326 covers electromagnetic compatibility for measurement and control equipment, and many process isolators additionally meet NAMUR NE 21, whose criterion A demands that the signal influence under EMC test stays within a tight band, with quality units claiming better than plus-or-minus 0.5 percent. For functional safety, IEC 61508 defines the SIL levels, and certified barriers commonly carry SIL 2 (SC 3).

Key Specification Parameters

Reading an isolator datasheet means filtering perhaps two dozen listed values down to the handful that actually decide selection. Seven parameters dominate: test voltage, working voltage, transfer error, temperature drift, response time and bandwidth, burden or supply requirement, and operating temperature range. Each is explained below, with the typical industrial values established by manufacturer datasheets and the standards in Chapter 4.

Test voltage (dielectric strength) is the AC voltage applied between ports for one minute without breakdown under IEC 61010-1, commonly 1.5 kV, 2.5 kV, or 3 kV AC. The Weidmuller ACT20M, for example, specifies 2.5 kV AC test voltage between input, output, and auxiliary power as a 3-port figure. Test voltage proves the barrier survives a transient; it is not a continuous operating rating, and engineers must not confuse the two.

Working voltage (rated insulation voltage) is the steady-state potential the barrier may carry indefinitely, typically 250 to 300 V AC/DC for general industrial isolators (the ACT20M rates 300 V AC/DC). For protecting personnel from a mains-referenced source, the working voltage drives the required creepage and clearance through IEC 60664-1, and reinforced insulation must be specified. For breaking a low-voltage ground loop only, a modest working voltage and a 1.5 kV test voltage are sufficient.

Transfer error is the deviation the isolator adds between input and output, in percent of full scale, typically 0.05 to 0.1 percent FS for quality units. It is the isolator's own contribution and adds to the upstream sensor error in the loop budget. A 0.075 percent transmitter feeding a 0.1 percent isolator yields a worst-case combined uncertainty near 0.175 percent FS if summed linearly, so the isolator must be chosen so the total stays inside the loop tolerance.

Temperature drift specifies how transfer error grows away from the calibration temperature, often around 0.005 percent FS per degree Celsius over a stated band such as -20 to +55 degrees C. In a hot cabinet this term can exceed the room-temperature transfer error, so for wide-temperature installations read the drift figure, multiply by the worst-case temperature excursion, and add it to the base error rather than quoting only the headline accuracy.

Response time and bandwidth govern how fast the output follows a step change and which frequencies pass. Process isolators commonly quote a step response of a few milliseconds to under one second to 90 percent, and a bandwidth that must be high enough to pass HART tones when required. For fast control loops or pulse and frequency signals, confirm the bandwidth explicitly; a slow isolator quietly adds lag to a control loop.

• Burden voltage (loop-powered units): the voltage the isolator consumes from the loop, often 2.5 to 5 V or more, which must be subtracted from the supply compliance along with the receiver burden and cable resistance.
• Supply and power loss (externally powered units): input range such as 19.2 to 30 VDC for the Phoenix MINI MCR-SL, or 5-36 VDC plus 90-250 VAC for wide-range 4-wire units, with power dissipation around 0.8 W per channel.
• Operating temperature: a typical cabinet rating of -20 to +65 degrees C (MINI MCR-SL) or -20 to +55 degrees C; verify the isolator covers the cabinet's real worst-case ambient, not just room temperature.
• Module width and channels: from 6.2 mm single-channel slim modules to multi-channel housings, a direct driver of cabinet density and DIN-rail length.
• Output load capability: the maximum load resistance the current output can drive at 20 mA, which sets how many series receivers a single output can feed.

A final note on accuracy semantics: as with transmitters, manufacturers sometimes fold non-linearity, hysteresis, and temperature effects into a single headline figure, and sometimes list them separately. For a critical loop, request the itemized data and build an error budget rather than trusting one percentage. Long-term stability, sometimes quoted as better than 0.1 percent over three years on premium units, belongs in that budget too.

Selection Decision Factors

To turn the preceding chapters into a specific model, follow the decision sequence below. As with most field instruments, selection errors come less from a single wrong value than from deciding a downstream detail before settling the architecture above it. These eight steps work as a fixed RFQ template for signal isolators.

1. Confirm you actually need isolation: identify the ground-loop or transient risk. If two grounds are at different potentials, a motor drive or VFD is nearby, or you are protecting an analog input card from the field, isolation is justified. If not, a non-isolated conditioner may suffice and costs less.
2. Define input and output signals: current, voltage, thermocouple, RTD, frequency, or HART-bearing 4-20 mA. Decide whether conversion is needed (mA to V) and whether HART must pass transparently. This narrows you to fixed-function or universal-input units.
3. Choose the power architecture: loop-powered if wiring simplicity matters and loop compliance has headroom; 3-wire or 4-wire externally powered if you need conversion, voltage output, transmitter excitation, or a splitter. Verify burden voltage against the supply for loop-powered units.
4. Set the isolation rating: a 1.5 kV test voltage and modest working voltage for functional ground-loop isolation; reinforced insulation sized per IEC 60664-1 to the overvoltage category and pollution degree for safety isolation from a mains-referenced source.
5. Fix the accuracy budget: select transfer error (0.05 to 0.1 percent FS) and temperature drift so that, summed with the upstream sensor and projected across the cabinet's worst-case temperature, the total stays inside the loop tolerance.
6. Check certifications and environment: hazardous-area duty needs an intrinsically safe isolated barrier with ATEX, IECEx, or NEPSI and the correct gas group and temperature class. Confirm EMC compliance (IEC 61326, NAMUR NE 21) and functional safety (IEC 61508 SIL) where the loop demands it.
7. Match the physical and electrical packaging: module width and channel count for cabinet density, DIN-rail bus power versus individual feeds, terminal type, and whether a multi-channel housing is acceptable given replaceability needs.
8. Build the total cost of ownership: purchase price plus wiring labor, cabinet space, spares strategy (universal modules reduce part numbers), and the downstream cost of a missed ground loop or a stripped HART signal. The cheapest module is rarely the lowest lifecycle cost.

One dimension that purchasing teams routinely undervalue is serviceability and supply continuity: whether the chosen series remains in production over the plant's life, whether spares ship quickly, whether configuration is by hardware switch or by software tool, and whether the vendor maintains local support. Mainstream series with broad distribution, such as Phoenix Contact MINI MCR-SL, Weidmuller ACT20M, Acromag isolators, PR electronics 3000 series, and the Pepperl+Fuchs K-System for intrinsically safe duty, are widely stocked and configurable, which lowers the long-run risk of an unobtainable replacement on a running line. Verify each candidate's exact isolation rating, transfer error, and certification against your installation before committing to a purchase order.