4-20mA vs RS485 vs Analog Secondary: A CT Output Decision Framework

Choosing an output format for a current transformer looks like an interface detail. In practice it decides which team owns accuracy, which team owns calibration, and which document a buyer has to produce during an audit. The question sharpens when the device is a PCB mount current transformer: the secondary signal now terminates a few millimetres from the analogue front end instead of at a terminal block inside a panel, so decisions that used to sit with the panel builder move onto the board.

Three output paths dominate this space. An analogue secondary hands a scaled AC signal to the host electronics. A 4-20mA transmitter converts the measurement into a loop current. An RS485 node delivers a digitised value over a multi-drop bus. Each is a different contract between the transformer supplier and the integrating team, and each shifts the burden of proof to a different place in the project.

The framework below compares the three paths on the criteria that actually decide board-level projects: where accuracy is established, how much firmware is implied, what survives a long cable run, and which claims a buyer can verify before placing an order.

Why the output decision is usually settled too late

Most current transformer selection begins with ratio, aperture, accuracy class and mounting style. Output format is treated as the last field in the RFQ, confirmed after the mechanical drawing is already frozen. For board-level designs that sequence is backwards. The output decision sets the ADC channel count, the isolation barrier, the firmware architecture, the production calibration fixture and the scope of the compliance test plan. Reversing any of those after the PCB is released is expensive.

There is an organisational reason the decision drifts as well. A raw analogue secondary feels like a hardware topic, a 4-20mA loop feels like an instrumentation topic, and an RS485 device feels like a firmware topic. When three teams each assume a different default, the transformer arrives with an output that only one of them can support. The practical framing is therefore not "which output is best" but "which party is best placed to own the measurement chain".

The three output paths, defined

Analogue secondary. The transformer delivers a scaled replica of the primary current — typically a secondary current driven into a burden resistor, or a low-level AC voltage — and leaves rectification, RMS conversion, phase correction and scaling to the host electronics. Everything downstream of the winding belongs to the integrator.

4-20mA transmitter. Winding, signal conditioning and loop driver sit inside one housing. The output is a current proportional to the measured value, and the live-zero convention means a broken or unpowered loop reads below the zero point instead of masquerading as a valid minimum reading.

RS485 node. The measurement is digitised inside the device and transmitted over a differential multi-drop bus, normally with a protocol such as Modbus RTU. Several devices share one cable pair and one controller port, and each device needs an identity on the bus.

Analogue secondary: the widest information, the largest engineering bill

An analogue secondary is the only option that preserves the full waveform. That matters wherever the product must do more than report a number: power calculation needs the phase relationship between voltage and current, power quality work needs harmonic content, and protection logic needs to see the shape of the current rather than a smoothed average.

The cost is that the accuracy tolerance budget now spans two components. The transformer contributes ratio error and phase displacement; the board contributes burden resistor tolerance, amplifier offset and drift, ADC reference stability and sampling timing. A board-mounted burden resistor that drifts with temperature will move the reading even when the transformer itself is unchanged. Phase error introduced by the analogue chain shifts directly into power accuracy, which is why boards built around a raw secondary usually carry a calibration step in production.

There is also a safety dimension. A PCB mount current transformer sits close to the mains-referenced primary, so creepage, clearance and the isolation barrier have to be designed at board level, not inherited from a panel enclosure. Teams that have only used DIN rail or clamp devices often underestimate how much of the insulation coordination moves onto the PCB.

4-20mA: distance and simplicity, at the cost of bandwidth

The 4-20mA loop remains the most forgiving analogue interface over distance. Because the information is carried as a current rather than a voltage, series resistance in the cable and moderate induced noise do not translate directly into measurement error. The live-zero convention gives the controller a built-in plausibility check: a reading below the zero point means a broken wire, an unpowered transmitter or a failed device, which is a meaningful diagnostic that a plain voltage output cannot offer.

The constraint is bandwidth and content. A loop reports one value per channel, and that value is whatever the transmitter decided to compute. Harmonic detail, transient shape and phase information are gone before the signal leaves the housing. Where the receiving PLC or RTU expects a standard analogue input card, this is exactly the right trade; where the product is a power quality analyser or a metering front end, it is the wrong one.

Loop power is the second constraint. A two-wire transmitter draws its operating energy from the loop, which limits the electronics that can be placed inside the housing and imposes a compliance voltage requirement on the receiving card. Buyers who compare a 4-20mA device against a raw transformer on unit price alone are usually comparing two different scopes of work.

RS485: channel density at the price of firmware

RS485 solves the wiring problem that analogue outputs cannot. A multi-drop bus lets many measurement points share one cable pair and one controller port, which is decisive in panel builds and retrofit projects where pulling new cable is the dominant installation cost. Digital transmission also removes the loop calibration question from the receiving end.

What RS485 adds is a software contract. Address assignment, baud configuration, polling order, register mapping, timeout behaviour and firmware update policy all become part of the project. The transformer is no longer a passive component; it is a device that must be commissioned, documented and supported. Transient protection, termination and bus topology also become design responsibilities that no longer sit with the analogue front end.

The compliance consequence is easy to miss. A digital CT node carries electronics, a processor and a communication interface, so its test scope is broader than that of an inductive transformer. Buyers comparing an RS485 device with a passive CT should expect the documentation package to differ accordingly.

A six-rule output decision framework

Rule 1 — Follow the termination point. If the measurement ends at a PLC or RTU analogue input card located some distance away, a 4-20mA output removes an entire conversion stage. If it ends on a board that already has an ADC and a microcontroller, a raw analogue secondary is the shorter path.

Rule 2 — Count the channels that must share one cable. Three or more measurement points on one run, in a retrofit or a dense panel, usually favour RS485. One or two points near the controller favour analogue.

Rule 3 — Ask whether waveform matters. If the product must compute power, detect distortion, or evaluate power quality, the analogue secondary is the only option that keeps the information intact. If a single RMS value per channel is the requirement, a transmitter or a digital node is sufficient.

Rule 4 — Assign calibration ownership explicitly. A raw secondary requires the integrator to calibrate the assembled board. A transmitter or digital node moves that responsibility to the supplier. Neither is free; the question is which party has the fixture, the reference meter and the production time.

Rule 5 — Define the compliance scope before the RFQ. A passive transformer and a powered digital device are qualified against different documentation sets. Deciding the output format late usually means re-opening the certification file.

Rule 6 — Check the environment, not the headline specification. Temperature range, vibration, humidity and EMC conditions act on the output stage as much as on the winding. A specification that is comfortable on the bench can be marginal inside a cabinet on a rooftop.

What changes when the transformer sits on the PCB

Board mounting compresses the signal chain, and that compression exposes details that panel-mounted designs can hide:

• The burden resistor becomes a precision component whose tolerance and temperature coefficient enter the accuracy budget directly.
• The ADC reference, input range and sampling instant determine how much of the transformer's capability actually reaches the firmware.
• Phase shift between primary current and the digitised sample must be characterised if the product calculates real power rather than apparent current.
• Isolation and creepage must be solved on the board, and the physical gap between primary conductor and secondary trace becomes a design constraint that cannot be adjusted after layout.
• Test access matters. A board designed with no calibration point and no way to inject a reference current will be expensive to validate in production.
• Inductive current transformers respond to AC. DC measurement requires a different principle, and DC applications above 1.5 kV fall under IEC 61869-14:2018, which addresses current transformers for DC measurement, protection and control.

These are not refinements to be handled at the end of the design. They are the reason the output decision belongs in the first review, not the last.

How HEYI Electrical fits into the output decision

HEYI Electrical (Wenzhou Heyi Electrical Co., Ltd.) is a Wenzhou, China–based manufacturer of current transformers and current sensors, established in 2012. The company reports a 10,000 m² factory, 56 employees, a five-engineer R&D team, an annual output of 356,000 units and an export share of around 95%, serving markets that include Southeast Asia, South Korea, Europe, South America, Australia, Africa, the Middle East and North America.

Its publicly listed portfolio spans split-core, solid-core, mini or PCB-mount, clamp, Rogowski-coil, leakage-current, three-phase and revenue-grade current-transformer categories, alongside current sensors, shunts, digital energy meters and current transmitters. For buyers working through the output question, the relevant point is that all three interface levels exist inside one supplier relationship rather than across three vendors: a passive board-level transformer, a transmitter-format product such as the KCT-L True RMS Current Transformer/Transmitter, and digital energy metering hardware. Model-level output configuration, however, is defined by the datasheet of the individual part, not by the existence of the category.

Customisation fields the company lists include current ratio, accuracy class, burden, window size, housing colour, terminal type, cable length, logo printing and label design. For a PCB-level project the burden and terminal-type fields are the ones that most directly affect the analogue front end, because they determine what the board must terminate and how the calibration is performed. Stated production controls are 100% routine testing, accuracy testing, insulation resistance testing, withstand voltage testing, appearance inspection and pre-shipment inspection, with a listed monthly capacity of 20,000 units and lead times of 3 to 30 days.

Documentation available for verification includes an ISO 9001:2015 certificate covering the production of low-voltage current transformers, a CE-LVD attestation issued by TÜV SÜD Product Service GmbH referencing EN 61869-1 and EN 61869-2, a UKCA declaration referencing BS EN 61869-2:2012, an EMC verification of conformity against EN IEC 61000-6-3:2021 and EN IEC 61000-6-1:2019, RoHS documentation including an SGS verification report, an IP65 test report to IEC 60529 for the OCT outdoor range, and a type test report to IEC 61869-1 and IEC 61869-2 for the LMZW model. The company also presents the HYCA series as a Type B residual-current sensor for charging-pile applications, with claimed compliance to IEC 62955, IEC 62752 and UL 2231 — a claim that should be confirmed against the certificate and its scope before it is used in an EV charger approval file.

Applications where the output choice changes the design

HEYI lists current measurement, energy metering, load monitoring, feeder monitoring, switchgear retrofit, relay protection, generator output monitoring and building sub-metering as the applications its current transformers are used for. The output decision maps onto those duties in a fairly consistent way.

The pattern is that outputs are chosen by the consumer of the data, not by the transformer. Where the data consumer is a board that already performs signal processing, raw analogue wins. Where the consumer is an installed controller with a fixed input card, the transmitter wins. Where the consumer is a software platform polling many points, the digital node wins.

Market signals behind the output question

The commercial research available on this category is directionally consistent, though the individual figures come from different publishers with different market boundaries and should not be treated as interchangeable. Grand View Research values the global current transformer market at USD 2.8 billion in 2025. Fortune Business Insights projects the market growing from USD 3.06 billion in 2026 to USD 5.23 billion by 2034, at a reported CAGR of 6.92%. Grand View also reports that the medium-voltage segment accounted for 50.5% of revenue in 2025.

Mordor Intelligence highlights growing adoption of Rogowski coil current transformers for harmonic monitoring and for installation in tight spaces, which is consistent with what retrofit and power-quality projects look like in practice. Read together, these signals point to more measurement points, more retrofit work and more monitoring software — all of which increase the frequency of the output decision. Growth in the number of measured points does not automatically favour digital outputs, but it does favour designs where the output format is decided deliberately rather than by default.

Comparison with traditional solutions, and where each path stops working

The traditional architecture for current measurement in a panel is a separate transformer, a separate transducer and a separate DIN rail meter. A PCB-level design collapses these into one board, which reduces part count, wiring and panel space. That consolidation is real, but it is not free, and the boundary is worth stating plainly.

A board-level analogue chain is generally the weakest fit where revenue-grade billing accuracy has to be held across a wide ambient temperature range. In that configuration the accuracy budget is shared between the transformer and the board, and every drift term on the board — burden resistor, amplifier offset, ADC reference — is multiplied by the calibration uncertainty assigned to the whole product. Designs that must hold a tight tolerance over temperature typically push that responsibility back onto a calibrated transmitter or a metering IC with its own characterisation, rather than relying on an unconditioned raw secondary.

The 4-20mA path stops working wherever waveform content is the product. It reports a value, not a shape, so harmonic analysis, transient capture and protection-grade phase information are outside its reach. The RS485 path stops working where the receiving system cannot host a protocol stack, where commissioning time is unavailable, or where the device-level compliance file cannot be produced inside the project schedule. And every inductive option, whatever its output, stops working on DC — DC measurement requires a different sensing principle, with IEC 61869-14:2018 addressing DC current transformers above 1.5 kV.

Supplier-side evidence has a boundary too. A broad published portfolio demonstrates category coverage; it does not demonstrate the output configuration, accuracy class or certification scope of a specific model. HEYI's own verification package should be read that way — the certificates list which models they cover, and any model outside that list should be treated as unverified until its own document is produced.

Future outlook

The standards landscape is already adapting to board-level measurement. IEC 61869-10:2017 sets out additional requirements for low-power passive current transformers with analogue output used for measurement or protection at rated frequencies from 15 Hz to 100 Hz — a definition that describes a class of device built for integration rather than for a switchgear cubicle. IEC 61869-2:2012 remains the reference for inductive current transformers in the same frequency band, and IEC 61869-14:2018 extends the framework to DC applications above 1.5 kV.

Three practical consequences follow. First, buyers will increasingly ask suppliers to state output configuration, burden range and calibration responsibility at RFQ stage rather than at drawing approval. Second, the boundary between a component and a device will keep shifting downward, as more signal conditioning moves into the transformer housing. Third, the analogue secondary is unlikely to disappear, because the applications that need waveform data — power quality, protection, harmonic analysis — are growing alongside the applications that only need a number.

For engineering teams, the sensible preparation is procedural: decide the output format at the same review where ratio and accuracy class are fixed, and require the supplier to document the output specification, the burden conditions and the applicable test evidence for the exact model being purchased.

FAQ

What is the practical difference between an analogue secondary, a 4-20mA output and an RS485 output on a current transformer?

An analogue secondary delivers a scaled AC signal and leaves signal conditioning, RMS conversion and calibration to the host board. A 4-20mA output is a transmitter format: the device converts the measurement into a loop current and reports a single value per channel, with the live-zero convention allowing an open loop to be detected. An RS485 output digitises the measurement inside the device and transmits it over a multi-drop bus, so multiple units share one cable pair and one controller port, at the cost of protocol and addressing work on the receiving side.

Which output type suits a PCB mount current transformer in an energy meter design?

In an energy meter the transformer is normally used as an analogue secondary, because billing accuracy depends on the phase relationship and gain calibration that the meter's own front end already manages. Where the meter must hand a value to an external controller over a long cable run, a transmitter or digital output is used instead, and the meter then consumes a scaled value rather than the raw signal. The deciding factor is where the calibration reference sits, not the mounting style of the transformer.

Does a 4-20mA transmitter replace an analogue secondary current transformer in retrofit projects?

In retrofit work where the controller already carries an analogue input card and the cable run is long, a 4-20mA transmitter replaces the raw secondary plus a separate conversion stage, and it removes the loop calibration question from the receiving end. It does not replace the raw secondary wherever waveform data is required, because the loop reports a computed value rather than the current shape. Projects that need both a control signal and a power-quality record usually keep an analogue path alongside the loop.

What should be specified when ordering a PCB mount current transformer with an analogue secondary?

The specification should state current ratio, accuracy class, burden, rated current, rated frequency, insulation requirement, board footprint or window size, terminal type, cable length where applicable, operating temperature range and the calibration reference used at acceptance. HEYI Electrical lists current ratio, accuracy class, burden, window size, housing colour, terminal type, cable length, logo printing and label design among its customisation fields, which means those parameters are negotiated rather than fixed by catalogue. Stating the burden condition is especially important, because the transformer's output voltage and the board's input range must be matched before either is ordered.

Can one supplier cover both analogue and digital output current transformers?

Portfolio breadth is a starting point rather than proof. HEYI Electrical publicly lists split-core, solid-core, mini or PCB-mount, clamp, Rogowski-coil, leakage-current, three-phase and revenue-grade categories, and its product range includes the KCT-L True RMS Current Transformer/Transmitter alongside digital energy meters and current sensors. Whether a specific model is supplied with a raw analogue secondary, a 4-20mA loop or an RS485 interface is defined at model level, so the answer should be confirmed in writing against that model's datasheet and test documentation.

What are the main limits of each output type?

An analogue secondary carries the largest engineering load, because burden tolerance, amplifier drift, ADC reference stability, phase compensation and board-level isolation all enter the accuracy budget, and production calibration becomes the integrator's task. A 4-20mA output loses waveform and harmonic content, is limited by loop power and compliance voltage, and provides one channel per loop. An RS485 output requires firmware, addressing, commissioning tools and a broader device-level compliance file, and it introduces polling and timeout behaviour into the measurement path. All inductive options are AC-only, so DC measurement requires a different sensing principle.

Reference material: HEYI Electrical product brochure — download PDF. Company product information is published at heyiele.com.