Technical explainer

Connecting a current transformer: polarity, secondary wiring and commissioning checks

22 September 2026 11 minute read

Getting a current transformer connection right comes down to three things done in order: the primary series connection, correct secondary polarity and wiring — including the shorting block and test switch — and single-point earthing. Commissioning then checks the work with continuity, polarity, ratio, insulation and burden tests before the breaker closes. This article covers the connection and the checks; what a CT is and how burden is defined are covered elsewhere and are only touched on here as needed.

What this covers, and what it does not

A CT's primary winding sits in series with the monitored conductor, and the secondary loop — the relay or meter coil, the lead wires and the CT's own winding — is what electrical engineers call the burden. Both ideas are explained in more depth in what actually drives a CT's price, and the distinction between a current transformer and a voltage-sensing instrument transformer is covered in how a CT differs from a potential transformer. Neither is repeated here. What follows is specific to installing and commissioning the CT itself.

Primary connection — series into the circuit

Two physical arrangements put a CT's primary into the circuit it monitors. A bar-type (or window-type) CT has no primary winding of its own: the busbar or cable being monitored simply passes through the CT's central window, and that conductor is the single-turn primary. A wound-type CT is different — the CT itself carries one or more primary turns wound onto the core, and the conductor being monitored connects to those turns rather than passing through a window.

Either way, the CT sits in series with the monitored conductor, and the impedance it adds is small enough that it does not materially change the current already flowing in that circuit — the CT is there to sense the current, not to influence it.

IEC 61869-2 gives the primary terminals standard designations: P1 and P2. By convention, P1 is the terminal at which primary current is defined to enter. That definition is the reference point everything else in this article checks against — a physical connection, or a drawing's polarity marks, is correct only if it agrees with which terminal is P1.

Secondary polarity — matching the mark to the connection

The same IEC 61869-2 convention names the secondary terminals S1 and S2, with S1 defined as the terminal carrying the instantaneous polarity that corresponds to P1. Get the P1/S1 pairing right and the CT reports current in the direction the protection or metering scheme expects; get it wrong and the sign of every measurement it feeds is inverted.

On a protection drawing, a CT is drawn with its own dedicated polarity-dot symbol, separate from the symbols used for a breaker or a relay. The direction that dot implies is what tells a protective element which current direction it has been wired to sense — it is not decorative.

Polarity matters most where two winding outputs are meant to combine. Connect two windings with opposite relative polarity into a circuit meant to sum or parallel their outputs and the result is not a blended reading — it is a very large circulating current, effectively a short circuit across the two windings. This is a general circuit-theory consequence, demonstrated in laboratory practice on power transformers rather than CTs specifically, but the same principle is exactly what CT paralleling and residual connections rely on getting right: ratio alone is not enough, relative polarity has to match too.

Wiring the secondary loop — shorting block and test switch

A CT secondary circuit cannot be treated like any other instrument wiring, because it cannot tolerate being opened while the primary carries current. If the secondary loop is broken under load, essentially all of the primary-referred current is forced through the core's magnetising branch instead of the external circuit, and the core drives whatever voltage it takes to push that current across the open point.

That voltage can be severe. Laboratory measurements of an open-circuited CT secondary have recorded peaks near 15 kV, against secondary winding insulation typically rated for an emergency peak of only around 3500 V. It is also not a clean sine wave: because flux crosses zero at every current zero-crossing, the open-circuit voltage appears as sharp, repeating pulses rather than a smooth waveform — which means a conventional RMS voltmeter can understate just how dangerous the condition is.

An accidental opening does three separate kinds of damage at once: it exposes hazardous voltage at the open point, it interrupts the current signal reaching the relay (a gap in protection, or a false trip on a differential scheme), and it can leave the core permanently magnetised, degrading its ratio and phase accuracy even after the circuit is reclosed.

The standard answer is the shorting terminal block and test switch fitted in the CT secondary circuit. A properly built test switch shorts the CT secondary before it breaks the path onward to the relay — a make-before-break sequence on its current-shorting poles — so isolating or testing a relay never leaves the CT open, even momentarily.

That guarantee is design-dependent, not universal. Comparative testing across several test-switch models recorded measurable voltage spikes during the operate cycle on some of them, meaning the secondary was briefly opened despite the switch being built for make-before-break operation. The practical conclusion is not to assume the switch protects the circuit — verify it, or at minimum know which model is fitted.

Earthing the secondary — one point, not two

Grounding a CT's secondary circuit serves two distinct purposes: it protects the equipment wired into that circuit, and it protects anyone who might come into contact with it.

It has to be done at exactly one point, and the reason is not caution for its own sake. While fault current is flowing through a station's ground grid, the grid's potential is not uniform from point to point. A secondary circuit grounded at two physical locations therefore sits across two points that are not at the same voltage during a fault, and the difference drives an unintended current through the circuit that was never meant to carry one.

Guidance is specific about where that single ground point should be: at the switchboard or relay panel — the first point where the circuit is actually used — rather than at the CT itself or at more than one location. The same single-point rule applies to spare, currently-unused secondary windings on a multi-tap or multi-core CT: they still need to be grounded, at one point, even though nothing is reading them yet.

Wiring several CTs to one relay — what it does to burden

At breaker-and-a-half or ring-bus line terminals, one protected line is often monitored by more CTs than the relay has current inputs. The traditional answer is to parallel those CTs' secondaries ahead of the relay, so a single input carries the summed current from more than one CT.

That convenience has a cost in security. Paralleled CTs are inherently less secure on an external (through) fault than CTs individually wired to their own relay inputs, because if the paralleled CTs saturate unequally, the mismatch produces a spurious differential current the relay cannot distinguish from a genuine internal fault. Documented cases record two real relay misoperations traced to exactly this mechanism.

Modern relay designs sidestep the problem by giving each CT its own dedicated current input rather than requiring external paralleling. The relay can then measure each CT's actual individual contribution and apply restraint logic accordingly, which is more secure than summing the CTs before the relay ever sees them.

Where paralleling is already in place, testing needs extra care: isolating one CT's contribution at a multi-CT terminal must not disturb the in-service contribution of the other CTs feeding the same summing point. And because burden adds as the loop's series impedances add — the same relationship that governs a single CT's loop — paralleling more circuits onto one secondary changes the burden picture the CT actually sees, which is the subject of the worked example below.

Residual and core-balance connections for earth-fault protection

A high-impedance bus differential scheme is the clearest illustration of paralleling secondaries at one summing point: every CT surrounding the protected bus zone has its secondary connected to a shared point, and for the sum to mean anything, every one of those CTs has to share the same transformation ratio.

If the ratios do not match exactly — and under normal accuracy tolerances a CT's ratio error alone can run to several percent — a spurious differential current appears at the relay even with no fault anywhere in the protected zone, sized by how far the mismatched CTs' errors sit from cancelling out.

A low-impedance, numerically compensated scheme avoids the requirement altogether: each CT gets its own dedicated input on a microprocessor relay, which digitally scales for CTs of different ratios and computes the differential current internally, rather than needing every CT around the zone built to one physical ratio.

Core-balance (zero-sequence) CTs, which route every phase conductor of a circuit through one ring-type CT to sense earth-fault current directly, are worth understanding conceptually at this stage but are not covered here step by step — no citable source describes the physical routing and bonding practice in installation detail, so this article stops at what the connection is and why it matters rather than improvising mechanics. What is documented is the commissioning difficulty: a residual or polarising connection is one of the hardest wiring errors to catch, because standard secondary-injection tests inject current straight into the CT windings and never exercise the actual primary wiring path, and under normal system conditions there is no zero-sequence signal present to expose a miswired polarising connection during a routine check.

Commissioning checks before energising

Commissioning answers a single question: how do you know the connection is right before the breaker closes? A recognised minimum test set covers class, polarity, ratio, an excitation test and insulation checks, and the same grouping — insulation resistance, ratio, polarity, winding and lead resistance, excitation, and burden — is what IEEE C57.13.1 uses to categorise field testing of relaying CTs. No specific ohmic insulation-resistance threshold is stated here; that figure is not attributed reliably enough to the standard's own text to publish as a requirement.

The table below carries that test set forward as a working checklist.

Worked example — burden across two lead cross-sections

Take one CT with a 5 A secondary and a rated secondary burden of 15 VA, feeding a meter that draws 2.5 VA and a test-switch block that adds 0.5 VA — 3 VA of fixed device burden regardless of the leads. The CT sits 30 m from the panel, so the lead run is 60 m out and back.

Lead resistance follows R = ρL / A, using copper resistivity ρ ≈ 0.0175 Ω·mm²/m, and the lead's VA contribution at rated secondary current is I²R.

2.5 mm² leads

• R = 0.0175 × 60 / 2.5 = 0.42 Ω
• Lead burden = 5² × 0.42 = 10.50 VA
• Total burden = 10.50 + 3.00 = 13.50 VA, against a 15 VA rating — inside the rating, with 1.50 VA (10%) of margin.

4 mm² leads

• R = 0.0175 × 60 / 4 = 0.2625 Ω
• Lead burden = 5² × 0.2625 = 6.56 VA
• Total burden = 6.56 + 3.00 = 9.56 VA, against the same 15 VA rating — 5.44 VA of margin, nearly four times the headroom of the thinner lead for the same run and ratio.

The device and test-switch burden do not change between the two cases; the entire difference is the lead's own resistance. On a long run, the cross-section of the secondary wiring is not a minor spec decision — it is the difference between a loop that clears rated burden comfortably and one that clears it by 10% and leaves no room for a test load or a future addition to the circuit.

Questions buyers ask

Which terminal is CT primary polarity, P1 or P2?

Under IEC 61869-2, P1 is defined as the terminal current is deemed to enter. S1 on the secondary carries the corresponding instantaneous polarity, and that P1/S1 pairing is what a physical connection or a drawing's polarity mark is checked against.

Can a CT secondary be left open while the primary is energised?

No. An open secondary forces primary-referred current through the core's magnetising branch, producing a dangerous voltage pulse, a gap or false trip in the protection signal, and possible permanent core magnetisation. The shorting block and test switch exist specifically to prevent this.

Why does a CT secondary need exactly one ground, not more?

Because station ground potential is not uniform during a fault, two ground points on the same secondary circuit sit at different voltages while fault current flows, and the difference drives an unintended current through the circuit. Guidance recommends one ground point, at the switchboard or relay panel.

What does a CT commissioning test actually cover?

Continuity and wiring verification, polarity, ratio, insulation, and burden, plus confirming that any unused CTs are shorted and grounded before the equipment is released for service.

Does paralleling several CTs into one relay affect accuracy?

Yes. Unequal saturation between paralleled CTs, or ratio errors that do not cancel, can produce a spurious differential current at the relay even when there is no fault in the protected zone.