Technical explainer

Current transformer against potential transformer: what the difference means in practice

12 August 2026 11 minute read

A current transformer measures line current and is wired in series with the conductor. A potential transformer — also called a voltage transformer — measures bus voltage and is wired in parallel across it. That single difference in what each device is excited by is why every other rule runs in the opposite direction for one against the other: connection, burden, accuracy, failure mode.

What each device measures, and why the winding is wired the opposite way

A CT is excited by a current source, not a voltage source. Its primary winding sits effectively in series with the line, and the CT's own series impedance is negligible against the impedance of the line itself. The practical consequence is that a CT's primary current is fixed by the load or the system, not by the CT — the CT has no say in how much current flows through it, only in how that current is represented on its secondary.

CTs and PTs share a common purpose despite measuring different quantities. Both scale a primary quantity down to a standardised, low-level signal that a protection relay or a meter can handle, and both electrically isolate the relaying or metering circuit from the primary power apparatus — a safety function for personnel and equipment as much as a metrology one.

In a three-phase system, each phase typically needs its own CT and its own PT. Voltage can also be taken phase-to-phase, directly, using two identical PTs and reading the difference between their two secondary voltages, rather than one PT per phase-to-neutral. It is a distinct, valid connection scheme, and worth recognising when reading a single-line diagram rather than assuming every voltage measurement is phase-to-neutral.

The two connection rules that catch people out

A CT secondary must never be opened under load

If a CT's secondary circuit is opened while the primary is still carrying current, the primary current that would normally be balanced by secondary current instead flows entirely through the magnetising branch. That drives the core into deep saturation and induces a dangerously high voltage across the open secondary terminals — a risk to winding insulation and to anyone touching the terminals.

The mechanism behind this is not just the symptom of saturation but a structural change in the CT's own equivalent circuit. As the core saturates, the magnetising impedance collapses; the core starts behaving like an air core, and the primary-to-secondary coupling that made the CT work in the first place is lost. The CT can no longer faithfully reflect the primary current to a now-disconnected secondary, and the energy has to go somewhere — into a voltage spike across the open terminals.

This is not the only secondary-circuit hazard worth planning for. A related but distinct concern is an ungrounded secondary accumulating charge rather than an open one saturating the core. The IEEE Power System Relaying Committee's dedicated guide on grounding instrument-transformer secondary circuits addresses that case, setting out single-point grounding practice and a minimum grounding-conductor size of 12 AWG copper under the US National Electrical Safety Code. It is a separate rule from the open-secondary one above, not evidence for it, and both belong in a secondary-circuit safety checklist.

In service, the open-secondary rule is enforced with a dedicated piece of test equipment: a make-before-break short-circuit switch shorts the CT secondary before a meter or relay is disconnected from the loop. That practice exists because a CT's normal, safe operating state is close to a short circuit, not away from one. A related fact makes the same point from a different angle — leaving a CT's factory-fitted secondary shorting shunt in place after installation does not stop the CT from working, but it does reduce the metered reading by roughly 50 to 80%. Low secondary-loop impedance is the CT's normal condition; any increase toward an open circuit is what degrades or endangers it.

A PT secondary must never be short-circuited

The opposite winding produces the opposite rule. A PT is a voltage source referred to the secondary, in the same equivalent-circuit family as an ordinary power transformer. A short-circuited PT secondary forces near-zero impedance onto a device that is still trying to hold its rated secondary EMF, driving secondary current to many times its rated value and risking transformer burn-out unless a fuse or breaker intervenes.

The enforcement tool mirrors the CT case but does the opposite thing: a separate potential switch is used to remove or isolate voltage from a PT-fed meter without ever shorting the PT secondary. Two devices, two opposite jobs, matching the two opposite failure modes.

Side by side

Burden and accuracy class, expressed differently for each

Burden is the impedance that everything electrically connected to a secondary winding presents to it — meter or relay input impedance, lead-wire resistance, test-switch and connection resistance. Exceed the transformer's rated burden and the resulting metering or protection error exceeds the transformer's stated accuracy class.

For a CT, secondary voltage is a direct consequence of burden. By the transformer EMF equation, the burden impedance sets the secondary voltage, and that secondary voltage sets the core flux; above the core's knee point, the CT saturates.

The units used to express burden split in practice. For metering-grade CTs, burden is customarily expressed in ohms of secondary-loop impedance — the higher the permissible ohms rating relative to the actual loop impedance, the better the resulting accuracy. For protection-grade CTs, burden is customarily expressed in volt-amperes at rated secondary current instead.

A PT's burden is expressed differently again. Under IEC 61869-3, standard rated burden values are fixed VA figures — 10, 25, 50, 100, 200 and 500 VA, for example. Unlike a CT, which is rated at a single fixed operating burden, a PT's accuracy class is defined over a burden range, typically 25 to 100% of rated burden. A PT has to hold its accuracy across a spread of connected load, not just at one design point.

What a CT enquiry must state: ratio, class and rating factor

A CT's ratio is conventionally referenced to a standardised secondary current — most commonly 5 A, with 1 A the other common standard — at rated primary current. A "400:5" ratio means 400 A on the primary produces 5 A on the secondary at rated conditions. This is distinct from the turns ratio whenever the primary has more than one physical turn.

US practice under ANSI/IEEE also divides CTs into two construction classes. Class T, the wound type, has non-negligible core leakage flux, so its performance can only be established by test — standard curves do not apply. Class C, the bar type, has negligible leakage flux, so its performance can be read from standard excitation curves published against IEEE C37.110. A rating such as "C100" states the secondary voltage up to which linear response is guaranteed for a given burden.

Rating factor is a separate figure again: the multiple of rated primary current the CT can carry while still meeting its certified accuracy class, defined at a stated ambient temperature — commonly 30 °C for CTs in open air, 55 °C enclosed, or 85 °C in pad-mounted equipment. Ratio, class and rating factor are three different questions, and an enquiry needs an answer to each.

These are also the specifications a quotation is actually built from, and they are what drives a current transformer's price — the companion piece covers how ratio, accuracy class and rating factor translate into cost, rather than repeating that ground here.

Wiring mistakes that quietly corrupt a reading

Correct CT polarity connection — matching the marked polarity terminal, the conventional H1-to-X1 marking — is required for correct metering and protective operation. Reversed polarity produces grossly incorrect metering, and when the circuit's power factor is not zero, the error is not obviously visible from the readings alone. A meter can display a plausible-looking number that is simply wrong.

The PT case has a parallel trap. Two identical PTs can measure line-to-line voltage directly, by taking the difference of their two secondary voltages. If the two PTs used are not identical in ratio or phase-angle error, the resulting measurement is distorted in a way a single phase-to-neutral PT measurement never exposes, because there is no second device's error to cancel against or reveal the discrepancy.

The common thread in both cases is that the error is invisible on the meter face. Specifying and verifying polarity, and confirming a matched PT pair, at commissioning is cheaper than chasing a billing dispute or a protection mis-operation later.

Accuracy limit factor against voltage factor

Ratio error is the percentage deviation between the secondary current scaled by the nominal turns ratio and the actual secondary current. While the CT is unsaturated, this error is essentially the magnetising current expressed as a percentage of secondary current.

Accuracy class itself is not a single figure that holds at every current. Under IEC 60044-1, a Class 1.0 CT is permitted roughly ±1.0% ratio error at 100% of rated primary current, widening to about ±1.5% at 20% of rated current and about ±3.0% at 5% of rated current. The same accuracy class only means what it says at the current level it was defined for, and gets worse as load falls away from rated current — worth checking against the actual load range an enquiry is being written for, not just the class number on its own.

This is where protection-grade and metering-grade CTs diverge sharply. A metering-grade CT need only stay accurate over roughly 5 to 125% of rated current, the everyday range a load actually occupies. A protection-grade CT must instead reproduce current linearly up to a much higher multiple, because its job is representing fault current to a relay, not everyday load current.

Accuracy limit factor, defined under IEC 61869-2, is the ratio between a protection CT's rated accuracy-limit primary current and its rated primary current — the multiple of rated current up to which the CT is guaranteed to stay within its stated composite-error limit before significant core saturation sets in. The labelling convention states both figures together: "5P10" means a maximum 5% composite error at 10 times rated current; "10P15" means a maximum 10% composite error at 15 times rated current. This is structurally different from a metering accuracy class, which is a single number — 0.2, 0.5 — valid at or below rated current, not a multiple-of-rated-current statement.

Voltage factor, defined under IEC 61869-3, is the PT's parallel concept, but for a different physical reason. It states the highest continuous multiple of rated primary voltage a PT must survive, for a specified rated time, without exceeding its accuracy or thermal limits. The multiple required depends on the primary system's earthing arrangement: on a system that is not solidly earthed, a single line-to-earth fault raises the voltage on the healthy phases toward line-to-line voltage until the fault clears, and the PT on those phases must survive that overvoltage for the fault's duration.

Stated plainly, the two are structurally parallel but physically different headroom multipliers. Accuracy limit factor is current-based and triggered by fault magnitude. Voltage factor is voltage-based and triggered by earthing arrangement and fault duration. Both answer the same question — how far above rated, and for how long, must this device keep working — but the axis and the trigger differ.

Worked example: does a CT stay within its class through a fault?

Take a 600:5 protection CT rated C200. Under the Class C convention, its secondary voltage stays within its accuracy certification up to 200 V at 20 times rated secondary current. The installed secondary loop — relay coil, lead run, test-block contact resistance — measures 0.85 Ω.

At rated primary current, 600 A primary gives 5 A secondary, and secondary voltage is V₂ = I₂ × Z = 5 A × 0.85 Ω = 4.25 V. That is far below the 200 V ceiling, so there is no saturation risk at normal load current.

At the accuracy-limit condition — 20 times rated secondary current, or 100 A, standing in for a fault at 20 times rated primary current — secondary voltage is 100 A × 0.85 Ω = 85 V. Still under the 200 V ceiling, so this CT stays within its C200 certification through the fault and continues to represent it linearly to the relay.

Contrast case. If the same loop instead measured 2.6 Ω — a longer lead run, or an added test switch in the circuit — the fault-condition figure becomes 100 A × 2.6 Ω = 260 V, above the 200 V ceiling. This CT would saturate before reaching 20 times rated current and would under-represent the fault to the relay, even though its nameplate ratio and class are unchanged. A C-class rating is only valid at or below the burden it was calculated for. The nameplate does not travel with a change in wiring.

Where a current transducer differs from a current transformer

A current transducer does not rely purely on magnetic induction the way a wound CT does. Two example principles illustrate the range. A Rogowski coil is an air-core coil whose output voltage is proportional to the rate of change of the primary current, which means the waveform has to be integrated to recover the actual current signal. A magneto-optic current transducer instead detects the Faraday-effect rotation of polarised light travelling through a fibre-optic loop around the conductor, a rotation proportional to the magnetic field and hence to current.

The Rogowski coil's air core, with no ferromagnetic material in it, is the structural difference from a wound CT's iron core that makes that rate-of-change output possible in the first place. Both principles are presented in the literature as non-contact alternatives to a conventional current transformer.

Set against that, a conventional CT's isolation is inherent, because the coupling is purely magnetic — this, together with signal-level scaling, is one of the two reasons an instrument transformer is used at all. A transducer changes how the measurement is taken, but galvanic isolation is not automatically preserved just because the function looks similar on a single-line diagram. A buyer specifying "current transducer" in place of "current transformer" is changing the sensing principle, not just the connection method, and it is worth confirming isolation requirements explicitly rather than assuming interchangeability.

Questions buyers ask

Can one instrument transformer serve both protection and metering?

Not to the same standard on both counts. A protection-grade CT is built to stay linear up to a high multiple of rated current, its accuracy limit factor, so it can represent fault current to a relay. A metering-grade CT is optimised for accuracy near rated load current, roughly 5 to 125% of it, and is not required to behave the same way during a fault.

Why must a CT secondary never be left open, even briefly?

With the primary still energised, an open secondary forces all the would-be secondary current through the magnetising branch, driving the core into deep saturation and inducing a high voltage across the open terminals — a risk to winding insulation and to anyone touching the terminals. This is why test practice uses a short-circuit switch before disconnecting a meter or relay.

What is the difference between rating factor and accuracy limit factor?

Rating factor is a thermal and accuracy multiplier on continuous current, defined at a stated ambient temperature. Accuracy limit factor, under IEC 61869-2, is specific to protection-class CTs and states the multiple of rated current up to which the CT stays within its composite-error limit during a fault. They answer different questions and both belong on an enquiry.

Why is a PT's burden a fixed VA figure rather than an ohms rating?

Because IEC 61869-3 defines PT rated burden as fixed VA values — 10, 25, 50, 100, 200, 500 VA — and requires the accuracy class to hold across a burden range, typically 25 to 100% of rated burden. That is a different scheme from a CT's ohms-or-VA figure tied to a single operating point.