Transformers · study.

Coupled coils and mutual inductance

A transformer contains two electrically separate windings linked by a shared magnetic path. The primary winding connects to the source. The secondary winding connects to a load, an instrument, or another circuit. Conductive isolation is retained between the windings; energy crosses the insulation through the time-varying magnetic field in the core. The arrangement requires a changing current or changing magnetic flux. A steady DC current establishes a static field after its switching transient and induces no continuing secondary voltage.

Each winding has self-inductance. A changing current in winding 1 produces a voltage in that same winding and a voltage in winding 2. With selected terminal references, the coupled-winding equations are

v_{1} = L_{1} \frac{d i _{1}}{d t} + M \frac{d i _{2}}{d t}, v_{2} = M \frac{d i _{1}}{d t} + L_{2} \frac{d i _{2}}{d t} .

The self terms $L_{1}$ and $L_{2}$ depend on turn count, geometry, and core permeability. The mutual inductance $M$ measures the flux linkage shared by the two windings. Its algebraic sign follows the chosen terminal references. Geometry alone does not give a signed $M$ until the reference directions have been stated. The dot convention records that sign on the circuit drawing and prevents a common error in coupled-circuit work: combining induced voltages with the wrong polarity.

Two windings share a high-permeability core. A changing primary current drives circulating core flux

Φ

that links both windings, while the copper circuits stay electrically isolated.

Faraday's law gives a direct physical definition. Let $Φ_{21}$ be the part of the magnetic flux produced by winding 1 that links winding 2, and let $N_{2}$ be the number of secondary turns. The induced voltage due to the primary current is

v_{2 \leftarrow 1} = \frac{d ( N _{2} Φ _{21} )}{d t} = M \frac{d i _{1}}{d t},

where the sign is included in the oriented flux linkage. A larger turn count raises both flux linkage and induced voltage. Tighter core coupling raises the fraction of the primary-produced flux that reaches the secondary. Air gaps, separated coils, fringing, and a core operated near saturation all reduce the ideal shared-path model.

Mutual inductance has a magnitude bound:

∣ M ∣ \leq L_{1} L_{2} .

The dimensionless coupling coefficient is conventionally written as

k = \frac{∣ M ∣}{L _{1} L _{2}}, 0 \leq k \leq 1.

An ideal common-core transformer is modeled with $k$ very close to one. The equality case says that all magnetic energy associated with either winding shares the same mode. It does not imply zero winding resistance, zero core loss, or unlimited power. Those requirements enter the transformer model separately. Small signal transformers, power transformers, RF transformers, and current transformers can all have a large $k$ while differing strongly in loss, voltage rating, and frequency range.

Reference directions and dots

Place a dot at one end of each winding. If a current enters the dotted terminal of winding 1 and rises, the induced voltage makes the dotted terminal of winding 2 positive. Equivalently, a rising current entering one dotted terminal produces a positive mutual-voltage term when the voltage reference on the other winding is from its dotted terminal to its undotted terminal. In that reference system the mutual terms in the two equations are $+ M d i / d t$ .

Dot convention. With both voltage references running from the dotted to the undotted terminal, a rising current

i_{1}

entering the primary dot makes the secondary dot positive, so both mutual terms are

+ M d i / d t

Changing one reference reverses one sign. A current arrow entering an undotted terminal is equivalent to a current leaving the dotted terminal. A voltage measured from undotted to dotted is the negative of the dot-to-undotted voltage. Neither choice is wrong. The error appears when a diagram uses one choice and the equations silently use another. Mark current arrows, voltage polarities, and dots before using mesh equations or phasors.

For sinusoidal steady state, derivatives become multiplication by $j ω$ in a phasor description. A coupled pair can therefore be written

V_{1} = j ω L_{1} I_{1} + j ω M I_{2}, V_{2} = j ω M I_{1} + j ω L_{2} I_{2},

when both current arrows enter dots and both voltages are dot-to-undotted. If one terminal convention is reversed, replace the relevant mutual sign. This compact form is valuable for finite-coupling circuits, tuned transformers, and leakage-aware models. The ideal-transformer relations developed below impose stronger assumptions than these general coupled-inductor equations.

Ideal transformer relations

The ideal transformer model separates a high-permeability shared core from all loss and leakage mechanisms. It assumes that every turn on either winding links the same core flux $Φ (t)$ , that winding voltage drops are negligible, and that the core requires negligible magnetizing current. These assumptions give a ratio element. They do not state that the secondary voltage exists independently of the primary source or that the output can supply arbitrary current.

A winding with $N$ turns has terminal-voltage magnitude and sign set by Faraday's law through the oriented flux linkage:

v = N \frac{d Φ}{d t} .

The sign convention used here treats the winding terminal voltage as the applied voltage balancing the induced back emf. Applying the same orientation on both windings gives

\frac{v _{2} ( t )}{v _{1} ( t )} = \frac{N _{2}}{N _{1}} = a, \frac{V _{2, rms}}{V _{1, rms}} = a .

The turns ratio $a = N_{2} / N_{1}$ sets voltage scaling. A step-up transformer has $a > 1$ . A step-down transformer has $a < 1$ . The waveform shape transfers with the ratio while the core remains in its linear operating region. A sinusoidal primary produces a sinusoidal secondary at the same frequency. A distorted source drives a proportionally distorted secondary voltage in the ideal model.

Ideal voltage scaling. Both windings link the same core flux

Φ (t)

, so each terminal voltage is proportional to its turn count:

V_{2} / V_{1} = N_{2} / N_{1}

. Here the secondary has more turns, so the connection steps the voltage up.

The instantaneous relation is stronger than an RMS ratio. If $N_{2} / N_{1} = 0.10$ and the primary is a 120 V RMS sine wave, the unloaded secondary is 12 V RMS in the ideal model. Its peak voltage is also one-tenth of the primary peak voltage. Confusing an RMS rating with a peak value changes a turns calculation by a factor of $2$ . Use one amplitude convention throughout a calculation. Voltage ratings on ordinary power transformers are almost always RMS values at a stated frequency and load.

With an open secondary, an actual primary still draws a small current. This is the magnetizing current $I_{m}$ . It establishes the alternating core flux required by the applied voltage. In a good power transformer it is generally much smaller than the rated load current, though it can be large enough to matter in a lightly loaded system. The ideal-ratio model sets it to zero in the load-current balance but does not deny that a real core needs magnetic field strength.

Under sinusoidal voltage, the idealized magnetizing branch is inductive. Its current lags the applied voltage by nearly a quarter cycle. Core hysteresis and eddy losses add an in-phase component, so a real no-load current is neither zero nor perfectly quadrature. A wattmeter reading on an unloaded transformer therefore measures core loss plus a small winding loss, even when the secondary terminals are unconnected.

Loading the secondary changes the primary current. A load current $I_{2}$ produces a magnetomotive force $N_{2} I_{2}$ whose core effect opposes the original source-driven flux. The source voltage constrains the required rate of change of core flux. The primary responds by drawing an additional load component $I_{1}^{'}$ such that, under the ideal approximation,

N_{1} I_{1}^{'} + N_{2} I_{2} = 0.

Thus the current ratio is

\frac{I _{2}}{I _{1}^{'}} = - \frac{N _{1}}{N _{2}} = - \frac{1}{a} .

The minus sign is the winding-reference sign. It records opposing ampere-turns, not a claim that input and output powers have opposite physical directions. Under passive load references, the source delivers positive average power to the primary terminals and the secondary load absorbs positive average power.

Ampere-turn balance under load. Secondary current

I_{2}

sets up an opposing magnetomotive force

N_{2} I_{2}

; the primary draws an added load component (shown here as

I_{1}

) so that

N_{1} I_{1}

balances

N_{2} I_{2}

and the required core flux is restored.

The complete primary current is the phasor sum of its magnetizing part and its reflected load part. At rated load, the latter is commonly much larger. A current probe on the primary will therefore show an increased current when a secondary load is connected, even though no conductive path joins the two circuits. This observation is a direct test of energy transfer by mutual induction and of the ampere-turn balance.

Reflected impedance and source loading

The transformer ratio converts a secondary load into an input load seen by the source. This result is often more practical than the separate voltage and current ratios. It allows a primary-side circuit to be analyzed without redrawing the secondary network. The reflected impedance depends on the square of the turns ratio, a consequence of one factor from the voltage ratio and another from the current ratio.

Use two different secondary-current symbols while deriving the result. Let $I_{2}$ enter the dotted secondary terminal, as required by the coupled-winding sign convention. Let $I_{L}$ leave that terminal into the external load. They are related by

I_{2} = - I_{L} .

In an ideal transformer, the ampere-turn equation becomes

N_{1} I_{1} - N_{2} I_{L} = 0, I_{1} = a I_{L},

where $a = N_{2} / N_{1}$ . The voltage relation is $V_{2} = a V_{1}$ . If the secondary load has impedance $Z_{L} = V_{2} / I_{L}$ , then

Z_{in} = \frac{V _{1}}{I _{1}} = \frac{V _{2} / a}{a I _{L}} = (\frac{N _{1}}{N _{2}})^{2} Z_{L} = \frac{Z _{L}}{a ^{2}} .

The equation applies to resistance, inductive impedance, capacitive impedance, or a network containing all three. The ideal transformer preserves the impedance angle while changing its magnitude by a positive real factor. If a secondary resistor is connected, the input is a resistor of the reflected value. If a secondary network has a positive imaginary part, the reflected network also has a positive imaginary part. The transformer does not reverse capacitive or inductive character.

Reflected impedance. A secondary load

Z_{L}

appears at the source terminals as

Z_{L} / a^{2}

with

a = N_{2} / N_{1}

, so a primary-side analysis can use the single reflected element without redrawing the secondary network.

The reflected-impedance rule should be used on a fully specified two-port drawing. Voltage polarity, the turns ratio direction, and which side is being reflected must be clear. Defining $a = N_{2} / N_{1}$ gives $Z_{in} = Z_{L} / a^{2}$ when the load lies on side 2 and the input lies on side 1. Reflecting a source-side impedance to side 2 multiplies it by $a^{2}$ . Reversing the naming of the windings without changing the formula is a common source of factors inverted by four or five orders of magnitude.

Power relation at an ideal ratio element

The ideal transformer has no independent source of energy. In phasor form, with passive terminal conventions selected on the load side, the input and output complex powers have equal magnitudes and opposite port signs. For real power delivered from the primary source to a secondary load,

P_{in} = V_{1, rms} I_{1, rms} cos ϕ_{1} = V_{2, rms} I_{L, rms} cos ϕ_{L} = P_{L}

in the lossless approximation. The primary load-current component inherits the load power factor when it is reflected through an ideal transformer. The magnetizing component adds a largely reactive no-load current in a physical transformer. Its presence changes the actual primary current without changing the current-ratio rule for the reflected load component.

A resistive secondary load permits a power relation without a phase angle. Combining $V_{2} = a V_{1}$ with $I_{1} = a I_{L}$ gives

V_{1} I_{1} = V_{1} (a I_{L}) = (a V_{1}) I_{L} = V_{2} I_{L} .

A step-down transformer decreases voltage and increases available secondary current by reciprocal factors. A step-up transformer increases voltage and decreases current. The phrase increases current needs the load condition. An open secondary has no load current, regardless of its turns ratio. A transformer cannot force a stated current into an arbitrary load because its winding resistance, leakage reactance, core limits, and source impedance constrain the operating point.

An ideal transformer can isolate a load from DC continuity, but it cannot transfer a DC power waveform through steady mutual induction. A rectangular drive transfers energy during its rising and falling intervals. The core flux changes during each nonzero applied-voltage interval, then remains at its latest value while the voltage is zero. Repeated unipolar pulses accumulate core flux and eventually drive a finite core into saturation. Power-conversion circuits use alternating polarity, controlled reset intervals, or a gapped magnetic design to keep the core within an allowed flux range.

Impedance matching

Many low-frequency power transformers are chosen for voltage conversion, whereas small-signal transformers are often chosen to place a source and load at compatible impedance levels. A source with a significant internal resistance delivers its maximum available sinusoidal power to a purely resistive reflected load equal to that resistance. If a load resistor $R_{L}$ is connected to secondary turns $N_{2}$ , select primary turns $N_{1}$ so that

R_{s} = (\frac{N _{1}}{N _{2}})^{2} R_{L}, \frac{N _{1}}{N _{2}} = \frac{R _{s}}{R _{L}} .

This condition describes a particular source model and a particular objective. It does not maximize efficiency when source resistance represents an unwanted loss. In a power supply, a low source resistance and a load that meets its voltage requirement are usually preferred. In a measurement system, input impedance, bandwidth, noise, phase error, and isolation may take priority over maximum transferred power.

Impedance matching. Choosing the turns ratio so that

R_{s} = (N_{1} / N_{2})^{2} R_{L}

makes the reflected load equal the source resistance, delivering the source's maximum available sinusoidal power without a direct conductive connection.

When an impedance has resistance and reactance, transform the entire complex value:

Z_{L} = R_{L} + j X_{L} ⟹ Z_{in} = \frac{R _{L}}{a ^{2}} + j \frac{X _{L}}{a ^{2}} .

The ratio therefore scales both real and imaginary components equally. A transformer does not tune a reactive load to resonance. It can place the required component values or source impedance in a more convenient range, but any cancellation of reactance must come from the surrounding circuit and its frequency dependence.

Polarity matters when transformer windings are connected in series. Two equal secondary windings can produce a doubled voltage when their marked polarities are series-aiding. Reversing one connection produces series opposition and a near-zero terminal voltage for equal voltages. Manufacturers mark a start lead, dot, or lead number to preserve this information. Verify it at low voltage before connecting windings in a higher-energy circuit.

Nonideal transformer behavior

The ideal model is a reference point, not a complete device model. A physical transformer contains resistance in both windings, finite magnetic permeability, flux that links only one winding, and a core with loss. At higher frequency it also contains distributed capacitance between turns, windings, and core. These effects determine the voltage regulation, no-load current, heat generation, frequency response, and maximum safe operating voltage.

A standard low-frequency equivalent circuit places primary winding resistance $R_{1}$ and primary leakage reactance $X_{σ 1}$ in series with the source. The core is represented by a shunt magnetizing reactance $X_{m}$ in parallel with a resistance $R_{c}$ . The resistance $R_{c}$ accounts for core-loss power at the applied voltage. Secondary winding resistance $R_{2}$ and secondary leakage reactance $X_{σ 2}$ lie in series with the load. Values on one side can be reflected to the other side using the square turns rule, reducing the model to a single circuit with no ideal-ratio symbol.

Low-frequency equivalent circuit referred to the primary. Series

R_{1}

and leakage reactance

X_{σ 1}

carry the load current; the shunt branch (core-loss resistance

R_{c}

in parallel with magnetizing reactance

X_{m}

) carries the no-load current, ahead of the ideal turns ratio

a

The circuit is an approximation with a frequency range and operating range. It assumes sinusoidal operation near the chosen test frequency and a core below severe saturation. At audio, switching, or radio frequencies, extra distributed elements may be needed. At high loading, winding temperature changes the resistances and the operating voltage drop. Engineering data sheets commonly report a regulation percentage, no-load current, short-circuit impedance, loss at a stated temperature, and a rated VA value instead of every lumped model parameter.

Winding resistance and copper loss

Each turn has resistance. Current produces an ohmic voltage drop and a heat rate

P_{cu, 1} = I_{1, rms}^{2} R_{1}, P_{cu, 2} = I_{2, rms}^{2} R_{2} .

At a fixed load current, this loss rises with the square of current. It therefore grows quickly as the transformer approaches overload. Resistance also reduces the terminal voltage under load. The induced secondary emf exists inside the winding, while the load sees that emf after subtracting the drops across secondary resistance and leakage reactance. The primary series terms similarly require a larger source voltage to hold the intended core flux.

Copper resistance rises with temperature. For ordinary copper near room temperature, the fractional resistance rise is about $0.0039$ per degree Celsius. A transformer tested while cool will therefore have a smaller winding loss than the same transformer after a long rated-load run. Ratings use a permitted temperature rise, insulation class, ambient temperature, enclosure, and cooling arrangement. A winding current below its headline rating can still be unsafe when airflow is blocked or ambient temperature is high.

Phasor construction of the terminal voltage. The induced secondary emf

E_{2}

loses the resistive drop

I_{2} R

and the leakage-reactance drop

I_{2} X_{s}

(a quadrature phasor), leaving a terminal voltage

V_{2}

below

E_{2}

for this lagging load.

High-frequency transformers face an additional copper effect. Alternating current crowds toward the surface of a conductor and toward regions influenced by nearby conductors. The resulting AC resistance can exceed the DC resistance. Foil windings, litz wire, interleaved layers, and conductor geometry are selected to limit those effects within a design band. A DC resistance measurement alone cannot predict full-load loss at a much higher operating frequency.

Leakage inductance

Some magnetic field closes through air and links one winding more strongly than the other. This portion is represented as leakage inductance. Its reactance increases with frequency and appears in series with the winding. It limits short-circuit current and causes load-dependent voltage drop. In a line-frequency power transformer, a modest leakage value can be desirable because it limits fault current. In a wideband signal transformer, leakage inductance restricts high-frequency response and is usually reduced by close winding placement and interleaving.

Leakage voltage drop is a phasor quantity. A resistive load current aligns closely with secondary voltage, so the resistive and leakage contributions form different directions in the voltage diagram. A lagging load can produce a larger positive regulation drop. A leading load can partly offset the series leakage contribution and can yield a terminal voltage above its no-load value. The sign of voltage regulation is therefore load-power-factor dependent, even when the transformer itself has only positive resistance and inductance.

Winding placement changes both leakage and capacitance. Placing primary and secondary layers adjacent reduces mean magnetic path length between them. Interleaving layers reduces leakage further, but increases interwinding capacitance and can transmit more common-mode noise. A design specifies an acceptable compromise according to insulation requirements, required isolation, source impedance, load impedance, and frequency band.

Core magnetization and saturation

An applied sinusoidal winding voltage with negligible series drop gives the familiar volts-per-turn relation on integration of Faraday's law

V_{rms} = 4.44 f N Φ_{pk} .

Here $f$ is frequency, $N$ is the driven winding's turn count, and $Φ_{pk}$ is peak core flux. The factor 4.44 belongs to a sine wave. It changes for a square wave, trapezoid, or any other waveform because the relevant constraint is the time integral of voltage. Frequency and applied voltage must therefore be specified together. Reducing frequency while holding RMS voltage constant increases peak flux. Increasing applied voltage at fixed frequency has the same effect.

The core's $B$ - $H$ curve has a high-slope region followed by a knee and a lower-slope saturated region. Below the knee, a small magnetizing current produces the required flux density. Near saturation, a much larger current produces little additional flux change. Primary current then becomes sharply non-sinusoidal and can exceed the winding or source rating even with an open secondary. The secondary voltage may distort at the same time.

The core magnetization curve. Below the knee a small magnetizing field

H

produces the required flux density

B

; beyond it, much larger

H

yields little further

B

, so the magnetizing current turns sharply non-sinusoidal.

Core saturation is controlled by several related quantities:

Applied volt-seconds. The integral $\int v d t$ determines flux change. A DC offset in the voltage waveform displaces the core operating cycle toward one limit.
Frequency. Lower frequency gives each half cycle more time to accumulate flux at a fixed voltage amplitude.
Turns and core area. More turns or larger effective core area reduce flux density for a stated voltage and frequency.
Temperature and material. Permeability, loss, and saturation flux density vary with core material and temperature.

An inrush transient occurs when energizing a transformer at an unfavorable source phase and with residual core flux. The new flux transient can add to the retained flux and cross the saturation knee for several cycles. Inrush current is limited by source impedance, winding resistance, and leakage. It is not reliably predicted by dividing rated voltage by a single magnetizing reactance measured under small-signal conditions.

Core loss and frequency range

Magnetic materials dissipate energy during each reversal of magnetization. Hysteresis loss corresponds to the area inside the material's $B$ - $H$ loop. Eddy currents are circulating currents induced inside conductive core material; their heating depends on changing flux and available current paths. Laminating a power-frequency core into thin insulated sheets interrupts large eddy-current loops. Ferrite cores have high electrical resistivity and are common at higher switching frequencies.

At a fixed sinusoidal voltage, increasing frequency lowers the required flux amplitude and often reduces the low-frequency saturation risk. Core loss may still rise because the material is cycled more times per second. Transformer core selection therefore requires a material loss curve at the intended frequency and flux density, not a simple rule that higher frequency is always safer. The winding AC resistance and parasitic capacitance also become more important as frequency rises.

Distributed capacitance provides a high-frequency path between windings and from a winding to a grounded core or shield. It can couple common-mode voltage even when the windings are galvanically isolated. Electrostatic shields reduce that coupling when properly connected, but a continuous turn around the core would behave as a shorted winding. Shields are constructed with a break so they do not form a closed conductive loop.

Regulation, efficiency, and ratings

Transformer nameplate voltage is normally a full-load RMS rating at a stated frequency. The open-circuit secondary voltage is often higher because almost no secondary copper or leakage drop occurs. Connecting a load increases current and produces a terminal drop that depends on the series impedance and on the load phase. Voltage regulation is reported as

VR = \frac{V _{nl} - V _{fl}}{V _{fl}} \times 100%,

where $V_{nl}$ is the no-load secondary voltage and $V_{fl}$ is the secondary voltage at specified full load. A positive value describes a falling terminal voltage for the stated load. The calculation must report the load power factor because leakage reactance changes the result for leading and lagging loads.

Secondary terminal voltage against load current. The voltage is highest at no load (

V_{n l}

) and falls as winding and leakage drops grow with current; the fractional drop to full load (

V_{f l}

) is the voltage regulation.

The same transformer can have different regulation values for different loads. A resistive load gives a drop associated with both winding resistance and leakage reactance. A lagging current gives a leakage-reactance contribution in a direction that usually increases the drop. A leading current can make the phasor subtraction smaller; some transformers show negative regulation at a sufficiently leading power factor. The result means the terminal voltage rises under the stated load condition. It does not indicate power gain.

Use a defined testing order. Measure the primary voltage at the transformer terminals, not at a distant source outlet. Measure secondary voltage at the load terminals, not at the ends of a separate lead pair. Lead resistance and measurement-point differences can otherwise be recorded as transformer regulation. Hold frequency and primary voltage constant, allow the unit to reach an appropriate temperature state, and specify the load current and power factor with the result.

Efficiency compares real output power with real input power:

η = \frac{P _{out}}{P _{in}} = \frac{P _{out}}{P _{out} + P _{core} + P _{cu} + P _{stray}} .

The core-loss term is approximately constant over a modest range of load currents when primary voltage and frequency are held fixed. Copper loss grows approximately as the square of load current. Stray loss includes eddy-current loss caused by leakage fields in conductors, clamps, and nearby metallic structure. It often increases more rapidly than the simplest $I^{2} R$ estimate. At small loads, fixed core loss dominates. At high loads, copper and stray loss dominate.

Loss against load current. Core loss stays nearly constant at fixed voltage and frequency; copper loss grows as the square of current. Their sum sets the thermal limit, and the efficiency peak falls where the two losses are equal.

With core loss treated as constant and copper loss proportional to $I^{2}$ , maximum efficiency occurs near the load current for which copper loss equals core loss. That condition is a design result, not a recommendation to operate every transformer at the maximum-efficiency point. Voltage regulation, temperature rise, overload margin, required output voltage, and duty cycle may select a different operating current. Efficiency also changes with load power factor because a lower-power-factor load can draw the same RMS current while delivering less real output power.

The apparent-power rating is

S_{rated} = V_{rated} I_{rated} .

It establishes a voltage-current thermal envelope for the windings and insulation. It is not automatically a real-power rating in watts. A 100 VA transformer may supply 100 W to a unity-power-factor resistive load at rated voltage and current, but only 80 W to a 0.80 power-factor load at the same rated RMS quantities. The winding heating is largely tied to RMS current, so the VA rating remains relevant for both loads.

Nameplate information also has a time basis. A continuous rating assumes a thermal equilibrium condition. A short-duration overload may be permissible under a published duty cycle because the winding and core have finite heat capacity. Repeated overloads can accumulate temperature rise if the cooling interval is too short. A small sealed transformer, an open-frame transformer, and a fan-cooled unit can have the same nominal VA value under different ambient-temperature and installation conditions.

Short-circuit impedance is another rating-related property. With the secondary shorted under a controlled test, only a small fraction of rated primary voltage may be needed to circulate rated current. The ratio of that applied voltage to rated voltage, often stated as a percentage, summarizes the transformer series impedance. It limits fault current and contributes to regulation. A transformer with lower short-circuit impedance has lower load drop but can deliver substantially larger fault current from a stiff source.

Measurement and validation

Transformer measurements begin with terminals, references, and an energy limit. Record the nameplate voltage, frequency, VA rating, winding identifiers, insulation class, and any stated tap connections. Apply only the voltage and frequency intended for the test configuration. A practical lab uses an isolated low-voltage AC source, current limiting, fused leads, and appropriately rated probes. Mains-connected windings, high-energy power transformers, and unknown equipment require a supervised setup and a documented isolation plan.

The turns ratio can be measured with a low-amplitude sinusoidal test applied to one winding while every other winding remains open. Measure the RMS terminal voltage on the driven winding and on each open winding. The observed ratio gives

\frac{N _{2}}{N _{1}} \approx \frac{V _{2}}{V _{1}} .

The approximation includes small winding and magnetizing drops. Repeating the test in both drive directions checks that the selected terminal pairs have been identified correctly. A result far from the expected ratio can indicate a wrong tap, a series connection error, a meter bandwidth limitation, or an input waveform that differs from the stated amplitude convention.

The polarity test identifies winding starts and ends. Energize one winding at a safe test level. Connect one lead from it to one lead from a second winding, then measure the voltage between the two remaining leads. The reading is close to the sum of the two winding voltages for a series-aiding connection and close to the difference for a series-opposing connection. Label the dots or leads after the measurement. This result is more reliable than inferring polarity from physical lead position on an unknown component.

The open-circuit test estimates the shunt branch. The secondary is left open, the primary receives rated-frequency voltage at its rated terminal value, and the test records $V_{1}$ , no-load current $I_{0}$ , and input real power $P_{0}$ . Because the no-load current is usually small, primary copper loss is small. The measured real power then approximates core loss at that voltage and frequency. The measured current can be resolved into an in-phase loss component and a quadrature magnetizing component after the measurement references and meter bandwidth are checked.

The short-circuit test estimates the series branch. The secondary is shorted with an approved low-resistance path. A reduced primary voltage is increased only enough to produce the stated test current. Record the applied voltage, current, and real power. The applied voltage is small compared with rated voltage, so core flux and core loss are small; the measurement is dominated by winding resistance and leakage impedance. This test is part of a controlled lab or manufacturer procedure. A direct secondary short on an energized mains transformer without current limiting can create destructive current and is not a diagnostic method.

Load testing checks the values that matter to the intended application. Record primary terminal voltage, secondary terminal voltage, primary current, secondary current, input real power, output real power, waveform frequency, and load type. Increase the load in planned steps up to the permitted rating. Compare the measured full-load voltage with the regulation prediction and compare total measured loss with the temperature rise. A discrepancy can indicate an unmodeled lead drop, an incorrect turns ratio, meter phase error, saturation, a load whose waveform is not sinusoidal, or a transformer outside the model's frequency range.

Measurements on high-impedance transformer circuits require attention to instrument loading. A meter with finite input capacitance or resistance changes the secondary load, especially in small-signal and high-frequency work. A current probe has its own phase and amplitude error. A two-channel oscilloscope with a shared earth-referenced ground can short two circuit nodes through its probe grounds if the test arrangement is not isolated. Use differential probes or a documented isolated measurement arrangement where the circuit requires it, and state probe scale factors in the measurement record.

Worked calculations and design checks

Worked example (step-down turns, currents, and reflected load). A 60 VA transformer has a 120 V RMS primary of $N_{1} = 1200$ turns and a 12 V RMS secondary. Find the secondary turns, the rated primary and secondary currents, and the primary resistance seen when the secondary drives a $2.4 Ω$ resistor.

The turns follow the voltage ratio:

N_{2} = N_{1} \frac{V _{2}}{V _{1}} = 1200 \cdot \frac{12}{120} = 120 turns .

The rated currents follow from the VA rating at each winding's voltage:

I_{2, rated} = \frac{60 VA}{12 V} = 5.0 A, I_{1, rated} = \frac{60 VA}{120 V} = 0.50 A .

A $2.4 Ω$ resistor draws $12/2.4 = 5.0 A$ at 12 V, so it is the rated load. With $a = N_{2} / N_{1} = 0.10$ , its reflected resistance is

R_{in} = \frac{R _{L}}{a ^{2}} = \frac{2.4 Ω}{( 0.10 ) ^{2}} = 240 Ω,

and the source-side current $120/240 = 0.50 A$ matches the VA calculation. A real unit draws slightly more primary current because core and winding losses add to the input power.

Worked example (peak flux density and a frequency change). A 120 V RMS, 60 Hz winding has 900 turns and an effective core area of $12 c m^{2} = 1.20 \times 1 0^{- 3} m^{2}$ . Neglecting the series drop, find the peak core flux and flux density, then the flux density if the same winding runs at 50 Hz and 120 V.

The volts-per-turn relation $V_{rms} = 4.44 f N Φ_{pk}$ gives

Φ_{pk} = \frac{120}{( 4.44 ) ( 60 ) ( 900 )} = 5.00 \times 1 0^{- 4} Wb, B_{pk} = \frac{Φ _{pk}}{A} = \frac{5.00 \times 1 0 ^{- 4}}{1.20 \times 1 0 ^{- 3}} = 0.417 T .

At fixed voltage, $Φ_{pk} \propto 1/ f$ , so lowering the frequency to 50 Hz raises the flux density by the frequency ratio:

B_{pk, 50} = 0.417 \cdot \frac{60}{50} = 0.500 T .

Compare this against the core material's allowable flux density and loss data at the operating temperature and waveform: reducing frequency at the same voltage has driven the flux density up by 20 percent.

Model selection and calculation checks

Transformer calculations require a model whose assumptions match the quantity being estimated. A turns ratio is sufficient for an unloaded secondary-voltage estimate at the intended frequency. It is insufficient for a full-load voltage estimate, a thermal rating calculation, an inrush prediction, or a high-frequency isolation measurement. Writing the active assumptions beside the calculation prevents a correct formula from being applied to the wrong operating condition.

Ideal ratio model. Use $V_{2} / V_{1} = N_{2} / N_{1}$ and the reciprocal current ratio for rated-voltage conversions, first-pass reflected-load calculations, and conceptual power flow. State that winding drops, magnetizing current, core loss, and leakage are being neglected.
Referred series model. Add reflected winding resistance and leakage reactance when computing load voltage, short-circuit current, or regulation. The load phase must accompany this model because the series drop is a phasor difference.
Shunt-core model. Include magnetizing reactance and core-loss resistance for no-load current, no-load real-power, and source-loading estimates. The values are valid near the voltage, frequency, and flux density at which they were measured.
Frequency-aware model. Add distributed capacitance, AC winding resistance, and material loss data when operating outside the low-frequency range. A 60 Hz ratio measurement cannot establish a transformer’s bandwidth or common-mode behavior.

Units provide a quick consistency check. Turns ratios and regulation are dimensionless. Impedance reflection multiplies ohms by a dimensionless squared ratio, leaving ohms. The product of RMS volts and RMS amperes has units of VA. Real power requires the relevant power factor or a direct time-domain wattmeter measurement. Flux is measured in webers, flux density in teslas, and the core cross-sectional area in square metres. A missing area conversion from square centimetres to square metres introduces a factor of $1 0^{4}$ in a flux-density calculation.

Use the following record for a calculation or test report:

terminal labels, dot locations, voltage polarity, and current-arrow directions;
RMS, peak, or instantaneous amplitude convention for every stated voltage and current;
operating frequency, source impedance, load impedance, and load power factor;
temperature state and allowed VA or current rating;
model parameters that were measured rather than assumed;
the location of voltage probes and current sensors.

Several recurring errors can be found from these entries. A voltage ratio greater than one paired with a secondary current ratio greater than one signals a failed reciprocal check. A computed reflected resistance that moves in the wrong direction often means that $N_{1} / N_{2}$ and $N_{2} / N_{1}$ were interchanged. A predicted secondary voltage above its no-load value under a resistive load calls for a review of voltage references and series drop signs. A current calculation above the VA rating requires a thermal and protective review even if the turns ratio remains correct.

The equivalent circuit is also an experimental claim. Compare its predicted no-load current with an open-circuit record, its predicted short-circuit voltage with the series test, and its predicted loaded secondary voltage with a load test. Agreement at one point does not validate the model at every voltage, frequency, or temperature. Saturation, nonlinear core loss, lead resistance, and waveform distortion can alter the parameters outside the test condition. Report the range where agreement was checked.