Transformer Winding Resistance Test: Kelvin (4-Wire) Bridge Method, Test Current Selection, OLTC Tap-by-Tap Testing & Three-Phase Unbalance Criterion (≤2%)

Abstract

The winding resistance test is arguably the single most informative routine field test on a power transformer. A single resistance measurement — compared to factory data, the adjacent phases, or prior years' results — can reveal loose bolted connections, broken conductor strands, OLTC contact degradation, high-resistance bushing draw-lead connections, and winding material quality issues. The test appears deceptively simple — inject DC current, measure voltage drop, divide by current — but the practical challenges of inductive winding saturation (the L/R time constant), test current selection, accurate temperature measurement, and correct interpretation of milliohm-level differences demand meticulous technique. This article covers the 4-wire (Kelvin) measurement principle that eliminates lead resistance error, the physics of current selection (why too much >10% rated current is prohibited and too little produces unstable readings), the OLTC tap-by-tap measurement protocol to verify every diverter switch and selector switch contact, and the established 2% phase-to-phase unbalance criterion for three-phase transformers.

1. The Kelvin (4-Wire) Measurement Principle

1.1 Why 2-Wire Resistance Measurement Fails

A transformer winding has a DC resistance on the order of:

Distribution transformers (<500 kVA, LV winding): 0.001-0.1 Ω
Power transformers (10-100 MVA, HV winding): 0.1-5 Ω
Power transformers (10-100 MVA, LV winding): 0.001-0.05 Ω

A standard 2-wire multimeter test lead has a resistance of 0.05-0.5 Ω per lead (depending on length, gauge, and connector quality). For an LV winding resistance of 0.005 Ω, the lead resistance is 10-100× the winding resistance — the measurement will read predominantly the lead resistance, not the winding.

The 4-wire (Kelvin) connection separates the current injection path from the voltage measurement path:

Current leads (C1, C2): Heavy-gauge leads (typically 4-10 mm² cross-section) connect the DC current source to the winding terminals. The current through these leads can be 5-50 A for power transformer testing. The voltage drop across these leads is irrelevant — it is not measured.

Potential leads (P1, P2): Light-gauge, shielded leads (typically 1-2.5 mm²) connect the voltmeter to the winding terminals as close as possible to the winding connection point. The voltmeter has an input impedance >10 MΩ → the current through the potential leads is negligible (<1 μA) → the voltage drop across the potential leads is effectively zero.

The measured voltage is the voltage drop across ONLY the winding, from P1 to P2. The lead resistances are eliminated from the measurement.

Connection rule: The potential leads must be connected as close as physically possible to the winding terminals, separate from the current lead connection points. If both are connected at the same clip, the contact resistance at the clip-to-terminal interface appears in both the current and potential paths — the Kelvin method eliminates lead resistance but not contact resistance.

1.2 Instrument Types

Modern transformer winding resistance test sets are microprocessor-controlled, high-current DC sources with integrated 4-wire measurement:

Instrument	Max Test Current	Winding Resistance Range	Application
Megger MTO210	10 A	1 μΩ - 2 kΩ	Distribution to medium power transformers
Megger MTO330	50 A	0.1 μΩ - 2.5 kΩ	All power transformers
Omicron CPC 100 (+ CP TD1)	6 A	1 mΩ - 2 kΩ	Multi-function test set, lower current but more versatile
Raytech WR50-12	50 A	1 μΩ - 500 Ω	Large power transformers, dual measurement (simultaneous HV + LV)

2. Test Current Selection

2.1 Minimum Current — Stability Criterion

The test current must be high enough to produce a stable, measurable voltage drop above the noise floor. A minimum test current of 1-5% of rated winding current is typically sufficient:

I_test_min ≥ 0.01 × I_rated

For a 50 MVA, 132 kV transformer: I_rated_HV = 50 × 10⁶ / (√3 × 132 × 10³) = 218.7 A → I_test_min ≈ 2.2 A. A test current of 5-10 A DC is easily achievable and provides a stable measurement.

For an LV winding rated at 33 kV, 875 A → I_test_min ≈ 8.75 A. A 10 A test set is adequate. For LV windings on large transformers (several thousand amperes), a 50 A test set may be required.

2.2 Maximum Current — Core Saturation Criterion

The test current must not exceed the transformer's rated continuous winding current by more than 10%:

I_test_max ≤ 0.10 × I_rated

Why: The DC test current produces a DC flux in the core. Under normal AC operation, the core flux oscillates around zero (symmetrical B-H loop). The DC test current biases the core flux in one direction:

If I_test > 10% × I_rated: The core may saturate from the DC bias. A saturated core has dramatically reduced incremental permeability, causing the winding inductance to drop by 90-99%. This makes the measurement more difficult (the L/R time constant becomes very short, and the measurement must be taken quickly), and the core may retain remanence after the test, causing high inrush on the next energization.
If I_test ≤ 10% × I_rated: The DC bias is modest, and the core remains in the linear region. This is the safe operating range.

For tap-changer testing: The test current must also be within the through-current rating of the OLTC. At extreme tap positions, the winding section in circuit may have reduced current-carrying capacity.

2.3 The L/R Time Constant Problem

The transformer winding is a large inductor (L = 10-500 H for power transformer HV windings) with a small series resistance R (0.1-5 Ω). The electrical time constant:

τ = L / R

For L = 200 H, R = 1 Ω → τ = 200 seconds. This means the current will reach only 63.2% of its steady-state value after 200 seconds, and 99% after approximately 5τ = 1,000 seconds (16.7 minutes).

Practical implications:

Waiting for the current to fully stabilize for every measurement point (especially for OLTC tap-by-tap testing with 17+ positions) would take hours
Modern test sets address this by: (1) injecting a higher initial voltage to force the current to stabilize faster (voltage-forcing or current-ramping mode), (2) using simultaneous magnetization of HV and LV windings (dual-injection method — the ampere-turns of the HV and LV winding are arranged to cancel, minimizing the net DC flux in the core and reducing the effective inductance by 90-95% — this is the most effective method for speeding up measurements), or (3) measuring the winding resistance at a predetermined time before full stabilization (e.g., at 90% of the steady-state current) and applying a correction factor — less accurate and not recommended for precision measurements

3. OLTC Tap-by-Tap Measurement

3.1 Why Test Every Tap?

The OLTC (on-load tap changer) has 9-35 tap positions depending on the voltage regulation range. Each tap position corresponds to a specific section of the regulating winding in the circuit. The winding resistance should increase smoothly and monotonically as the tap position adds more winding turns:

Broken or high-resistance diverter switch contacts: A single tap position showing 5-50% higher resistance than the adjacent taps indicates a degraded contact at that specific position — the most common OLTC failure precursor
Selector switch misalignment: An anomalous resistance jump at one tap indicates the selector switch contact is not fully engaging at that position (mechanical misalignment, worn contact fingers)
Regulating winding fault: A resistance anomaly repeating at multiple tap positions indicates a problem in the regulating winding itself (shorted turns, broken strand)

3.2 Measurement Protocol

Start at the lowest tap position (Tap 1, typically the maximum turns ratio)
Measure R at Tap 1. Record: Tap position, measured resistance, test current, winding temperature, time of measurement.
Operate the OLTC to the next tap position. Allow the tap change to complete (observe the "tap change complete" indication).
Wait for the test current to stabilize at the new tap position (the L/R time constant changes because the winding inductance changes — at maximum tap, the full winding is in circuit, giving the longest time constant; at minimum tap, a reduced winding section has lower inductance and shorter time constant).
Measure R and record.
Repeat for all tap positions.

Safety interlock: The OLTC motor drive should be electrically isolated during the resistance test — the induced voltage from mutual inductance between the test winding and other windings, or induced from nearby energized equipment, can damage the OLTC motor controller electronics.

4. Three-Phase Unbalance Criterion (≤2%)

4.1 Calculating Phase Unbalance

For a three-phase transformer (delta or wye connected), the winding resistance is measured for each phase at the same tap position and temperature. The phase unbalance is:

Unbalance (%) = (R_max - R_min) / R_avg × 100

Where:

R_max = highest phase resistance
R_min = lowest phase resistance
R_avg = (R_A + R_B + R_C) / 3

4.2 Acceptance Criterion

Per IEC 60076-1 (Clause 10.2) and IEEE C57.12.90:

Phase unbalance ≤2% for windings in good condition with no broken strands or poor connections.

Interpretation:

Unbalance	Interpretation	Action
≤1%	Normal — within manufacturing tolerance and measurement uncertainty	Accept
1-2%	Slight unbalance — may be within measurement uncertainty if the winding temperature is not perfectly uniform	Repeat measurement with temperature stabilization; investigate if persistent
2-5%	Significant unbalance — likely a defective connection (loose bolted joint, high-resistance tap changer contact, partial broken strand in a multi-strand conductor)	Investigate — internal inspection of connections, tap changer maintenance
>5%	Severe unbalance — one phase has a defective winding or connection that will overheat under load	Do not energize — rectify the defect before returning to service

4.3 Temperature Uniformity

The most common cause of a false phase unbalance alarm is non-uniform winding temperature. If the transformer has been exposed to uneven solar heating (one side exposed to direct sunlight, the other shaded), or the cooling has been uneven, the phases may be at different temperatures. A temperature difference of 1 °C produces a resistance difference of approximately 0.4% for copper (temperature coefficient α ≈ 0.00393/°C at 20 °C).

Before concluding a phase unbalance defect: Measure each phase's temperature (winding resistance thermometer or the winding temperature indicator (WTI) for each phase) and verify that all phases are within 2-3 °C of each other. If not, allow the transformer to thermally stabilize (4-8 hours with cooling fans off, shielded from direct sun) before re-testing.

FAQ

Q: Why does the winding resistance measurement take so long to stabilize?

The transformer winding is a large inductor with a small resistance — the L/R time constant is 100-500+ seconds for power transformer HV windings. The DC test current takes approximately 5τ (500-2,500 seconds = 8-40 minutes) to reach 99% of its steady-state value. Modern test sets speed this up by: (1) applying a higher initial voltage to force the current ramp (constant-power mode), (2) using simultaneous dual-injection (HV and LV windings energized with opposing ampere-turns to cancel the core flux, reducing the effective inductance by 90-95% — cuts the stabilization time from 20 minutes to 30-60 seconds), or (3) measuring before full stabilization with a correction algorithm. For routine testing without dual-injection capability, allow 5-10 minutes per measurement point. The wait is tedious but necessary — a measurement taken before 90% stabilization will read 5-10% higher than the true DC resistance and will be irreproducible.

Q: What is the relationship between winding resistance and temperature?

The resistance of copper follows: R(T₂) = R(T₁) × (234.5 + T₂) / (234.5 + T₁). For aluminum: R(T₂) = R(T₁) × (228 + T₂) / (228 + T₁). For comparison with factory data (typically measured at 20-25 °C), correct the field measurement using the formula. The temperature used should be the average winding temperature, measured by: (1) the winding resistance itself (if the cold resistance at a known temperature is available — this is the most accurate), (2) the WTI (winding temperature indicator) reading, or (3) the top-oil temperature + an estimated winding-to-oil gradient (typically 5-15 °C depending on the load and cooling mode). Always record the temperature measurement method and the assumed gradient.

Q: What current should I use for a winding resistance test?

The test current should be 1-10% of the rated winding current: sufficiently high (≥1%) to produce a stable voltage measurement above the noise floor, but not so high (>10%) that the core saturates from DC bias. For a winding rated at 500 A, a test current of 5-50 A DC is acceptable. In practice, most field test sets are limited to 5-10 A (portable units) or 50 A (large trailer-mounted units). Use the highest current the test set can deliver, up to the 10% limit — higher current produces a larger voltage signal, improving measurement stability.

Q: Why test winding resistance on every OLTC tap?

The OLTC diverter switch contacts experience arcing during every tap change. Over tens of thousands of operations, the arcing contacts erode and develop increased contact resistance. The resistance increase typically starts at one or two specific tap positions (the positions where the transition impedance passes the most current, or where the diverter switch contact mechanism has a slight mechanical misalignment). A single high-resistance tap position — while all others are normal — is the earliest detectable sign of OLTC contact degradation. Testing only the nominal tap misses this early warning signal. Tap-by-tap testing adds 30-60 minutes to the test time but provides a complete map of OLTC contact condition.

Q: What reference temperature should I correct my winding resistance measurements to?

The standard reference temperature for transformer winding resistance is 75 °C for oil-immersed transformers (IEC 60076-1) or 85 °C for dry-type transformers. However, for trending with historical factory data, correct to the temperature at which the factory test was performed (typically 20-25 °C — check the factory test report). The factory test report should state both the measured resistance and the temperature. If comparing with field data from a prior year, correct both datasets to the same reference temperature (20 °C is the most common for asset management databases).

Q: If I reverse the test current polarity, should the resistance be identical?

Yes — within the measurement uncertainty of the test set (±0.2% for modern instruments). The winding resistance is an ohmic property and is independent of current polarity. If a significant difference (>0.5%) is observed between positive and negative polarity measurements, the cause is typically: (1) thermal EMF (thermoelectric voltage) at the test lead connections — the junction of dissimilar metals (copper lead to brass terminal) generates a small DC voltage (microvolts to millivolts) when a temperature gradient exists across the junction; this adds to the measured voltage in one polarity and subtracts in the other, (2) residual core magnetization from prior DC testing (the core remanence interacts with the test current), or (3) a rectifying contact (a high-resistance connection that behaves as a semiconductor junction — this indicates a serious connection defect). The standard practice is to measure in both polarities and average the results, eliminating the thermal EMF error.

References / Standards

Reference	Title
IEC 60076-1:2011	Power transformers — Part 1: General
IEEE C57.12.90-2021	IEEE Standard Test Code for Liquid-Immersed Distribution, Power, and Regulating Transformers
IEEE C57.152-2013	IEEE Guide for Diagnostic Field Testing of Fluid-Filled Power Transformers, Regulators, and Reactors
IEC 60076-3:2018	Power transformers — Part 3: Insulation levels, dielectric tests and external clearances in air

*Authored by Du Fu, Production Engineer at ZY POWER. The winding resistance test is the definitive electrical connectivity check for every current-carrying path in the transformer. Never energize a transformer without a complete, temperature-corrected winding resistance measurement at all tap positions.*