The test itself is simple: inject a voltage on one winding, measure the induced voltage on the other, and compare the measured ratio with the expected ratio for the tap position. The engineering content is in the expected value, the acceptable deviation at that position, and what a deviation implies.

Why ratio is measured tap by tap

A power transformer tap changer changes the number of active turns in one winding. Each tap therefore has its own expected ratio, and each tap exercises a different set of selector contacts and a different length of tap winding. Testing one position verifies one contact set; testing the range verifies the mechanism, the selector and the winding together.

The practical case for the full run is that the failure modes are position-specific. A diverter contact that has eroded after years of arcing operations shows up on the taps that use it. A selector that stops fractionally short of its seated position affects the taps on one side of the range. A tap winding with a damaged turn affects the taps that include it. None of these produce an obvious deviation at the mid position.

Running the range also establishes the pattern that makes a single deviation meaningful later. A technician who has seen the same transformer produce a smooth progression across sixteen taps knows immediately when one step sits outside that progression, and that recognition is faster and more reliable than any single limit check.

Acceptance limits by tap position

The limit applied to a ratio measurement is a tolerance on deviation from the declared ratio for that tap. Most transmission specifications apply a band of 0.5 percent at the principal tap, and many accept up to about 1 percent at the extremities where the tap winding contributes a larger share of the total turns and manufacturing tolerances accumulate.

The comparison should be made against the factory or commissioning value for the same tap wherever one exists, because that value already includes the design and manufacturing reality of the unit. Comparing only against a printed nameplate ratio is a weaker test, since nameplate values are nominal and are commonly rounded.

The phase-to-phase row deserves emphasis. Because the three phases are measured at the same tap within the same session, the spread between them should be smaller than the absolute deviation band, and a phase that differs from its neighbours warrants investigation even when all three remain inside the absolute limit.

Separating ratio error from tap changer position error

When a ratio deviates, the first question is whether the winding has changed or the tap changer is not where the indicator says it is. The two produce similar numbers and completely different responses, so the diagnostic step is to identify the pattern across the tap range rather than to examine the single deviating reading.

A tap position error produces a ratio that matches a neighbouring tap. If the tenth tap reads like the ninth, the selector is probably sitting one position off, and the ratio test has just detected an indication fault rather than a winding fault. This is one of the more useful functions of the test, because the discrepancy is otherwise invisible without opening the compartment.

A winding or connection error produces a deviation that persists across the whole range and shifts with the tap change in a way the expected values do not. A partial tap winding connection fault produces a deviation concentrated in the taps that include that winding, and a phase-to-phase difference that is present at every position usually means an internal lead or connection issue rather than a tap changer problem.

Where the pattern is ambiguous, the position check is worth doing before any conclusion. Verifying the indicated position against the mechanism, either through the position transmitter or by operating the changer through a full cycle and re-measuring, separates the two cases at low cost.

Vector group and polarity as companions

Ratio and vector group are usually tested in the same pass because a three-phase ratio test establishes the voltage relationships between the windings, and the phase displacement is part of that relationship. Recording only the magnitude ratios and discarding the phase information throws away half the result.

The vector group defines the phase displacement between the primary and secondary windings, expressed as a clock notation, and it is fundamental to paralleling. Two transformers with matching ratios and mismatched vector groups cannot be operated in parallel, and the mismatch is rarely obvious from the ratio magnitudes alone, because a delta-star connection can produce ratios that look plausible while the phase relationship is wrong.

Polarity checks matter most on single-phase units and on instrument transformers, and they matter most after any work that involved reconnecting leads, bushings or winding terminations. A reversed polarity produces a ratio that is numerically plausible but physically wrong, and it will produce circulating currents the moment the unit is paralleled with another.

What a single-tap deviation implies

A deviation isolated to one tap narrows the cause list sharply. The winding is common to all taps, and the measurement chain is unchanged between taps, so a deviation that exists at exactly one position points to something that is unique to that position: a selector contact, a diverter contact, or a tap lead.

The isolated deviation is also the case most likely to be a measurement artefact, so it warrants a repeat before any escalation. An unstable supply, a lead that was disturbed, a contact that was not fully made during the test, or a tap changer that had not completed its operation when the measurement began all produce the same signature.

If a repeat confirms the deviation, the next step is normally to exercise the tap changer over its full range and re-measure, which reseats the contacts. A deviation that disappears after a full cycle is usually a contact that was not seated. A deviation that persists is a physical defect in the tap changer or its connections.

Recording and comparing across the tap range

The record that supports comparison is a table rather than a single verdict. One row per tap, one column per phase, with the measured ratio, the expected ratio, the deviation in percent, and the phase displacement. Read down the deviation column and the pattern is visible immediately; read across the rows and the phase comparison appears.

The record should also carry the test voltage, the winding connections used, the tap direction convention, the ambient temperature where the instrument specifies a temperature effect, and the instrument model and serial number. Tap direction conventions are worth stating explicitly, since a transformer that raises voltage in the opposite direction from the convention used by the tester produces a table that is internally consistent but inverted.

Where a plant runs a repeat pattern of the same transformer type, storing the tap tables from every unit gives a reference set that makes an outlier obvious. That reference set is more useful than any single acceptance table because it captures the behaviour of the specific design, including its legitimate deviations at the extremities. The declared characteristics these measurements are checked against, including the tolerances on ratio, are defined in IEC 60076-1.

Instrument versus measurement error

Instrument error is the smallest term in a ratio measurement. A modern ratio tester with an accuracy specification well inside 0.1 percent contributes a fraction of a 0.5 percent acceptance band. The larger terms are the test connections, the excitation level and the state of the tap changer during the test.

Excitation level matters because ratio measurement assumes the winding is operating in its linear region. A test voltage that is too low relative to the core characteristics can produce a measurement that is influenced by core behaviour rather than by turns. Most instruments apply a voltage chosen to avoid this, and the instrument documentation states the expected range.

Connection quality is the most common practical source of error. A clip that is not fully seated, a bushing that is not clean, or a lead that shares a path with a grounding conductor all add impedance that appears as a ratio deviation. Checking the connections and repeating the measurement costs minutes and resolves the majority of marginal results.

Instruments that measure ratio and phase displacement together, such as those in the transformer ratio test equipment range, reduce the number of separate connections and therefore reduce the opportunity for this class of error.

Method notes from ratio tester suppliers such as OMICRON state the excitation levels a given instrument applies and the accuracy it claims, which is the information needed to decide whether an observed deviation is meaningful. On the tap changer side, the mechanism documentation published by Maschinenfabrik Reinhausen is the reference for selector and diverter behaviour, and the condition assessment material coordinated through CIGRE study committees puts a single tap table into the wider diagnostic context. Applied field experience on the same question is also collected by Transformer Consultants.

When ratio testing justifies an outage

Ratio testing justifies an outage in three situations: after transport or installation, before paralleling two transformers, and after any event that could have displaced a winding or the tap changer. In those cases the test either confirms a condition that must be verified before energisation or resolves a specific doubt.

For a unit already in service with a stable history, the ratio test usually travels with the rest of the outage programme rather than justifying an outage on its own, because a single ratio measurement is not a trend tool. The value of the test in a running fleet comes from the full tap table, which needs the unit isolated long enough to operate the changer through its range.

On a tap changer with a long service history, the ratio table is often the fastest way to decide whether a maintenance intervention is needed. A smooth progression across the range with consistent phase agreement supports deferring the work; a step change at one position or a widening phase spread supports scheduling it before the next operating season.