The turns ratio test looks like the simplest measurement in transformer work. Apply a small voltage to one winding, measure what comes out the other, divide, compare to nameplate. A handheld instrument does it in seconds and prints a pass/fail.
That simplicity is exactly why it gets done badly. People read the ratio number, see it’s within tolerance, and move on — ignoring the two other numbers the instrument gives them, skipping taps, missing the difference between a global error and a single-phase error, and never demagnetizing the core before they start. The test that should catch a wrong winding, a shorted turn, a bad tap contact, or a reversed connection ends up confirming only that the ratio is roughly right.
This is a practical guide to running and interpreting a TTR test properly — on the bench at commissioning, in the field on a maintenance round, or in the factory during acceptance. The physics is straightforward; the value is in reading all the information the test actually provides.
Table of Contents
What the Test Measures, and What It Doesn’t
A transformer transfers voltage in proportion to its turns ratio. If the high-voltage winding has 26.48 times as many turns as the low-voltage winding, then voltage applied to the HV side appears on the LV side reduced by that factor. The TTR test exploits this directly: inject a low AC voltage (typically under 250 V, often much less) on one winding, measure the induced voltage on the other, and the ratio of those two voltages is the turns ratio.
One precise point that matters for three-phase work: what you actually measure is the voltage ratio, not the turns ratio. For single-phase transformers and for straightforward Yy or Dd three-phase units, the two are the same. But for Yd, Dy, zig-zag, and other configurations where the winding connection introduces a phase shift, the measured voltage ratio differs from the turns ratio by a factor that depends on the vector group. The instrument (or the test engineer) applies a correction factor based on the vector configuration to get from measured voltage ratio to the value that should be compared against nameplate. Get the vector group wrong in the setup and the “ratio error” you see is an artifact of the correction, not a real fault.
What the test doesn’t measure: insulation strength, dielectric condition, moisture, oil quality, mechanical integrity. A transformer can pass TTR with flying colors and still fail a dielectric test or have wet insulation. TTR proves the windings have the right number of turns and are connected correctly. That’s a necessary condition for a healthy transformer, not a sufficient one.
The Three Numbers, Not One
A modern TTR instrument presents three quantities for every measurement. Reading only the first is the most common way the test gets underused.
Turns ratio (or voltage ratio). The headline number. Compared against the calculated value from nameplate. The acceptance tolerance is ±0.5% deviation — this is consistent across IEC 60076-1, IEEE C57.12.00, and IEEE C57.152. A measured ratio more than half a percent off the calculated value is a fail.
Excitation current. The current the instrument has to push into the energized winding to magnetize the core during the test. This is the diagnostic gold that gets ignored. Excitation current is sensitive to core condition and to winding faults — a shorted turn, even one, dramatically increases the excitation current because the shorted turn acts as a loaded secondary, drawing current the instrument has to supply. A ratio that’s within tolerance but an excitation current that’s abnormally high or unbalanced between phases is telling you something is wrong even though the ratio “passed.”
Phase displacement (phase angle). The angular difference between the input and output voltages. This verifies the vector group and the winding polarity. A transformer wound or connected with a polarity error can produce a ratio that’s numerically correct but a phase angle that’s wrong — and that’s a serious fault that the ratio number alone won’t catch.
A thorough test engineer reads all three on every measurement. The ratio tells you the turns count is right. The excitation current tells you the winding and core are healthy. The phase angle tells you the connection is correct. Skipping the last two throws away most of the test’s diagnostic power.
Before You Test: Demagnetize
A point that’s routinely skipped and routinely causes confusion: residual magnetism in the core distorts the excitation current measurement.
If the transformer was recently energized, or had DC injected into it (a winding resistance test, for instance, leaves the core magnetized), there’s residual flux in the core. That residual flux changes how the core responds to the TTR excitation, throwing off the excitation current reading and making it non-comparable to previous tests or to other phases.
The fix is to demagnetize the core before the excitation current measurement. Many modern instruments have a demag function. The sequence matters in a test plan: winding resistance (which uses DC and magnetizes the core) should come after TTR, or the core should be demagnetized between them. A test engineer who runs winding resistance first and then wonders why the TTR excitation currents look strange has answered their own question.
The ratio measurement itself is less sensitive to residual magnetism than the excitation current, but for clean, comparable, trendable data, demagnetize first.
Testing a Three-Phase Transformer: The Pattern Tells the Story
On a three-phase transformer, you measure all three phases — A-a, B-b, C-c (or the appropriate pairs for the vector group). The single most useful interpretive skill is reading the pattern across the three phases, not just each value in isolation.
The normal excitation pattern. On a three-limb core, the magnetic path isn’t symmetric — the center phase has a shorter, lower-reluctance path than the two outer phases. So the excitation currents come in a characteristic pattern, often described as high-low-high: the two outer wound phases draw similar, higher excitation currents, and the center phase draws the lowest. This is normal and expected. A test engineer who sees an asymmetric excitation pattern and panics may just be looking at normal three-limb core behavior. The flag is when the pattern is wrong — when it isn’t high-low-high, or when one outer phase differs sharply from the other.
Uniform error on all three phases. If all three ratios are off by a similar percentage, the problem is usually global, not internal to the transformer. A wrong test voltage setting. A systematic error in the nameplate data entered. A wrong HV-LV pairing in the connection. Before condemning the transformer, check the setup — a uniform error usually means the test is wrong, not the transformer.
Error on a single phase. This is the meaningful fault signature. If B-b is off but A-a and C-c are fine, the fault is isolated to the B phase leg. Shorted turns on the B winding. A bad tap-changer contact on the B phase. A connection problem specific to that phase. A single-phase ratio error, with a corresponding excitation current anomaly on the same phase, is a strong, specific finding — it points at one leg of the transformer.
Correct ratios but failed interphase / vector check. If the individual phase ratios are right but the vector group comes out wrong, the internal connections are wrong. The classic example: a transformer nameplated Dy11 that tests as Dy1. The ratios are fine because the turns counts are right, but a winding set has been connected with reversed polarity, shifting the phase relationship. This is a build error, and it’s exactly the kind of thing TTR exists to catch before energization — a Dy1 connected where Dy11 was specified will not parallel correctly and may cause serious circulating currents.
Testing Every Tap
The tap changer is one of the most fault-prone parts of a transformer, and TTR is the test that exercises it. Testing only the principal tap is a half-test.
Run the ratio on every tap position, every phase. On each tap, two things should be true:
- The measured ratio at that tap matches the calculated ratio for that tap within ±0.5%.
- All three phases show the same ratio on that tap.
The progression across taps should be smooth and monotonic — each tap step changes the ratio by the designed increment. A tap that’s off, or a tap where the ratio jumps unexpectedly, points at a tap-changer contact problem: a worn contact, a misaligned selector, a contact not making properly. A tap that reads open (no ratio) means the contact isn’t connecting at all.
A practical note for transformers with non-linear tap arrangements: on some designs the per-tap ratio change isn’t uniform, and a strict ±0.5% per-tap check can flag taps that are actually fine. In those cases the criteria become: the ratio at both extreme taps (highest and lowest) is within tolerance of calculated, and for any given tap, all three phases agree. The phase-to-phase agreement on each tap is often the more reliable indicator than the absolute per-tap value.
New Transformer vs In-Service Transformer: Different Jobs
The same test serves two different purposes depending on the transformer’s life stage, and the interpretation differs.
New transformer (factory acceptance, commissioning). The job is conformity verification. The measured values are compared directly against nameplate and design data. Any deviation beyond ±0.5%, any wrong phase angle, any wrong vector group, renders the unit non-compliant — it shouldn’t be energized until resolved. There’s no history to compare against; the nameplate and design are the reference. The TTR result becomes the baseline record for the rest of the transformer’s life.
In-service transformer (maintenance, diagnostics). The job is dual: still check against the standard tolerance, but now also compare against the transformer’s own history. The rate of change between successive tests is often the more sensitive diagnostic than the absolute value. A transformer that read dead-on at commissioning and now reads at the edge of tolerance has changed — and the change, even within tolerance, is the signal. A common practical guideline is that a variation of more than about 2% from previous results (at the same temperature) warrants investigation, even if the absolute value still passes.
Sudden changes in ratio or phase angle versus historical records point to winding movement or mechanical displacement — the kind of thing caused by through-fault short-circuit forces, rough transport, or seismic stress. Trending excitation current upward over successive tests points to a developing internal abnormality. This trend analysis is only possible if the baseline and intermediate records exist and were taken under comparable conditions — which is why the commissioning TTR record matters years later.
Common Faults and Their Signatures
Pulling the interpretation together, here’s what specific faults look like on a properly read TTR test:
Shorted turns. Ratio may shift slightly (fewer effective turns), but the dominant signature is sharply elevated excitation current on the affected phase. The shorted turn is a closed loop acting as a loaded secondary. This is one of the most important faults TTR catches, and it’s invisible if you only read the ratio.
Open winding or open tap contact. No ratio reading, or a wildly wrong one, on the affected phase or tap. The instrument can’t establish the measurement because the circuit isn’t complete.
Tap-changer contact problems. Ratio errors that appear on specific taps, often with erratic or non-repeating readings as the contact makes and breaks. Re-running the tap may give different results — itself a sign of a marginal contact.
Wrong winding turns (manufacturing error). Ratio off by a consistent amount on the affected winding, repeatable, with normal excitation current. The turns count is simply wrong. Caught at factory acceptance, this is a build defect; the transformer goes back.
Reversed polarity / wrong connection. Ratios correct, phase angle wrong, vector group result doesn’t match nameplate. A connection or winding-direction error. Serious for paralleling and protection.
Core problems (shorted laminations). Ratios usually normal, but excitation current elevated and unbalanced in a way that doesn’t fit the normal high-low-high pattern. TTR isn’t the primary core test, but the excitation current reading flags core issues as a side benefit.
Practical Setup Points That Affect Results
A few things that change the numbers and trip people up:
Test voltage. Ratio error can vary slightly with applied test voltage, especially near the low end. When comparing a current test to a historical one, use a comparable test voltage. Comparing a 100 V test to an old 10 V test and attributing the difference to winding change is a mistake.
Lead resistance and connections. Poor lead connections produce inconsistent or noisy readings. Inconsistent results that change when you wiggle the leads are a connection problem, not a transformer problem. Clean, tight connections first.
Step-up vs step-down excitation. The convention of which winding you energize affects how some instruments present results. Be consistent, and know which way your instrument is configured, especially when comparing to historical data taken the other way.
Temperature. For trending, note the temperature. The ratio itself is fairly temperature-stable, but excitation current and the comparison guidelines assume comparable temperature between tests.
Safety. Even though the test uses low voltage, the transformer is a large inductor and capacitor. Follow lockout/tagout, ensure the unit is isolated and grounded before connecting, and be aware of capacitive discharge when disconnecting. The test voltage is low; the stored energy in a large transformer is not.
The Takeaway
The turns ratio test is simple to perform and easy to underuse. The discipline that separates a confirmatory glance from a real diagnostic test is small and concrete: read all three numbers, not just the ratio. Demagnetize before measuring excitation current. Test every tap and every phase. Read the pattern across phases, not just the individual values. Compare against history, not just tolerance, on in-service units. And when a result is off, use the pattern — global vs single-phase, ratio vs excitation vs phase angle — to localize the fault rather than just recording a fail.
Done that way, TTR catches wrong windings, shorted turns, bad tap contacts, reversed connections, and developing winding movement — a remarkable amount of diagnostic coverage from a test that takes seconds and uses a voltage you could touch. Done as a glance at the ratio number, it confirms almost nothing. The test is the same; the value is in how it’s read.