Winding Resistance Test for Transformers: Method and Results

Of all the electrical tests performed on a power transformer, the winding resistance test is one of the simplest to perform and one of the most revealing. It measures the direct-current resistance of each winding, phase by phase, and compares the results with each other, with the factory test values, and with earlier site tests. A winding is, after all, a long length of copper or aluminium conductor with a set of bolted joints, leads and tap changer contacts in series with it. Anything that degrades those connections — a bolt that was never torqued properly, a contact that has worn, a joint that has overheated — shows up as extra resistance. The test finds those problems while the transformer is still out of service, when they are cheap to fix, instead of after energisation, when a hot joint can become a failure.

This guide explains the winding resistance test in plain language: why it is done, how the measurement works, how results are corrected to a reference temperature, how to interpret imbalance between phases, and what to do when a reading looks wrong. As with everything on this site, always work to your project specification, the manufacturer's instructions and your site safety rules; the general guidance here never overrides those documents. Winding resistance testing is done on a de-energised transformer, with the windings isolated, earthed and proved dead, by competent persons under permit. If you are new to transformer testing as a whole, start with our power transformer testing and commissioning guide, which puts this test in context alongside insulation, ratio and oil tests.

What the test is actually measuring

A transformer winding has a very low resistance — typically a fraction of an ohm on a large unit, and a few ohms at most on a small distribution transformer. That resistance comes almost entirely from the conductor itself: the length of the winding wire, its cross-section and its material. In series with the conductor are the internal leads, the bushings, the tap changer contacts and every bolted joint between them. The test measures the total of all of these, from one terminal to another, so it is really a test of the whole current path through the winding, not just the wire in the core.

That is why the test is so sensitive to connection problems. A healthy bolted joint adds almost nothing to the total. A loose or corroded joint adds resistance that can easily exceed the resistance of the winding itself, and it does so on one phase only, which makes it stand out when phases are compared. A tap changer contact that is pitted or misaligned adds resistance only on the taps it serves, which is why the test is repeated across tap positions. And a winding fault such as a shorted turn, while it usually shows up more clearly in other tests, also changes the resistance, which is one more reason the test is never skipped.

It is worth being clear about what the test does not do. It does not test insulation — that is the job of insulation resistance and tan delta tests. It does not prove the turns ratio — that is the ratio test. And it cannot see inside the winding the way frequency response analysis can. Its strength is narrow and specific: it proves the integrity of the current path, phase by phase and tap by tap, and it does so with equipment that is portable, quick and unambiguous.

How the measurement is made

The test uses a direct-current source and a four-terminal measurement, often called a Kelvin connection. Direct current is used because alternating current would be affected by the winding inductance and would not give a clean resistance reading. The test set injects a steady DC current through the winding — typically a modest current, chosen so the reading is stable without heating the winding — and measures the voltage drop across the winding with separate sense leads. Because the voltage is measured at the winding terminals themselves, the resistance of the test leads and their contact points does not contaminate the result. This four-wire method is essential: with only two wires, the lead resistance would swamp the tiny winding resistance being measured.

One practical complication is inductance. A large power transformer winding is a big inductor, and when DC is first applied, the current takes time to build up — sometimes minutes on a very large unit — before it stabilises and the resistance reading settles. Modern test sets handle this by using a higher test current and by detecting when the reading has stabilised, but the test engineer still needs patience: a reading taken while the current is still rising will be wrong. The same inductance matters at the end of the test. When the current is switched off, the stored magnetic energy must be discharged safely through the test set's discharge circuit, never by simply pulling a lead off, which would produce a dangerous voltage spike. Following the test set's discharge procedure, and waiting for it to complete before touching anything, is a basic safety discipline on this test.

Connections matter as much as the instrument. Test leads must be clean and firmly clamped to the bushing terminals, and the winding under test must be properly isolated from the rest of the installation. On a three-phase transformer, each phase is measured in turn, and on each winding — high voltage, low voltage, and tertiary where fitted. The neutral, where accessible, is part of the measurement path and its connection should be checked too. Sloppy connections at the test leads themselves are the most common cause of a "bad" reading that turns out to be nothing at all, so a suspicious result is always re-measured with fresh, clean connections before any conclusion is drawn.

Temperature correction to 75°C

Copper and aluminium resistance rises with temperature — by roughly four tenths of a percent per degree for copper — so a winding measured cold in the morning will read differently from the same winding measured warm in the afternoon. Factory acceptance tests are reported at a standard reference temperature, conventionally 75°C for oil-filled transformers, and site tests must be corrected to the same reference before any comparison is meaningful. Without correction, a perfectly healthy transformer tested on a hot day could appear to fail against its factory values, and a developing fault could hide behind a cold morning's low reading.

The correction needs one thing the test set cannot measure: the average temperature of the winding itself. In practice this is estimated from the oil temperature, measured at the top and bottom of the tank and averaged, after the transformer has been out of service long enough for the winding and oil temperatures to equalise — typically after standing de-energised for several hours, ideally overnight. The standard correction formula scales the measured resistance by the ratio of the reference-temperature constant to the measurement-temperature constant, using the material constant for copper or aluminium as appropriate. The arithmetic is simple, but the temperature estimate is where errors creep in: a winding that has recently carried load, or one side of which sits in the sun, is not at the oil temperature, and the correction will be wrong. Good practice is to record the top oil temperature, the bottom oil temperature, the ambient temperature and the time since de-energisation alongside every reading, so that anyone reviewing the results later can judge how reliable the correction is.

There is one more subtlety worth knowing. On a transformer with an on-load tap changer, the resistance changes slightly with tap position because the number of turns in circuit changes. That is expected and is not a fault. What is compared across taps is the pattern: resistance should change smoothly and gradually from tap to tap, following the turns count. A single tap that jumps out of that smooth pattern is the signature of a contact problem on that tap, and it is one of the most valuable things this test reveals.

Interpreting the results: balance, factory values and trends

Three comparisons turn raw readings into a verdict. The first is phase balance: on a healthy three-phase winding, the three phase resistances should be close to each other, allowing for the geometry of the winding — the middle phase of a three-limb core, for example, is physically shorter and legitimately reads slightly lower. Project specifications commonly allow a small percentage deviation between phases, and any phase that sits clearly outside that band deserves investigation. The second comparison is against the factory test values, corrected to the same temperature: a site reading that matches the factory value confirms the winding arrived and was installed without damage, while a significant departure points to transport damage, a connection left undone, or a tap changer problem.

The third comparison, and often the most powerful, is trend over time. A joint that is slowly deteriorating will show a resistance that creeps upward from one maintenance test to the next, long before it becomes a hot spot or a failure. That is why the results of every winding resistance test should be recorded in the transformer's permanent test history, with temperature, tap position and instrument details, rather than filed as a pass-or-fail tick. A single test proves the transformer is acceptable today; a history of tests proves it is not deteriorating, which is a much stronger statement.

When a reading is genuinely high, the investigation follows a logical order. Re-measure with clean connections to rule out a test artefact. Check the tap position and repeat on neighbouring taps to localise the problem to a contact or to the whole phase. Check the external connections — bushing terminals, cable lugs, neutral links — since these are accessible without opening the transformer. If the problem is internal, the next steps depend on the design: inspecting tap changer contacts, checking diverter and selector mechanisms (see our OLTC guide), and, in the worst case, internal inspection under outage. What should never happen is energising a transformer with a known unexplained resistance imbalance in the hope that it will "bed in" — a high-resistance joint under load only gets hotter, and hotter joints only get worse.

Common mistakes and how to avoid them

Most bad winding resistance results are not bad transformers; they are bad tests. Measuring before the current has stabilised gives a high, drifting reading. Dirty or loose test connections give a high reading on one phase that disappears when the leads are cleaned and re-clamped. Testing a winding that has not thermally stabilised, or guessing the winding temperature instead of measuring oil temperatures, makes the temperature correction meaningless. Testing with the tap changer between positions, or recording the wrong tap position, produces a result that cannot be compared with anything. And forgetting to discharge the winding properly at the end of the test is not a measurement error at all — it is a safety failure, and it is the one mistake on this test that can injure someone.

The defence against all of these is unglamorous: a written test procedure, a checklist that includes thermal stabilisation time and oil temperatures, clean leads, patience while readings stabilise, and results recorded with full context — date, tap, phase, temperature, instrument and tester. Winding resistance testing rewards discipline more than skill, and the teams that get reliable results are the teams that treat the checklist as part of the test, not as paperwork after it.

Frequently asked questions

Why is winding resistance corrected to 75°C?

Copper resistance rises with temperature, so readings taken at different winding temperatures cannot be compared directly. Correcting all readings to the standard reference temperature of 75°C puts factory tests, site tests and later maintenance tests on the same basis.

What does high winding resistance on one phase mean?

A single phase reading clearly above the others usually points to a high-resistance connection in that phase circuit: a loose bolted joint, a poor crimp, a damaged lead, or a worn or misaligned tap changer contact. It should be investigated, not averaged away.

Should winding resistance be measured on every tap position?

On transformers with an on-load tap changer, resistance is normally measured on all tap positions, or at least on the extreme and nominal taps plus any tap in service, because a faulty tap contact only shows up on the taps it serves. Off-circuit tap changers are measured on the tap in service and checked again if the tap is moved.