In this guide
A transformer that "works" is not the same as a transformer you understand. When you connect a transformer to a load, two things eat into its performance: the iron core wastes energy magnetising and re-magnetising itself sixty times a second, and the copper windings waste energy as heat because they have resistance. If you want to predict a transformer's efficiency or its voltage drop under load, you need numbers for both — and that is what the open-circuit, short-circuit and polarity tests give you.
These three tests are the standard experiments in every electrical machines laboratory, they appear in project reports wherever a custom transformer is involved, and they are among the most-asked viva topics in the subject. This guide explains what each test actually measures, why it is performed on the side it is performed on, and how the readings turn into efficiency figures.
What the tests are really after: the equivalent circuit
A real transformer can be modelled as an ideal transformer plus two parasitic branches:
- A shunt branch across the input — a resistance Rc (core/iron loss) in parallel with a reactance Xm (magnetising current). This branch draws a small current even with no load connected.
- A series branch — winding resistance R and leakage reactance X, which cause the voltage drop you see when load current flows.
You cannot measure Rc, Xm, R and X directly with a multimeter — they are tangled together inside the machine. The open-circuit test isolates the shunt branch; the short-circuit test isolates the series branch. That separation is the entire point.
Open-circuit test: measuring the iron loss
Setup: Energise one winding at its rated voltage with the other winding left open (no load). Measure input voltage, current and power with a wattmeter.
With the secondary open, no load current flows, so the copper loss (I²R) is negligible — the only current drawn is the small no-load current, typically 2–6% of rated current. Almost all the wattmeter reading is therefore the core loss (hysteresis + eddy currents), which depends on voltage and frequency, not on load.
Which side? The test is done on the low-voltage (LV) side at rated LV voltage, because test equipment for lower voltages and the moderate no-load current is easier to arrange safely, and rated flux is established with ordinary lab supply voltages. The high-voltage winding stays open.
What you get:
- Iron loss Pi (watts) — constant at a given voltage, present whether the transformer is loaded or not.
- No-load current I0 and, from it, the shunt-branch parameters Rc and Xm if you work through the phasor math.
A detail students miss: because the wattmeter reads nearly pure reactive magnetising current at a very low power factor, the test needs a low-power-factor wattmeter — an ordinary wattmeter gives a poor, inaccurate reading here.
Short-circuit test: measuring the copper loss
Setup: Short-circuit one winding through an ammeter, and apply a reduced voltage to the other winding — just enough to circulate rated current in the windings. Measure voltage, current and power.
Because the applied voltage is small (typically 5–10% of rated), the core flux is small and the iron loss is negligible. The wattmeter reading is therefore essentially the full-load copper loss (I²R in both windings).
Which side? The test is done on the high-voltage (HV) side, with the LV winding shorted. Rated current on the HV side is numerically smaller than on the LV side (same power, higher voltage), so the test current is easier for lab equipment to supply and the reduced test voltage is a modest, safe value.
What you get:
- Full-load copper loss Pcu (watts) — varies with the square of the load current, so at half load it is one-quarter of this value.
- The impedance voltage (%Z): the reduced voltage expressed as a percentage of rated voltage. This single number governs how much the secondary voltage sags under load and how large the short-circuit current would be — it is the figure printed on distribution transformer nameplates.
Worked example: from test readings to efficiency
Efficiency needs the two losses plus the output power. Suppose a single-phase transformer is rated 5 kVA, 230/110 V, and the lab tests give:
- Open-circuit test (230 V on LV side): wattmeter reads 120 W → iron loss Pi = 120 W.
- Short-circuit test (rated current circulated): wattmeter reads 180 W → full-load copper loss Pcu = 180 W.
At full load with a resistive load (power factor 1):
- Output = 5,000 W
- Total loss = 120 + 180 = 300 W
- Efficiency = 5000 / (5000 + 300) = 94.3%
At half load: output = 2,500 W, iron loss stays 120 W, copper loss falls to 180/4 = 45 W → efficiency = 2500 / 2665 = 93.8%. Notice the efficiency peaks where iron loss equals copper loss — that load point (here at √(120/180) ≈ 0.82 of full load) is called the maximum-efficiency condition, a favourite viva question.
(The numbers above are illustrative lab-scale values, not measurements from any particular machine — run the same arithmetic on your own lab readings.)
Polarity test: the one that prevents explosions
Transformers have dotted terminals marking which ends of the windings are instantaneously at the same polarity. If you parallel two transformers — common in labs and in distribution practice — and their polarities are mismatched, the secondaries fight each other and circulate a huge current limited only by winding impedance. The polarity test verifies the markings.
Simple voltmeter method: Connect one LV terminal to the adjacent HV terminal, apply a small AC voltage to the HV winding, and measure:
- V1: voltage across the HV winding
- V2: voltage across the LV winding
- V3: voltage between the two free terminals
If V3 = V1 − V2, the connected terminals have the same instantaneous polarity (subtractive polarity — the standard for distribution transformers). If V3 = V1 + V2, it is additive. Either is fine as a test result; what matters is that the actual polarity matches the nameplate dots before you parallel anything.
Sumpner's test: full-load heat without a full load
Sometimes you need the temperature rise at full load — for a large transformer, arranging a genuine full load is impractical. Sumpner's (back-to-back) test connects two identical transformers with primaries in parallel across rated voltage and secondaries in series-opposition. A small injected voltage circulates full-load current while each core sees rated flux, so both iron and copper losses appear as heat simultaneously. It needs two identical machines, which is why it is a special test rather than a routine lab one — but knowing it exists answers the "how would you test a 500 kVA transformer?" viva question.
Safety rules for the machines lab
- Short-circuit test means high current. Tighten every connection before energising; a loose shorting link arcs.
- Never exceed rated current in the SC test — you are deliberately pushing full current through the windings.
- Open-circuit test on the LV side still exposes the HV terminals at transformed voltage. Treat every terminal as live.
- Keep one hand in your pocket around energised panels, and never work alone on live equipment.
- Record meter ranges and multiplication factors before switching on — most lab "wrong readings" are range errors, not faulty transformers.
Frequently asked questions
Why is the OC test done at rated voltage but the SC test at reduced voltage?
Core loss depends on flux, which depends on applied voltage — so the OC test must use rated voltage to measure the true iron loss. The SC test only needs rated current to measure copper loss, and rated current is reached at a small fraction of rated voltage because the shorted winding is nearly a dead short.
Can I do the OC test on the HV side instead?
Physically yes, but you would need a high-voltage supply at rated HV voltage just to measure a tiny no-load current — impractical and more dangerous in a student lab. The LV side gives the same core-loss information.
Do these tests work for three-phase transformers?
Yes — the same principles apply per phase, with two-wattmeter or three-wattmeter power measurement. Your lab manual will specify the connection; the interpretation of the readings is identical.
What is the difference between %Z and the leakage reactance?
%Z is the magnitude of the total series impedance (resistance + leakage reactance combined), expressed as a percentage of rated voltage. It is what limits short-circuit current: a 5% impedance transformer fed from an infinite bus would see roughly 20 times rated current on a dead short.
Why does efficiency peak below full load?
Iron loss is constant while copper loss grows with the square of load. Total loss relative to output is minimised where the two are equal — for most distribution transformers this lands around 70–85% of full load, which is exactly where they spend most of their service life.
Limitations
These are steady-state, sinusoidal-supply tests. They do not characterise inrush current (the large magnetising surge at switch-on), harmonic behaviour with non-linear loads, or insulation strength — those need separate inrush, harmonic and high-voltage tests. The equivalent-circuit parameters are also temperature-sensitive: winding resistance rises as the machine heats up, so precision work corrects copper loss to a reference temperature (usually 75 °C). For student lab work and project documentation, the uncorrected values are the accepted standard.