Percent impedance is the one number on a transformer nameplate that tells you two things at once: how far the output voltage sags under load, and how much current the unit can push into a short circuit. On a 100 VA to 10 kVA dry-type transformer it is also the number most often missing from the plate, or copied from a datasheet written for 2,000 kVA switchgear. Here is what percent impedance measures and how to use it.
Percent impedance, written %Z, is the fraction of rated voltage you must apply to one winding — with the other short-circuited — to drive rated current through both. A unit marked 6% needs 6% of rated primary voltage to circulate full-load current through a bolted secondary short.
That impedance comes from two things: the resistance of the conductor, and the leakage reactance created by flux that links one winding without linking the other. The nameplate collapses both into one per-unit figure — convenient for calculation, useless for diagnosis. At small kVA ratings the split matters more than the total.
Every %Z value comes from one routine factory test. The low-voltage winding is shorted with a heavy link, a variable voltage is applied to the other winding, and the voltage is raised slowly until the ammeter reads rated current. That voltage, divided by rated voltage and multiplied by 100, is the percent impedance. The same test measures load loss. Because resistance rises with temperature, %Z is declared at a reference temperature, commonly 75 °C or 85 °C; a cold bench reading will be slightly low.
European and Asian plates print impedance voltage ou Uk; North American plates print %IZ; test reports say short-circuit voltage. Treat “Uz 4.5%” as %Z. Our guide to reading a transformer nameplate covers the neighbouring fields.
On a 100 VA to 10 kVA general-purpose transformer, %Z is frequently not printed at all. Standards for small transformers focus on insulation, temperature rise and dielectric strength, not on declaring impedance. Many units ship with kVA, voltage ratio, frequency, temperature rise, insulation class and a serial number, and nothing else.
| Nameplate marking | What it tells you | Typical format |
|---|---|---|
| %Z, %IZ, IMP | Percent impedance at rated current | 4.5% |
| Uk, Uz, impedance voltage | Same quantity, different name | Uz = 4.5% |
| kVA / VA | Base for rated current, which %Z is referenced to | 1000 VA |
| Temperature rise | Thermal limit, and the reference temperature for %Z | 60 K, Class F |
| Tap voltage | %Z is normally declared on the principal tap only | 0 / +5% |
If your plate has no impedance field, do not substitute a catalogue value from a similar unit — ask for the test report, or measure it.
Published impedance tables start at 15 kVA, where the normal band is roughly 1.5% to 10%. Below 15 kVA there is no standard band — exactly where most small step-up and step-down buyers operate. Measured values tend to look like this:
| Rating | Typical %Z | Share that is resistance | What dominates |
|---|---|---|---|
| 100 – 250 VA | 6 – 12% | 60 – 80% | Winding resistance |
| 500 VA – 1 kVA | 5 – 9% | 50 – 70% | Resistance, with growing reactance |
| 2 – 5 kVA | 4 – 7% | 40 – 60% | Balanced |
| 7.5 – 10 kVA | 3 – 6% | 30 – 50% | Leakage reactance starting to lead |
%Z generally falls as kVA rises, because a larger core allows more turns of thicker conductor and per-unit resistance drops. More importantly, the composition flips. A 2,000 kVA distribution transformer runs at an X/R ratio of 10 to 40, so %Z ≈ %X is a fair assumption. A 500 VA transformer runs closer to X/R of 1 to 3, so much of its impedance is plain resistance that heats the winding.
The drop from no load to full load, in percent, is the resistance component times the power factor plus the reactance component times the sine of the power factor angle:
Vdrop% ≈ %R × cos φ + %X × sin φ
Take a 1 kVA, 230/115 V transformer with %Z of 6.0% and measured %R of 4.2%. The reactance component is √(6.0² − 4.2²) = 4.3%. At full load and a lagging power factor of 0.8, the drop is (0.8 × 4.2) + (0.6 × 4.3) = 5.9% — about 108 V on a 115 V secondary, enough to upset equipment with a ±5% tolerance. At 60% load it drops 3.5% and lands near 111 V.
| Load (% of rating) | Approx. drop at 0.8 PF | Tension secondaire |
|---|---|---|
| 0% | 0% | 115 V |
| 25% | 1.5% | 113.3 V |
| 50% | 3.0% | 111.6 V |
| 75% | 4.4% | 109.9 V |
| 100% | 5.9% | 108.2 V |
This is why oversizing fixes most voltage complaints: a 2 kVA unit carrying 1 kVA runs at half load, so the drop halves with it. On motor loads the same impedance produces the starting dip, covered in sizing a step-up transformer for motor loads.
On large transformers %Z is primarily a fault-current limiter, and the calculation is one division: rated secondary current divided by %Z in per unit.
Isc = I_rated ÷ (%Z ÷ 100)
For the 1 kVA, 115 V unit above, rated secondary current is 8.7 A and %Z is 6%, giving about 145 A. That is correct and practically misleading, because it assumes an infinite source. In a real installation the branch breaker, supply cable and source impedance sit in series with the transformer, and at this scale their combined impedance is comparable to the transformer’s own. Measured fault current on a 1 kVA unit fed from a 10 A branch is a fraction of the calculated figure.
So you do not need a short-circuit study to protect a 10 kVA transformer; you need a primary device that tolerates inrush. Small units draw a magnetising surge of roughly 10 to 20 times rated current for 10 to 100 milliseconds, so a breaker sized at rated current trips on that surge. Size the primary device at 125% to 150% of rated primary current with a C-curve or D-curve magnetic threshold above the inrush peak. %Z barely enters that decision: inrush is set by the core’s saturation flux and the point on the voltage wave where the switch closes.
Transformers sharing a bus share load inversely in proportion to their impedance. A 4% unit and a 6% unit of identical kVA will not split evenly: the 4% unit carries roughly 60% and reaches its rating first. Match impedance within about 10%, along with voltage ratio and polarity.
Nameplate matching is not enough. Declared impedance carries a tolerance band under IEC 60076-1 of roughly ±7.5% to ±10%, so two units both marked 6% can legitimately measure 5.5% and 6.5% — near the edge of the pairing window before you connect them. If parallel operation matters, compare measured values from both test reports.
You need a variable transformer rated for the primary current, a true-RMS ammeter in the primary circuit, a voltmeter across the primary, and a shorting link rated for the secondary current.
Two companion tests belong in the same session. Measuring DC resistance gives you %R and lets you separate the components. A turns ratio test confirms the winding is intact first, since a displaced winding changes %Z while a shorted turn changes the ratio.
| Misreading | Why it happens | Correct approach |
|---|---|---|
| Using a 2,000 kVA catalogue value on a 1 kVA unit | Published tables start at 15 kVA and above | Measure, or use the test report |
| Treating %Z as %X | True for power transformers, false below about 10 kVA | Measure %R separately |
| Reading %Z as the voltage drop | %Z equals full-load drop only at zero power factor | Apply the %R/%X formula at the actual PF |
| Assuming a nameplate match means a parallel match | The tolerance band allows a ±7.5–10% spread | Compare measured values |
Because %Z is not a standard declaration below 15 kVA, the only way to get a usable number is to ask. State the kVA and voltage ratio, request the measured percent impedance and its reference temperature, ask for the resistance and reactance split if the load is power-factor sensitive, and require the test report with the shipment. For parallel operation, send the existing measured %Z and ask for a match within 10%.
Percent impedance is a voltage-drop figure first and a fault-current figure second, and at 100 VA to 10 kVA it is usually dominated by winding resistance rather than leakage reactance. Read it as the fraction of rated voltage that disappears at full load, apply the resistance and reactance split at poor power factor, and treat the fault-current formula as an upper bound, not a design value.
If you need impedance declared rather than assumed, our DT-1000VA step-up and step-down transformer shows the format of the data we supply, and the wider transformer range covers 100 VA to 10 kVA in single and three-phase builds. Send the load profile and we will confirm the rating that keeps the output in tolerance.
Applying 5% of rated primary voltage, with the secondary shorted, drives rated current through both windings. Equivalently, the unit loses about 5% of rated voltage across its own impedance at full load and near-zero power factor.
Yes. Percent impedance, impedance voltage, short-circuit voltage, Uk and Uz all describe the same quantity, and the percentage equals the per-unit value: 5% is 0.05.
General-purpose dry-type units typically measure 5% to 9%, with half to two-thirds from winding resistance. There is no standard band below 15 kVA, so the test report is the only authority.
Only the transformer-limited value: rated secondary current divided by %Z in per unit. That assumes an infinite source, so it overstates real fault current — dramatically below about 10 kVA.
Standards covering small transformers do not require an impedance declaration. Ask for the routine test report, which carries the measured value, or measure it yourself.
