Percentage impedance is the one nameplate figure that determines something the buyer never sees directly: how much fault current the transformer will deliver into a short circuit. It also sets the voltage drop the loads will experience, and it constrains which transformers can be run in parallel. Yet it is routinely left blank on enquiries and filled in by the supplier’s default.
This guide explains what transformer impedance measures, how to calculate the fault level it produces, what the typical values are by rating, and how the two competing effects — regulation and fault current — should be traded off before the LV switchgear is ordered.
Percentage impedance (written uk%, %Z or Z%) is the fraction of rated primary voltage required to circulate rated current through the transformer when the secondary is short-circuited.
The test that defines it is straightforward in principle: short the secondary, raise the primary voltage slowly, and record the voltage at which secondary current reaches its rated value. Express that voltage as a percentage of rated primary voltage and you have uk%. It is measured at low voltage, at low flux, and it is a routine measurement for every unit — unlike the short-circuit withstand test described later, which destroys the unit if it fails.
Physically, the number is dominated by leakage reactance: the portion of the magnetic flux that does not link both windings. It depends on winding geometry, the spacing between windings and the height of the coil, not on the amount of copper in any simple way. That is why two suppliers can offer the same kVA at the same impedance and still produce different efficiency, temperature rise and short-circuit strength.
For a fault at the transformer secondary terminals, the prospective symmetrical short-circuit current is:
Isc = In × 100 ÷ uk%
In = kVA × 1,000 ÷ (√3 × line voltage)
The relationship is inverse, which makes impedance the single most powerful lever on fault level. At 4% impedance the transformer delivers 25 times its rated current into a bolted fault; at 6% it delivers about 16.7 times. Halve the impedance and you double the fault current — and quadruple the mechanical forces on the windings and busbars.
| Rating (kVA) | Full-load current at 400 V (A) | Isc at 4% (kA) | Isc at 6% (kA) |
|---|---|---|---|
| 100 | 144 | 3.6 | 2.4 |
| 250 | 361 | 9.0 | 6.0 |
| 400 | 577 | 14.4 | 9.6 |
| 630 | 909 | 22.7 | 15.2 |
| 1,000 | 1,443 | 36.1 | 24.1 |
| 1,600 | 2,309 | 57.7 | 38.5 |
| 2,000 | 2,887 | 72.2 | 48.1 |
| 2,500 | 3,608 | 90.2 | 60.1 |
Two practical notes on reading this table. First, the figures assume an infinite upstream source; a real HV network has its own impedance, so the actual fault current is slightly lower, which means the table errs on the safe side for switchgear selection. Second, these are symmetrical RMS values — the first-cycle peak includes a DC offset and can reach roughly 2.5 times the symmetrical figure, which is what busbar bracing must withstand.
The cost consequence is direct: moving a 2,500 kVA unit from 6% to 4% takes the prospective fault level from 60 kA to 90 kA, and 90 kA switchgear is not a small premium over 60 kA. That single decision can outweigh the transformer price difference several times over.
Impedance is not arbitrary. IEC 60076-5 and the harmonised distribution standards establish conventional values by rating, and most utilities specify from this pattern:
| Rating range | Typical uk% | Why this value |
|---|---|---|
| Up to 630 kVA | 4% | Small units cannot generate dangerous fault levels, so impedance is kept low to favour voltage regulation |
| 631 – 2,500 kVA | 6% | The compromise point where fault level becomes the dominant constraint in LV switchgear selection |
| Above 2,500 kVA | 6 – 10% | Design dependent; higher ratings need higher impedance to keep fault current within switchgear ratings |
| Medium and large power transformers | 10 – 20% | Transmission-level fault levels would otherwise be unmanageable |
Non-standard values are available on request, and are sometimes necessary — but they add cost and lead time, and they must be agreed before the LV switchgear is specified. Retrofitting a different impedance after the switchgear order is placed is not a commercial option.
Where the utility sets a minimum uk%, it is protecting its own network and the customer’s switchgear. Three reasons appear repeatedly in connection agreements:
If your connection offer specifies a minimum impedance, that figure overrides any general table. Confirm it before the enquiry goes out — a supplier quoting a standard 4% unit against a 6% requirement will be rejected at the design review stage, often after the switchgear has been ordered.
The same impedance that limits fault current also causes voltage to fall as load rises. Approximate regulation can be calculated from the resistive and reactive components of impedance:
%VR ≈ loading × (ukr × cos φ + ukx × sin φ)
Where ukr = load loss in watts ÷ (10 × kVA), and ukx is derived from uk% and ukr. For a 630 kVA unit with 7,500 W of load loss, ukr is about 1.19%. At 0.85 power factor that gives:
| Loading | Regulation at 4% impedance | Regulation at 6% impedance |
|---|---|---|
| 50% | 1.5% | 2.1% |
| 75% | 2.3% | 3.1% |
| 100% | 3.0% | 4.1% |
A 1.1 percentage-point difference at full load looks small, but it accumulates along the distribution chain: transformer regulation plus cable drop plus the drop at the far end of a long feeder. On a 400 V system where the total permitted drop to the last load is typically 5%, the transformer choosing 6% impedance instead of 4% can consume a visible share of that budget — and push cable sizes up to compensate.
Tap changers recover some of this. An off-circuit tap changer set on the correct tap compensates for a steady voltage offset but not for load-dependent drop; an on-load tap changer does both continuously. How the tap range and type are specified for a given supply is covered in the oil-immersed selection guide.
When two transformers run in parallel, load sharing is inversely proportional to their impedance: the lower-impedance unit takes the larger share. If the mismatch is significant, one unit reaches its rating while the other is still lightly loaded, and the apparent spare capacity on site does not exist in practice.
The practical rules for parallel operation:
If the parallel pair is being bought together, specify the impedance in the enquiry for both units and require the routine test reports to confirm the measured values. Two units “both 6%” can differ by 10% in practice, and the tolerance in IEC 60076-1 permits it: the standard allows ±7.5% on declared impedance for the principal tapping where the declared value is 10% or above, widening to ±10% below that.
Measuring uk% is a routine test. Proving the transformer survives a short circuit is a type test under IEC 60076-5, and it is a different matter entirely.
The unit is subjected to a series of full-current short circuits — typically three, at reduced voltage, with the current limited by the test laboratory’s reactor rather than by the transformer. The test verifies both thermal withstand and mechanical withstand: the windings must not deform, the impedance must not change beyond the permitted tolerance after the test, and the unit must still pass its dielectric tests.
Three things buyers should understand about it:
How the test programme is structured, what the routine test report should contain, and where to draw the acceptance limits are covered in the factory test and FAT guide.
Impedance sets the fault level; the fault level sets the required breaking capacity of every device downstream; and the transformer’s own damage curve sets the upper limit for the protection that sits at its terminals. Three checks follow:
| Check | What it confirms |
|---|---|
| Breaking capacity of LV devices | Every breaker and fuse must exceed the prospective fault current at its installation point, using the transformer contribution plus motor contribution |
| Thermal withstand of cables and busbars | I²t let-through must be inside the cable’s short-circuit withstand for the clearing time |
| Transformer damage curve | The primary protection must clear a secondary fault before the transformer’s thermal damage curve is reached — otherwise the transformer protects the breaker instead of the reverse |
This is why impedance should be fixed early. It is not simply a transformer parameter; it is an input to the protection study, the switchgear specification and the cable sizing. Change it after any of those are ordered and the whole downstream design has to be rechecked.
The rating itself has to be settled before any of this, since impedance values are indexed to kVA. The kVA sizing method covers that calculation, and the nameplate reference shows where impedance, losses and temperature rise are declared on the plate you will receive. For dry-type units, the losses that determine ukr are also what the efficiency class regulates, so the two specifications should be written together rather than in sequence.
Percentage impedance is a system parameter wearing a transformer label. It sets the fault level your switchgear must survive, the voltage drop your loads will see, and whether two units can share load properly. Specify it explicitly, from the connection agreement and the fault level calculation, before the LV equipment is ordered — and ask for the measured value on your serial number, not just the design value.
If you can supply the rating, the secondary voltage and the utility’s requirement, the engineering team can confirm which impedance class keeps the fault level inside your switchgear while staying inside the regulation budget.
It means that 5% of rated primary voltage, applied with the secondary short-circuited, is enough to circulate full-load current. Practically, it means the transformer will deliver 20 times its rated current into a bolted fault at its terminals, and that its voltage will fall by roughly 3–4% at full load depending on power factor.
Divide the full-load secondary current by the per-unit impedance: Isc = In × 100 ÷ uk%. For a 1,000 kVA transformer at 400 V, full-load current is about 1,443 A, so at 6% impedance the prospective fault current is roughly 24 kA. The result is a symmetrical RMS value and should be increased by a factor of about 2.5 for first-cycle peak when checking mechanical withstand.
Neither, in isolation. Lower impedance gives better voltage regulation and lower losses in the supply path, but raises the fault current and therefore the cost of switchgear, busbars and cable withstand. Higher impedance limits fault current and protects downstream equipment but worsens regulation and may force larger cables. The right value is the lowest one that keeps the fault level inside the equipment you are prepared to buy.
Load sharing becomes inversely proportional to impedance, so the lower-impedance unit carries more than its share and can overload while the other is still lightly loaded. Common practice limits the difference to about ±10% of the lower value. Vector group and ratio must match exactly regardless; impedance matching only governs how well the pair shares load.
It should not. A measurable change in impedance indicates winding deformation, typically after a through-fault or a mechanical shock, which is why impedance is measured at commissioning and compared during later diagnostic testing. IEC 60076-5 sets a limit on the change permitted after the short-circuit withstand test for exactly this reason.
