Technical explainer
Three-phase transformer connections: delta, wye, and what each one decides
A three-phase transformer's winding connection is chosen independently on each side: delta or wye on the primary, delta or wye on the secondary. That gives four standard combinations — delta-delta, delta-wye, wye-delta, wye-wye — and each one settles a different question: the voltage and current relationship at the terminals, whether a neutral point exists at all, the phase shift between the two sides, and how the unit behaves under unbalanced load. A delta-delta connected transformer is one answer among the four, not the default one.
How delta and wye change what you get at the terminals
The winding connection sets a fixed arithmetic relationship between line and phase quantities, and that relationship is different on each side.
In a wye winding, each winding sits between one line conductor and the common star point, so the voltage across the winding — the phase voltage — is the line voltage divided by √3. In a delta winding, each winding sits directly between two line conductors, so phase voltage and line voltage are the same number. Current behaves the opposite way: a wye winding carries the full line current, while a delta winding's phase current is the line current divided by √3, because each line conductor feeds two windings at once.
That is the entire vocabulary the rest of this article works with. It says nothing about how the transformer is built — dry-type versus oil-immersed construction is a separate question from the one this article answers.
The four connections compared
| Connection | Secondary voltage relationship | Neutral available? | Phase displacement (primary → secondary) | Harmonic / unbalance behaviour | Typical application |
|---|---|---|---|---|---|
| Delta-delta | Secondary line voltage in phase with primary line voltage | No, on either side | 0° | Third-harmonic magnetising current stays confined to the delta loop; secondary line voltages stay near-constant even under unbalanced load | Three-wire distribution feeders; industrial loads with no need for a neutral |
| Delta-wye | Secondary shifted from primary by a nominal 30° | Yes, on the wye (secondary) side | ±30°, direction set by winding convention | Delta primary confines third-harmonic current locally | Step-up at generating stations; general step-down to a four-wire feeder |
| Wye-delta | Secondary shifted from primary by a nominal 30° | Yes, on the wye (primary) side only | ±30°, direction set by winding convention | Delta secondary confines third-harmonic current locally | Step-down at a receiving substation from a higher transmission voltage |
| Wye-wye | Secondary line voltage in phase with primary line voltage | Yes, on both sides (if brought out) | 0° | Star point shifts under unbalanced load, unbalancing phase voltages; third-harmonic components add rather than cancel unless corrected | Rarely chosen for major power transformation despite lower insulation cost |
Delta-delta: no neutral, strong under unbalance
A delta-delta connection introduces no phase displacement at all: the secondary voltage triangle sits in the same rotational position as the primary, so primary and secondary line voltages stay in phase.
Neither side has a neutral point, which is the connection's defining limitation. A delta-delta connected transformer cannot directly feed a three-phase, four-wire load that needs a line-to-neutral connection, because there is no neutral to connect to.
What it gives up in flexibility, it makes up in tolerance for uneven loading. Under an unbalanced load, a delta-delta bank's three secondary line voltages stay close to constant — a genuine advantage over wye-wye, where the same imbalance shifts the star point and pulls the phase voltages apart.
Wye-wye: cheap insulation, rarely chosen
A wye winding only has to insulate for the phase voltage, which is the line voltage divided by √3 rather than the full line voltage. For a small, high-voltage unit that looks like an attractive saving.
The saving comes with a condition attached: the connection only behaves correctly when the load is balanced. An unbalanced load shifts the star (neutral) point, and that shift unbalances the three line-to-neutral phase voltages even though the line-to-line voltages barely move.
There is a second problem, independent of loading. Each phase's magnetising current carries a third-harmonic component, and because there are exactly three cycles of that component for every one cycle of the fundamental, the three phases' third-harmonic components land in phase with each other in a wye winding rather than cancelling. Left uncorrected, they can add up to a third-harmonic voltage larger than the fundamental itself.
Two fixes exist. Solidly grounding the star point gives the imbalance current and the third-harmonic current somewhere to return through. Adding a delta-connected tertiary winding — typically sized at roughly a third of the main winding's capacity — gives the third-harmonic component a closed loop to circulate in instead. Between the balanced-load requirement and the harmonic behaviour, wye-wye is rarely selected for major three-phase power transformation, despite its insulation advantage.
Delta-wye and wye-delta: the step-up / step-down pair
Delta-wye — delta on the primary, wye on the secondary — is the standard connection for stepping up at a generating station. The delta primary needs fewer turns at the comparatively low generator voltage, and the wye secondary's grounded neutral supports four-wire service at the higher transmission voltage.
Wye-delta is the mirror image, used to step down at the receiving end. The wye primary's grounded neutral suits the higher incoming transmission voltage, and the delta secondary serves distribution loads that have no need for a transmission-side neutral.
Under the ANSI/IEEE "standard 30-degree" convention, a delta-grounded-wye step-down transformer has its high-voltage line-to-line voltage and line current leading the low-voltage side by 30°; a step-up connection of the same delta-y transformer type reverses which side leads.
The delta-grounded-wye step-down connection is the one usually feeding a four-wire wye distribution feeder, specifically because the wye secondary gives three independent single-phase circuits, which lets single-phase loading be balanced across the bank. A three-wire delta feeder, by contrast, is typically served from a delta-delta bank at the substation.
Worked example: sizing a delta-wye transformer
Take a 1000 kVA, 50 Hz distribution transformer with an 11 kV delta primary and a 400 V grounded-wye secondary — round figures chosen for this article, not taken from any source.
Primary side (delta). Line voltage and phase voltage are the same quantity on a delta winding: 11,000 V either way. Rated line current follows from S = √3 × VL × IL:
IL(primary) = 1,000,000 / (√3 × 11,000) ≈ 1,000,000 / 19,053 ≈ 52.5 A
Because it is a delta winding, phase current is the line current divided by √3:
Iphase(primary) = 52.5 / √3 ≈ 30.3 A
Secondary side (wye). Line voltage is 400 V, but phase voltage — the line-to-neutral voltage each winding actually sees — is line voltage divided by √3:
Vphase(secondary) = 400 / √3 ≈ 231 V
Rated line current on the secondary:
IL(secondary) = 1,000,000 / (√3 × 400) ≈ 1,000,000 / 693 ≈ 1443 A
On a wye winding, phase current equals line current, so the winding itself also carries approximately 1443 A.
Turns ratio. The turns ratio is set by the two winding (phase) voltages, not by the two line voltages:
Turns ratio = 11,000 / 231 ≈ 47.6
That is a different number from the line-voltage ratio, 11,000 / 400 = 27.5. A reader who assumes the nameplate line-voltage ratio is the winding turns ratio will be off by a factor of √3, because the √3 applies on the wye side and not on the delta side — which is exactly the distinction the rest of this article has been building toward.
Phase displacement and why it blocks paralleling
Any delta-wye or wye-delta transformer imposes a structural phase shift of a nominal 30° between primary and secondary line quantities. It is a consequence of how a line voltage relates to a phase voltage differently on the delta side than on the wye side — not a feature deliberately designed into the windings.
The direction of that shift, lead or lag, depends on which winding-to-line labelling convention is used, and reverses if the sense of the delta connection changes. That is why engineers distinguish "+30°" and "−30°" connections rather than treating the shift as fixed in one direction.
The practical consequence is that a delta-wye or wye-delta bank cannot be run in parallel with a wye-wye or delta-delta bank of the same nominal ratio: the two banks' secondary phase angles will not coincide even when the voltage magnitudes match. That mismatch leaves a voltage difference across the paralleling point even with no load connected, and closing onto it drives a circulating current between the units. Two delta-wye banks — or two wye-delta banks — can only be paralleled with each other if both are wired to produce the same direction of shift, which is worth checking explicitly when specifying units meant to run together.
Reading the vector group off the nameplate
Vector-group notation is how a nameplate records both the connection and the phase displacement. Under IEC 60076-1, an upper-case letter gives the higher-voltage winding's connection, a lower-case letter gives the lower-voltage winding's, and an "N" or "n" is appended if that winding's star point is brought out. A clock-hour figure from 0 to 11 follows: the high-voltage phasor is fixed at the 12 o'clock position, and each hour of the low-voltage phasor's position represents 30° of displacement.
ANSI/IEEE C57.12.00 records the same physical relationship differently — descriptively, for example as "delta-wye, 30° lag" — rather than by clock hour. The same transformer's nameplate therefore reads differently depending on whether it follows IEC or ANSI/IEEE practice, even though the underlying 30°-multiple relationship is identical physics. Worth checking before assuming two nameplates from different markets describe different hardware.
Third-harmonic current and why the delta winding matters
The payoff comes first: a transformer with a delta winding somewhere in it delivers a sinusoidal output voltage even though the current magnetising its core is not. A delta winding is what makes that possible, by giving the third-harmonic component of that exciting current a closed local path to travel instead of forcing it out onto the line conductors.
The reason is a matter of timing rather than geometry. Across one cycle of the fundamental, the three phases are staggered evenly apart; a third-harmonic wave, though, completes three full cycles in that same span, so the stagger that keeps the phases distinct at the fundamental collapses once the waveform is viewed at that tripled rate — each phase's third-harmonic component ends up aligned with the other two rather than spread out. With nothing left to drive it toward the line conductors, that aligned component has nowhere to go but around the closed delta loop itself, leaving the core flux, and so the terminal voltage, sinusoidal.
The same closed-loop property explains a second effect. Facing zero-sequence excitation — the common-mode component associated with third-harmonic magnetising current or a single-line-to-ground fault — an ungrounded wye winding has no closed path for that current and blocks it outright, the equivalent of an open circuit. A delta winding is the opposite case: its closed loop gives the same current somewhere to flow, so it behaves as a short circuit instead. That is the underlying reason a delta winding, sitting across the core from a wye winding, stops both third-harmonic current and ground-fault zero-sequence current from passing through to the far side. It is also where the transformer impedance sets the downstream fault level: what the connection blocks or passes through determines what fault current is even available on the far side to begin with.
When one transformer fails: the open-delta option
If one unit of a delta-delta bank built from three separate single-phase transformers is lost, the remaining two can keep delivering three-phase power in an open-delta, or V-V, arrangement — at reduced capacity, until the third unit is replaced.
The capacity available is 1/√3, or 57.7%, of the original closed bank's rating. That is not two-thirds, which is the intuitive but wrong guess from "two of three units remain." The shortfall happens because in V-V operation each secondary line current becomes equal to the full secondary phase current of that unit, rather than √3 times a proportionally smaller phase current as in the closed bank — so the two remaining units cannot jointly reach two-thirds of the original rating without exceeding their own individual current rating.
Questions buyers ask
What is the difference between a delta-delta and a delta-wye transformer connection?
Delta-delta keeps primary and secondary line voltages in phase, with no neutral on either side. Delta-wye shifts the secondary by a nominal 30° from the primary and gives the wye side a neutral suited to four-wire loads.
Can I parallel a delta-wye transformer with a delta-delta transformer?
Not at the same ratio without a problem: the delta-wye bank's secondary is displaced 30° from its primary, so its secondary phase angle will not match a delta-delta bank's even with equal voltage magnitudes, which drives circulating current if the two are closed together.
Why does a delta-wye transformer show a 30° phase shift?
The shift is a structural result of how line voltage relates to phase voltage differently on the delta side than on the wye side — not a deliberate design choice. Its direction depends on winding-labelling convention.
What happens if one transformer fails in a delta-delta bank?
The remaining two can continue supplying three-phase power in an open-delta (V-V) arrangement, but only at 57.7% of the original bank's rated capacity, not two-thirds.
How do I read a transformer vector group like "Dyn11"?
The letters give the HV/LV connection — upper-case for the higher-voltage winding, lower-case for the lower-voltage winding, with "n" added if the low-voltage neutral is brought out. The number is a clock hour: the high-voltage phasor is fixed at 12, and each hour represents 30° of low-voltage-side displacement.
Standards referenced
- IEC 60076-1, Power transformers — Part 1: General (vector-group clock notation)
- ANSI/IEEE Std C57.12.00 (30-degree connection convention; additive/subtractive polarity)
Sources
- Aditya Engineering College (AEC), Unit 5: Three Phase Transformers, course notes. aec.edu.in
- Philadelphia University, Electric Machines I – Three Phase Transformers (Dr Firas Obeidat). philadelphia.edu.jo
- Rohini College of Engineering & Technology, EE8301 Electrical Machines-I, §2.6. rcet.org.in
- NIT course material, Ch. 33: Transformer — Three Phase. nit-edu.org
- MIT OpenCourseWare 6.061, Introduction to Electric Power Systems, Class Notes Ch. 4 "Introduction to Symmetrical Components" (J.L. Kirtley Jr.). ocw.mit.edu
- Iowa State University, EE653 Modeling Three-Phase Transformers (Dr Zhaoyu Wang). iastate.edu
- Purdue University, ECE 595 Power Distribution System Analysis, Lecture 9: Three-Phase Transformer Models (Vassilis Kekatos). purdue.edu