Delta connection Δ

What it is (precise)

Take three windings — of a transformer, a motor, a filter bank — and connect them end to end: the finish of phase A to the start of phase B, B to C, C back to A. The three junctions are the line terminals, and the triangle they form is the delta. Two ratios follow directly. Each winding sits directly across two lines, so winding voltage equals line-to-line voltage — there is no √3 on the voltage side.

Each line terminal is fed by two windings at once, so line current is the phasor difference of two winding currents 120° apart: √3 times the winding current, about 1.732. The wye (star) connection is the exact mirror — √3 on voltage, none on current — and that entry carries the other half of the story.

Multiply through and the connection drops out of the power arithmetic: 3 × V(winding) × I(winding) × cos φ equals √3 × V(L-L) × I(line) × cos φ for delta and wye alike, so a 690 V, 5 MVA nameplate commits the machine to the same line current — about 4,180 A — whichever way its windings are wired. What the connection does decide is what each winding endures.

A delta winding is insulated for the full line-to-line voltage but carries only 58% of the line current; the equivalent wye winding sees 58% of the voltage and the full current. That trade is why the connection matters to the people who wind transformers and size copper, and why a thermal or ampacity check that confuses winding current with line current is off by 73% before it starts.

On paper the delta is the D in the transformer connection codes — capital for the higher-voltage winding, lowercase for the lower, so Dyn11 has the delta on the HV side and YNd11 on the LV — and a delta circuit is inherently three-wire: three phases, no fourth conductor to run. How the letters pair up, what the clock numbers mean and which combinations grid, inverter vendor and protection scheme will each accept is the transformer vector group entry's territory; this page owns what the triangle itself does.

The closed loop: where triplen currents go

The delta's defining quirk is that it is a closed circuit all by itself. Zero-sequence currents — components equal in magnitude and in phase on all three phases, which is exactly what the triplen harmonics (3rd, 9th, 15th…) of transformer magnetizing current look like — cannot leave through the line terminals, because at every corner the identical currents in the two adjoining windings cancel. Around the loop, though, those same in-phase components add, so the current circulates inside the triangle instead.

The plant-level consequences follow: a delta winding blocks zero-sequence and triplen currents from passing between transformer windings, which is why star-star units carry a supplementary stabilizing delta and why a fault study sees an open circuit looking into a delta winding — the sequence theory behind both lives under zero-sequence.

The circulating current is still real current in real copper, heating the winding that traps it. And the trap is selective: the 5th, 7th, 11th and 13th orders a PCS actually emits are not zero-sequence and pass straight through — see harmonics for the inverter's spectrum and what manages it.

No neutral point

A delta has three terminals and nothing else — no fourth point to ground, no line-to-neutral voltage to tap. That absence works through a storage plant's design twice. First, auxiliaries: control cabinets, lighting and small UPS supplies want single-phase line-to-neutral service, and a delta-fed board cannot provide it; the neutral has to be manufactured by a transformer winding somewhere, one more reason the auxiliary architecture deserves its own line on the one-line diagram rather than an assumption.

Second, grounding: a bus fed only by delta windings has no earth reference at all. On a storage one-line this appears as a Dyn11 step-up unit with the delta on the 34.5 kV side plus a separate grounding transformer on the collector bus, its manufactured neutral taken to earth through a resistor. Which side of the transformer the delta occupies, and what that decides about where the ground source sits, is the vector-group decision; the delta itself simply cannot supply one.

Left floating, a delta system fails in a characteristically deceptive way. The first line-to-ground fault draws almost nothing — there is no return path except stray capacitance — so nothing trips and the plant keeps running. What changes is the stress: the two healthy phases now stand at full line-to-line voltage above earth, √3 times their normal level, and every arrester, cable termination and instrument transformer on the bus carries that stress until the fault is found.

A second fault on another phase then completes a phase-to-phase short through the earth. The engineering answer is not to leave it floating: establish the earth reference the interconnection requires, fit ground-fault detection that can see a high-impedance first fault, and put the grounding transformer's in-service status on the commissioning checklist before the collector bus is first energized.

Where the delta appears in a plant

In a BESS the delta's home is the step-up transformer. The PCS-side low-voltage winding is one place it sits, where it isolates the inverter from collection-system zero sequence and leaves the earth reference on the MV winding. Medium-voltage practice varies by market: many plants run grounded wye on the MV winding (YNd11), grounding the collection system directly, while others put the delta on the MV side (Dyn11) and add the collector grounding transformer described above.

Star-star units, where they appear, typically carry a buried stabilizing delta — a third winding whose whole job is the zero-sequence path of the previous section. Reconciling which arrangement the utility, the inverter vendor and the protection scheme will each sign off is exactly what the transformer vector group entry walks through; the point here is that every one of those choices is a decision about where to put a triangle.

The other habitat is motor loads. The plant's chillers, coolant pumps and HVAC compressors are three-phase induction motors, and a motor nameplate routinely offers both connections of one winding: 400/690 V Δ/Y means the winding is rated for 400 V — connect it in delta on a 400 V system, in star on a 690 V one, and the √3 between the two systems is absorbed by the connection change.

The same ratio powers the classic star-delta starter: start connected in star so each winding sees 1/√3 of rated voltage and the motor draws one-third of its delta starting current, then switch to delta to run. Reading those nameplates correctly is auxiliary-commissioning work, and getting the connection wrong is not cosmetic — the pitfalls below quantify it.

Common pitfalls

The √3 errors come in mirrored pairs. On the voltage side, remembering that three-phase has a √3 in it somewhere and dividing a delta system's 690 V by 1.732 invents a 400 V line-to-neutral supply that does not exist.

On the current side, quoting winding current where line current belongs — or the reverse — misses by 73%, enough to turn a passing cable-ampacity or winding-thermal check into a failing one. The discipline is the same as everywhere else in three-phase work: establish which quantity a number is before it goes into a formula, and remember that on a delta the √3 lives on the current.

The motor nameplate is the pitfall with smoke. Connect a 400/690 V Δ/Y motor in delta on the 690 V board and every winding sees √3 times its rating; connect it in star on the 400 V board and each winding gets 1/√3 of rated voltage, torque falls with voltage squared to about a third, and the loaded motor stalls or overheats trying.

Two further traps close the list. Planning a line-to-neutral auxiliary tap from a delta bus assumes a terminal that is not there — the fix is a transformer, and it belongs on the one-line early. And an ungrounded delta's first ground fault announces itself to nothing; if the detection scheme was value-engineered out, the first indication is the second fault.