Why √3 keeps showing up in three-phase AC
It is not a universal alternating-current multiplier. It is the length of a chord between two equal rotating quantities that sit one-third of a turn apart—geometry that becomes a line-to-phase rule only when the system is balanced and sinusoidal.
By Jordan Gold and Valery Gold
Walk into almost any shop class that mentions three-phase power and you will hear the same shortcut within the first ten minutes: multiply by the square root of three. Line voltage is phase voltage times √3. Line current in a delta is branch current times √3. Apparent power is √3 times line voltage times line current. The constant arrives so early, and works so often, that it starts to feel like a property of electricity itself—as if alternating current somehow “contains” √3 the way a circle contains π.
It doesn’t. √3 is what you get when you subtract two equal-length arrows that are 120° apart and then ask for the length of the result. That is a geometry fact first, a circuit fact second, and a power fact only after you insist on balance. Confusing it with the other famous root in AC work—√2, which converts a sine wave’s peak to its RMS value—is the fastest way to turn a clean derivation into a superstition.
This essay follows that geometry into wye voltages, delta currents, and the three-phase power formulas, with the same worked numbers used in the companion training deck. Every formula below sits next to a small animated model you can pause. Prefer reduced motion? Your system setting starts those models paused.
First, keep √2 and √3 in different pockets
A single-phase voltage written as a sine of peak magnitude Vpk has an instantaneous value that swings between ±Vpk. The quantity meters report for heating and most power calculations is the root-mean-square:
For a pure sinusoid centered on zero, that integral collapses to a memorable shortcut: Vrms = Vpk/√2. University of Louisville’s ECE 220 notes state it plainly, and they also warn that the √2 factor does not automatically transfer to other waveshapes—the integral definition still does. UMass Amherst’s open textbook makes the same separation.
That √2 story is about the intensity of one waveform. It has nothing yet to say about three phases, line-to-line spans, or delta nodes.
Three equal arrows, one-third of a turn apart
Balanced three-phase sources are three sinusoids of equal magnitude and frequency, displaced by 120°—exactly one third of a cycle. In peak form:
Under those equal-magnitude, equal-frequency, sinusoidal assumptions, the three waveforms sum to zero at every instant. KTH’s EI1120 Topic 13 notes show the phasor identity 1 + e−j120° + e−j240° = 0 behind that cancellation. The animated tracks below keep a shared time cursor so you can see the sum sit on the axis while the individual waves trade peaks.
Where √3 is born: a vector difference
Line-to-line voltage is not “phase plus phase.” It is a difference. MIT’s Kirtley notes write vab = va − vb and then reduce the balanced case to a line-to-line set whose magnitude is √3 times the line-to-neutral magnitude, leading the corresponding phase by 30°.
The same length falls out of the law of cosines. Take two equal RMS phasors of length V separated by angle θ. Their difference has squared length
At θ = 120°, cos θ = −1/2, so the right-hand side becomes 3V² and the length is exactly √3 V. Electrical PE Review’s √3 article walks the same tip-to-tail construction for exam audiences; the algebra matches Kirtley.
Sweep the angle in the model. Only the lock at 120° prints √3 on the ruler. That is the whole plot: √3 is a special value of a continuous geometric function, not a stamp applied to every AC problem.
Wye: the voltage multiplier
In a wye (star) connection, a phase quantity is a line-to-neutral span. A line quantity is a line-to-line span—the chord just derived. With balanced sinusoidal RMS values,
and, for the positive ABC convention with VAN at 0°, VBN at −120°, and VCN at +120°, the phasor VAB = VAN − VBN lands at √3 VLN ∠ +30°. Kirtley’s equations (24)–(26) are the time-domain version of that 30° lead. In a balanced wye the line current equals the phase current; the √3 factor is carrying the voltages, not the currents.
Delta: the current multiplier
Delta flips the story. Each branch already sees line-to-line voltage, so VLL = Vbranch. The √3 factor moves onto the currents. At a delta node the line current is a signed difference of two branch currents. With the reference arrows IAB (A→B), IBC (B→C), ICA (C→A) and IA entering node A from the line, Kirchhoff’s current law at A reads IA + ICA = IAB, so
Fiore’s LibreTexts §9.3 states the magnitude rule directly for balanced delta connections. The accompanying 30° lag of IA behind IAB (positive ABC, these arrows) is the current-side twin of the wye voltage’s 30° lead—another reminder that arrow conventions matter as much as the √3 sticker.
Power: one S, two writings
For one sinusoidal phase, average real power is Vrms Irms cos φ, and the displacement power factor is cos φ. Instantaneous p(t)=v(t)i(t) still wiggles at twice line frequency; negative lobes are energy returned to the source, not “negative RMS.” MIT OCW 6.622 Lecture 4 is careful that true power factor equals that cosine only when the current is also a pure sinusoid—otherwise true PF = (distortion factor)×(displacement factor).
Under balanced three-phase sinusoidal conditions the three double-frequency ripples cancel, leaving steady total real power, and the same apparent power can be written two ways:
Then P = S cos φ and Q = S sin φ, with S² = P² + Q². Schneider Electric’s Electrical Installation Guide lists the three-phase forms with the √3 line quantities and calls cos φ the displacement power factor for sinusoids. This essay’s sign convention follows the lagging-current / inductive Q>0 habit common in power-system teaching; leading current flips the sign of Q.
Worked numbers, without mixing peak and RMS
Wye example. Given VLN = 230 V RMS, IL = 10 A, PF = 0.80 lagging. Internally keep VLL = 230√3. Then S = 3·230·10 = 6900 VA = 6.90 kVA, which matches √3·VLL·IL. Real power P = 0.80·6900 = 5520 W. With sin φ = 0.60 for this lagging displacement factor, |Q| = 4140 var.
Delta example. Given VLL = 400 V RMS, Iph = 6 A, PF = 0.80 lagging. Line current IL = 6√3. Apparent power from either writing is S = 3·400·6 = √3·400·IL = 7200 VA. Then P = 5760 W and Q = 4320 var (lagging positive).
What the shortcut refuses to cover
Unequal phase magnitudes, angles that are not 120°, and non-sinusoidal currents all push you outside the √3 sticker. Unbalance needs symmetrical components or a full nodal solve. Harmonics need the broader power-factor definition from the MIT notes above. The quiz slide in the training deck deliberately breaks balance and adds a distorted trace so the shortcut fails in public—because the point of learning √3 carefully is knowing when to put it down.
References
- University of Louisville, ECE 220 Lesson 10 — sinusoids, ω=2πf, T=1/f; RMS definition; sine Vrms=Vpk/√2; warning that √2 does not transfer to arbitrary waveshapes.
engineering.louisville.edu/raise/EE220/L10.html - UMass Amherst, Applied Electrical Engineering Fundamentals §2.5 — RMS as √(mean square); sine factor; broader integral definition.
openbooks.library.umass.edu/funee/chapter/2-5/ - MIT / LibreTexts, Kirtley, Introduction to Electric Power Systems §3.3 — vab=va−vb; |VLL|=√3|VLN|; +30° lead; 120 V LN → 208 V LL example.
eng.libretexts.org/…/3.03:Line-Line_Voltages - Electrical PE Review (Zach Stone, P.E.), “Square Root 3 and Three Phase Power” — √3 from 120° phasor difference; wye voltage / delta current; S=√3 VLL IL; RMS phasor magnitudes.
electricalpereview.com/square-root-three-3-electrical/ - KTH EI1120, N. Taylor, Topic 13: Three-phase AC systems (PDF) — balanced sum to zero; |uab|=√3 Up; three-phase power forms.
kurs.ets.kth.se/…/ch13_threephaseac.pdf - Fiore, AC Electrical Circuit Analysis, LibreTexts §9.3 — delta: Vph=VLL, IL=√3 Iph.
eng.libretexts.org/…/9.3:Three-Phase_Connections - Schneider Electric, Electrical Installation Guide — Definition of reactive power — S, P=S cosφ, Q=S sinφ; three-phase √3 U I forms; displacement PF for sinusoids.
electrical-installation.org/…/Definition_of_reactive_power - MIT OCW 6.622 Power Electronics, Lecture 4 (Perreault), Spring 2023 — true PF = distortion × displacement when harmonics are present.
ocw.mit.edu/…/mit6_622_s23_lec04.pdf
Sources that failed to load during verification (Fiore MVCC PDF timeout; UC3M Module 4 HTTP 500) are omitted here on purpose.
There is a quieter lesson underneath the algebra. √3 shows up because three-phase power chose a particularly elegant compromise: equal magnitudes, equal spacing, and a closed set that sums to zero. The constant is the receipt for that compromise. Treat it that way—earn it from the triangle, spend it only while the assumptions hold—and the shortcut stays honest.
Take the animated deck with you
Twelve interactive Flux Lab slides, offline, with the same models and worked examples, plus learning objectives, a glossary, keyboard help and a self-check quiz. No account. No CDN.
Download the animated training deck (.html)Save the HTML file, then open it in a browser. The complete animated training deck works offline.
