Two-Wattmeter Method: Measuring Three-Phase Power

A three-phase induction motor's nameplate says 15 kW. Your lab bench has two single-phase wattmeters, a 415 V three-phase supply, and no three-phase power analyser in sight. Buying one is not an option — and it turns out you do not need one. The two-wattmeter method is the standard way to measure total three-phase power with two ordinary single-phase wattmeters, and it works for star or delta loads, balanced or unbalanced, as long as the system is three-wire (no neutral current).

This guide covers the connections (the one thing you must get exactly right), the short derivation that makes the result trustworthy, the power-factor formula hidden in the two readings, what to do when one meter reads negative, the lab procedure step by step, and the mistakes that waste an afternoon.

Why two wattmeters are enough

A wattmeter has two coils: a current coil (low resistance, wired in series with the line, like an ammeter) and a pressure/voltage coil (high resistance, wired across two points, like a voltmeter). Its deflection is proportional to V × I × cos θ — voltage across its own terminals, current through its own coil, and the angle between them.

The deep reason two instruments suffice is Blondel's theorem: in an N-wire system, N−1 wattmeters correctly measure total power, provided the voltage coils share a common reference point. For a three-wire system that means two wattmeters. No assumption about balance is needed for the power result — the two readings always add up to the true total power. (Balance only matters for the power-factor formula later.)

Intuition: in a three-wire system the three line currents must sum to zero (no neutral to carry the remainder), so the third wire's power contribution is already encoded in the other two lines' measurements. Once you see the derivation below, this stops feeling like a trick.

The connections: the one picture to get right

Label the three supply lines R, Y, B (any three phases; L1/L2/L3 works the same).

• Wattmeter W1: current coil in series with line R; voltage coil connected between R and Y.
• Wattmeter W2: current coil in series with line B; voltage coil connected between B and Y.

Line Y is the common reference for both voltage coils — the middle line carries no current coil. Each wattmeter's terminal markings (±, or an asterisk/star on electrodynamometer instruments) go toward the supply side for the current coil, and the corresponding voltage terminal toward the same line — this sets the polarity so upscale readings mean positive power under normal lagging loads.

Two wiring discipline rules:

1. Current coil in series, voltage coil in parallel — never swapped. A current coil across the line is a near-short across 415 V. This is the wiring mistake that destroys instruments, so draw the full circuit on paper and check it against the diagram before energising.
2. Keep both wattmeters' voltage-coil common point on the same line. If W1's voltage coil spans R–Y but W2's spans B–R, the readings do not add up. Same common line, always.

For background on the instrument transformers you will meet at higher currents, see the CT and PT guide, and for reading the panel and wiring itself, the three-phase wiring basics guide.

The math: why W1 + W2 is the total power

Take a balanced load with power-factor angle φ (current lagging voltage by φ). With the connections above, the phasor analysis gives:

• W1 = V_L × I_L × cos(30° − φ)
• W2 = V_L × I_L × cos(30° + φ)

(V_L and I_L are line voltage and line current.) Adding them:

W1 + W2 = V_L I_L [cos(30° − φ) + cos(30° + φ)] = V_L I_L × 2 cos 30° cos φ = √3 V_L I_L cos φ

which is exactly the standard three-phase power formula. The derivation assumed balance, but the result P = W1 + W2 holds for unbalanced three-wire loads too — that is Blondel's theorem doing its work.

Where do the ±30° come from? Each wattmeter's voltage coil sees a line-to-line voltage (R–Y, B–Y), which leads the corresponding phase voltage by 30°, while its current coil sees the line current lagging the phase voltage by φ. The angle each meter actually multiplies is the difference: 30° ∓ φ. Sketch the phasor diagram once and this clicks permanently.

The hidden bonus: power factor from two readings

Subtracting instead of adding:

W1 − W2 = V_L I_L [cos(30° − φ) − cos(30° + φ)] = V_L I_L × 2 sin 30° sin φ = V_L I_L sin φ

Divide by the sum and the V_L I_L cancels:

tan φ = √3 × (W1 − W2) / (W1 + W2)

So two readings give you total power and power factor — no separate pf meter needed. (This formula does assume a balanced load; the power sum does not.)

Worked example. A 400 V motor test gives W1 = 8 kW, W2 = 5 kW.

• P = 8 + 5 = 13 kW
• tan φ = √3 × (8 − 5) / (8 + 5) = 1.732 × 3 / 13 ≈ 0.400 → φ ≈ 21.8° → power factor ≈ 0.93 lagging
• Cross-check current: I = P / (√3 × V_L × pf) = 13000 / (1.732 × 400 × 0.93) ≈ 20.2 A — which should roughly match your clamp meter.

Reading the ratio at a glance: W1 = W2 means pf = 1 (φ = 0). W2 = 0 with W1 positive means pf = 0.5 (φ = 60°). Between those, the bigger the split between W1 and W2, the lower the power factor. And when W2 goes negative, you are below pf 0.5 — which brings us to the classic lab moment.

When W2 reads negative (power factor below 0.5)

Look at W2 = V_L I_L cos(30° + φ). When φ exceeds 60° (pf below 0.5), the angle inside the cosine passes 90° and the reading goes negative. On an analog electrodynamometer wattmeter the pointer slams against the zero stop; on a digital meter it shows a minus sign.

This is normal, not a wiring fault. The fix: reverse the connections of one coil of W2 (voltage coil is easiest and safest), take the reading, and subtract it:

P = W1 − |W2|

Example: W1 = 6 kW, W2 reversed reads 2 kW → P = 6 − 2 = 4 kW. Then tan φ = √3 × (6 − 2) / (6 − (−2))... careful: use the signed values, W2 = −2 kW: tan φ = √3 × (6 − (−2)) / (6 + (−2)) = 1.732 × 8 / 4 ≈ 3.46 → φ ≈ 73.9° → pf ≈ 0.28. A heavily reactive load, exactly the situation that drives W2 negative.

Why does reversing work? You flipped the sign of the product the meter multiplies, so the deflection is positive again; the true (negative) value is minus the reversed reading, and Blondel's sum uses the true value.

The four-wire problem: when two wattmeters are NOT enough

The two-wattmeter method assumes no neutral current — a three-wire system. If your load has a neutral connection carrying current (unbalanced single-phase loads on a 4-wire supply), the theorem's precondition fails and you need the three-wattmeter method: one wattmeter per phase, each voltage coil from its line to neutral, P = W1 + W2 + W3. For a balanced four-wire load, one wattmeter × 3 suffices.

Lab rule of thumb: motor and three-phase heater loads are three-wire — two wattmeters. Mixed lighting/socket loads with a neutral — three wattmeters.

The lab procedure, step by step

1. Draw it first. Sketch supply → W1 current coil (line R) → load, supply → W2 current coil (line B) → load, voltage coils R–Y and B–Y, and mark every ± terminal toward the supply.
2. De-energise and wire. Current coils in series with their lines; voltage coils across the line pairs; common reference line Y on both voltage coils. Double-check that no current coil sits across the supply.
3. Zero-check analog meters and note each meter's range (choose current ranges above the expected line current — a 15 kW motor at 415 V draws roughly 25–30 A at typical power factors).
4. Energise, read both meters, note the signs. If W2 pins below zero, de-energise, reverse W2's voltage-coil leads, re-energise, and record it as negative in your calculation.
5. Compute: P = W1 + W2 (signed), tan φ = √3(W1 − W2)/(W1 + W2), pf = cos φ.
6. Cross-check against the nameplate: measured P should be in the right neighbourhood of the expected shaft-plus-losses power, and the computed current should match a clamp meter within a few percent. A factor-of-√3 error almost always means line voltage was used where phase voltage was needed, or vice versa.

If you are sizing the correction capacitors for a lagging load you just measured, the power-factor capacitor sizing guide takes it from here.

Six mistakes that waste an afternoon

1. Voltage coils on different common lines — W1 across R–Y, W2 across B–R. The readings are then meaningless together. Same common line, always.
2. Adding a reversed reading instead of subtracting. After reversing W2's coil, its true value is negative. P = W1 − |W2|, not W1 + |W2|.
3. Using it on a four-wire system with neutral current. Two wattmeters cannot see neutral power. Check for a neutral connection before you start.
4. Ignoring meter polarity. Swapping both coils' polarity on one meter double-negates back to correct — but swapping only one inverts that meter silently. Mark ± terminals as you wire.
5. Wrong range, pegged pointer. An undersized current range saturates and reads low; an oversized range wastes resolution. Estimate line current first.
6. Forgetting the CT secondary rule at high current. Above direct-metering ranges you measure through current transformers — and a CT secondary must never be open-circuited while energised (dangerous voltages develop). Short it when the meter is disconnected. See the CT/PT guide.

FAQ

Can I use digital / multifunction meters instead?
Yes — a digital power meter on each of the two positions reads signed power directly, so the reversal step becomes just reading the minus sign. The connections and the maths are identical.

Does it work for delta-connected loads?
Yes. The method sees only the three lines; star vs delta inside the load does not matter, for balanced or unbalanced three-wire loads.

Can I get reactive power too?
For a balanced load, Q = √3 × (W1 − W2). (That is the same subtraction that feeds the tan φ formula, scaled.) For unbalanced loads, the two readings do not separate P and Q cleanly.

Why do analog wattmeters have ± or starred terminals?
The deflection direction depends on the relative polarity of the two coils. The marks define "positive" power flow (supply → load); wire them toward the supply and upscale means the load is absorbing power.

My two readings are equal. Is something wrong?
No — W1 = W2 means the load is at unity power factor (a pure heater load, for instance). Equal readings are the signature of pf = 1.

Does harmonic distortion break it?
Electrodynamometer wattmeters respond to true instantaneous V×I, so the power sum stays correct even with harmonics; the tan φ power-factor formula, though, assumes sinusoidal balance and degrades as distortion rises.

Limitations

• Three-wire only. Any neutral current invisible to the two meters corrupts the total — use three wattmeters.
• The power-factor formula assumes balance and sine waves. The power sum is general; tan φ = √3(W1−W2)/(W1+W2) is not.
• No per-phase information. You get the total and (under balance) the average angle — not which phase is misbehaving. For per-phase diagnosis, meter each phase.
• Instrument limits still apply. Frequency range, burden, and CT accuracy class bound the result exactly as in any metering setup.

Where to go from here

• CTs and PTs Explained — measuring high voltage and current safely, and the never-open-a-CT-secondary rule.
• Three-Phase Wiring Basics — star/delta, line vs phase quantities, the panel itself.
• Power-Factor Capacitor Sizing — what to do with the power factor you just measured.
• Reading Electrical Single-Line Diagrams — find the feeders you are about to meter.
• More electrical fundamentals in the Electrical branch hub.