Millman's theorem solves a very common case in one go: several branches connected in parallel between the same two nodes, each made up of a source with its resistance (or impedance) in series. Instead of setting up meshes or a system of equations, the voltage between the two nodes is obtained with a single formula.
Essentially it is nodal analysis applied to a circuit with only two nodes, or, equivalently, converting each branch into its Norton equivalent and adding them up: the Norton currents add and the conductances add.
Branches in parallel between nodes A and B. A branch may have no source (a load).
Millman's theorem in direct current
2. Statement
The voltage between nodes A and B is the sum of each source multiplied by the conductance of its branch, divided by the sum of all the conductances:
where \(G_k = 1/R_k\) is the conductance of each branch. Once \(V_{AB}\) is known, the current in each branch (taken from B to A through the source) is:
A resistance in series with a current source has no effect: it appears in neither the numerator nor the denominator.
4. Steps to solve
Check that all branches are connected between the same two nodes A and B.
Take B as the reference and give each source a sign: positive if its + terminal faces A.
Calculate the conductance of each branch, \(G_k = 1/R_k\) (adding up any series resistances in the same branch first).
Apply Millman's formula to obtain \(V_{AB}\).
Calculate the current in each branch with \(I_k = (E_k - V_{AB})/R_k\) and interpret the sign: if it is negative, the current enters through the + terminal (the source absorbs energy, for example a battery being charged).
Check that the currents add up to zero.
Example 1: two batteries in parallel supplying a load
A battery of \(E_1 = 12\) V with internal resistance \(R_1 = 2\,\Omega\) and another of \(E_2 = 6\) V with \(R_2 = 3\,\Omega\) are connected in parallel across a load \(R_3 = 6\,\Omega\).
Check: \(2 = 0.67 + 1.33\ \checkmark\). Battery 2 has a negative current: it is being charged by battery 1. That is why batteries of different voltages should not be connected in parallel.
Millman's theorem in alternating current
5. Statement with phasors
In sinusoidal AC the theorem is identical, replacing voltages with phasors and conductances with complex admittances \(\overline{Y}_k = 1/\overline{Z}_k\):
Always work with complex numbers: magnitudes cannot be added directly, because the phasors have different angles.
6. Steps to solve in AC
Express all sources as phasors (magnitude and angle) with the same reference.
Calculate the impedance of each branch \(\overline{Z}_k = R_k + jX_k\) and its admittance \(\overline{Y}_k\).
Calculate the numerator \(\sum \overline{E}_k \overline{Y}_k\) (in rectangular form so the terms can be added).
Calculate the denominator \(\sum \overline{Y}_k\) and divide (easier in polar form).
Obtain the branch currents and, if required, powers: \(\overline{S}_k = \overline{V}\,\overline{I}_k^{\,*}\).
Example 2: two generators in parallel supplying a load
Two generators of \(\overline{E}_1 = 230\angle 0^\circ\) V and \(\overline{E}_2 = 220\angle 0^\circ\) V, both with internal impedance \(\overline{Z} = j2\,\Omega\), supply a \(10\,\Omega\) resistive load.
Both generators supply the same active power (same real part of the current), but the one with the higher EMF delivers reactive power and the other absorbs it: the voltage difference only causes reactive power to circulate between them.
7. Key application: unbalanced star load without neutral
If an unbalanced star-connected load is connected without a neutral, its star point O' no longer coincides with the supply neutral O. The three phases are three branches in parallel between O and O', so Millman directly gives the neutral shift voltage:
With a neutral of negligible impedance (\(\overline{Y}_N \to \infty\)) there is no shift; without a neutral, \(\overline{Y}_N = 0\). The actual voltages across each load are \(\overline{V}_{AO'} = \overline{V}_A - \overline{V}_{O'O}\), and likewise for the others. If the load is balanced, the numerator is zero and there is no shift.
Example 3: unbalanced star without neutral
400 V supply (230 V phase): \(\overline{V}_A = 230\angle 0^\circ\), \(\overline{V}_B = 230\angle -120^\circ\), \(\overline{V}_C = 230\angle 120^\circ\) V. Resistive loads \(R_A = R_B = 10\,\Omega\), \(R_C = 20\,\Omega\), no neutral.
The least loaded phase (highest impedance) suffers a 20% overvoltage, and the other two fall below the rated voltage. Currents: \(I_A = I_B = 21.1\) A, \(I_C = 13.8\) A, and their phasor sum is zero.
Why the neutral must never be broken: in an installation with single-phase loads spread across the phases, a broken neutral causes exactly this situation. Appliances on the least loaded phase may receive voltages close to 400 V and be damaged. That is why the neutral is not fused and is never disconnected without disconnecting the phases at the same time.
Batteries and power supplies in parallel: how current is shared between them and detection of reverse currents (one battery charging another).
Generators and transformers in parallel: load sharing and reactive power circulating due to voltage or impedance differences.
Unbalanced three-phase systems: neutral shift, broken neutral and actual voltages on star loads without neutral.
Electronics: resistive adders and signal mixers (the output is a weighted average of the inputs), weighted-resistor digital-to-analogue converters.
Photovoltaic installations: estimating the common voltage of strings or panels connected in parallel with slightly different voltages.
10. Precautions and common mistakes
Only two nodes. All branches must run between exactly the same two nodes. If there are elements connecting branches to each other, simplify the circuit first (or use nodal analysis).
Source signs. Each source takes its sign depending on whether its + terminal faces A or B. A wrong sign is the most common mistake.
Branches without a source count in the denominator. Forgetting the load in the sum of conductances gives a wrong voltage.
Ideal voltage source (R = 0): its conductance is infinite and the formula cannot be applied directly; in that case \(V_{AB}\) is simply that voltage. Two different ideal sources in parallel are an impossible circuit.
Do not add magnitudes in AC. Numerator and denominator are added in rectangular form; only convert to polar at the end.
Same frequency. In AC all sources must have the same frequency; with different frequencies, apply superposition and solve each frequency separately.
Linear circuits. Like all these theorems, it is only valid for linear elements.