When you are only interested in what happens in one part of a circuit (a load, a connection point, a device that is going to change), there is no need to solve the whole network every time. Thévenin's and Norton's theorems let you replace the whole circuit, as seen from two terminals A–B, with a very simple equivalent that behaves exactly the same towards the outside.
Thévenin: any linear network seen from two terminals is equivalent to a voltage source \(V_{Th}\) in series with an impedance \(Z_{Th}\).
Norton: the same network is equivalent to a current source \(I_N\) in parallel with an impedance \(Z_N\).
Both equivalents are two ways of describing the same thing, and converting from one to the other is immediate.
The two equivalents of the same network seen from terminals A–B.
2. The three parameters of the equivalent
The whole equivalent is defined by three quantities, and knowing any two of them is enough:
Identify terminals A–B and remove the load (or the part of the circuit you want to study). Mark the polarity: \(V_{Th} = V_A - V_B\).
Calculate \(V_{Th}\): solve the circuit with A–B open (voltage divider, mesh, nodal analysis, superposition…) and find the voltage between A and B.
Calculate \(Z_{Th}\), depending on the type of sources:
Independent sources only: turn them off (voltage source → short circuit; current source → open circuit) and calculate the equivalent impedance seen from A–B.
With dependent sources:do not turn them off. Calculate \(Z_{Th} = V_{Th}/I_N\), or turn off only the independent sources and connect a test source \(V_p\) across A–B: \(Z_{Th} = V_p / I_p\).
Draw the equivalent: \(V_{Th}\) in series with \(Z_{Th}\).
Reconnect the load and calculate what is required:
Identify A–B and remove the load. Set the direction of \(I_N\) (from A to B through the short circuit).
Short-circuit A–B and calculate the current flowing through that short circuit: that is \(I_N\).
Calculate \(Z_N\): exactly as for \(Z_{Th}\).
Draw the equivalent: \(I_N\) in parallel with \(Z_N\).
Reconnect the load (current divider):
\[ I_L = I_N\,\frac{Z_N}{Z_N + Z_L} \]
Practical tip: it is often easier to calculate two of the three quantities (\(V_{Th}\), \(I_N\), \(Z_{Th}\)) and derive the third. Pick the two that take the least work.
Example 1: DC circuit
A 12 V source feeds a divider made of \(R_1 = 4\,\Omega\) (in series) and \(R_2 = 6\,\Omega\) (in parallel with the output). From the midpoint, a resistor \(R_3 = 2.6\,\Omega\) goes to terminal A. We want the current in a load \(R_L = 7\,\Omega\) connected between A and B.
Thévenin voltage (with no load, no current flows through \(R_3\)):
A \(230\angle 0^\circ\) V source feeds \(Z_1 = 10\,\Omega\) in series, with \(Z_2 = j10\,\Omega\) between A and B. Which load absorbs the maximum power, and how much is it?
This table shows pairs of terms, concepts or elements which, when swapped, let you derive from one
statement another one that is also true. This is called
duality and it is useful for proving theorems, especially in electrical engineering.
6. Applications
Variable loads: if many load values need to be tried, the equivalent is calculated once and each case is solved in one line.
Maximum power transfer: in DC, \(R_L = R_{Th}\); in AC, \(Z_L = Z_{Th}^*\). It is the basis of impedance matching in audio, antennas and communications.
Short-circuit calculations in power networks: the network seen from the fault point is reduced to its Thévenin equivalent, and the short-circuit current is \(I_{cc} = V_{Th}/Z_{Th}\). Breakers and protection devices are sized from it.
Modelling real sources: a battery, a generator or a transformer secondary is represented as \(V_{Th}\) plus its internal impedance (for transformers, the short-circuit impedance). This explains the voltage drop as the load increases.
Electronics: amplifier input and output impedances, transistor biasing (the base divider is replaced by its Thévenin equivalent), Wheatstone bridge sensors.
Laboratory measurement: \(V_{Th}\) is measured with a voltmeter on open circuit and \(I_N\) with an ammeter on short circuit (if safe); it also lets you estimate the "loading" error introduced by a measuring instrument.
7. Experimental determination
METHOD
WHAT IS MEASURED
RESULT
Open circuit + short circuit
\(V_0\) and \(I_{cc}\)
\(V_{Th} = V_0\), \(Z_{Th} = V_0/I_{cc}\)
Open circuit + known load
\(V_0\) and \(V_L\) with \(R_L\)
\(R_{Th} = R_L\left(\dfrac{V_0}{V_L} - 1\right)\)
Half voltage
\(R_L\) such that \(V_L = V_0/2\)
\(R_{Th} = R_L\)
The known-load methods avoid the short circuit, which on a real source can cause dangerous currents.
8. Precautions and common mistakes
Linear networks only. The part being replaced must be linear (resistors, inductors, capacitors, sources, linear dependent sources). The load itself can be non-linear (a diode, a lamp); it is then solved with the equivalent plus the load characteristic.
Do not turn off dependent sources. When calculating \(Z_{Th}\), only independent sources are turned off. Also, the variable controlling the dependent source must be inside the network being replaced.
Turn sources off correctly: voltage → short circuit; current → open circuit. Mixing them up is the most common mistake.
Watch polarities and directions. \(V_{Th}\) is \(V_A - V_B\) and \(I_N\) flows from A to B through the short circuit. With a consistent convention, \(Z_{Th} = V_{Th}/I_N\) must have a positive real part in passive networks.
Equivalent only towards the outside. The equivalent reproduces voltages and currents at A–B, but not the power or the internal losses of the original network. On open circuit, the Thévenin equivalent dissipates nothing and the Norton one does, while the real network may do something else entirely.
In AC, valid at a single frequency. \(Z_{Th}\) depends on frequency; if the frequency changes (or there are harmonics), the equivalent must be recalculated for each one.
Limiting cases: if \(Z_{Th} = 0\) (ideal voltage source) there is no Norton equivalent; if \(Z_N \to \infty\) (ideal current source) there is no Thévenin equivalent.
Maximum power ≠ maximum efficiency. With \(Z_L = Z_{Th}^*\) the efficiency is only 50%. In power networks the opposite is sought: \(Z_{Th}\) much smaller than the load.
Three-phase systems: with a balanced load, work with the single-phase equivalent (per phase, in equivalent star).