Maximum Power Transfer With a Complex Source: Derivation
A load takes the most power from a source Z_S = R_S + jX_S when X_L = -X_S and R_L = R_S, which means Z_L = Z_S*. The available power is |V_S|^2/(8 R_S), so a 2 V peak source behind 50 ohm can deliver at most 10 mW.
Engineer’s note: To get maximum power from a complex source, the load has to be its complex conjugate: same resistance, opposite reactance, so an inductive part is cancelled by a capacitive one and vice versa. A lot of everyday matching work comes down to grinding that leftover reactance to zero.
For practitioners
Every matching problem starts from the same model: a linear source with an open-circuit voltage and a fixed internal impedance, driving a load you are free to choose. If the source impedance is purely resistive, the answer is the familiar R_L = R_S. When the source has reactance, the answer becomes the complex conjugate. This page derives that result one variable at a time, so you can see what each half of the condition does.
The setup: power into a complex load
Take a sinusoidal source phasor V_S (peak amplitude) with internal impedance Z_S = R_S + jX_S. It drives a load Z_L = R_L + jX_L. Because the circuit is a single series loop, the current is
The load reactance stores energy and gives it back, so it absorbs no average power. Only R_L absorbs power. With peak phasors, the average power is (1/2)|I|^2 R_L:
The source is fixed, so V_S, R_S and X_S are constants. The two free variables are R_L and X_L. Because they appear in different places in the expression, you can optimize them one at a time.
Step 1: pick X_L to cancel the source reactance
X_L appears in only one place, the term (X_S + X_L)^2 in the denominator. That term is a square, so it can never be negative. Whatever value R_L has, P_L is largest when that term is zero:
This is the reactance half of the condition. An inductive source (X_S > 0) needs a capacitive load, and a capacitive source needs an inductive load. When the total loop reactance is zero, the loop is resonant at the operating frequency. The current is then limited only by the total resistance R_S + R_L, and no voltage is wasted driving current through a net reactance.
Note that this step does not depend on R_L. If R_L is fixed by something outside your control, cancelling the reactance is still the best move.
Step 2: pick R_L with the reactances cancelled
With X_L = -X_S, the power becomes
Now maximize f(R_L) = R_L / (R_S + R_L)^2. Use the quotient rule:
The derivative is positive for R_L < R_S, zero at R_L = R_S, and negative for R_L > R_S. So the single stationary point is a maximum:
Put both steps together and you get the conjugate match:
The reason for the optimum is a trade-off. A small R_L draws a large current but turns little of it into load power, because P = (1/2)|I|^2 R_L. A large R_L takes a large share of the voltage but chokes the current. The product peaks where the two resistances are equal.
Available power and a worked example
Substitute R_L = R_S into the reduced expression. The denominator becomes 4R_S^2, which gives the available power of the source:
No passive load can draw more than this from the source. If you work with RMS phasors, the factor 1/2 is gone and the result is |V_rms|^2 / (4 R_S). Mixing up the two conventions is the most common way to get an answer that is off by a factor of two.
Worked example: V_S = 2 V peak and R_S = 50 Ω. X_S can take any value, because the load cancels it.
- Choose the load: X_L = -X_S and R_L = 50 Ω, so Z_L = Z_S*.
- Loop resistance: R_S + R_L = 100 Ω. Loop reactance: zero.
- Current: |I| = 2 V / 100 Ω = 20 mA peak.
- Load power: (1/2)(0.02 A)^2 (50 Ω) = 0.01 W.
- Check with the formula: |V_S|^2 / (8 R_S) = 4/400 W = 10 mW.
The same 20 mA peak flows through R_S, so the source's internal resistance also dissipates 10 mW. The source puts out 20 mW in total.
How fast power falls off with resistance mismatch
Divide the reduced P_L by P_avail. With the reactances still cancelled and r = R_L/R_S, you get
The curve is symmetric on a log axis. Swapping r for 1/r gives the same value, so a load at half the source resistance loses as much as a load at twice the source resistance.
- r = 1: 1.0, the conjugate match.
- r = 2 or r = 0.5: 8/9 = 0.889. In the example, that is about 8.89 mW out of 10 mW.
- r = 4 or r = 0.25: 16/25 = 0.64, or 6.4 mW.
The peak is flat. A factor-of-two error in resistance costs only about 11 percent of the power. That is why a match that is "close" is often good enough, and why the expensive part of a design tends to be the reactance cancellation across a band, not the exact resistance. The same loss, expressed through the reflection coefficient, is the mismatch loss covered in VSWR Explained: Reflection Coefficient and Mismatch Loss.
What the theorem does and does not say
At conjugate match, the same current flows through R_S and R_L, and the two resistances are equal. So they dissipate equal power, and the efficiency in this source model is 50 percent. With reactances cancelled, the general efficiency is R_L / (R_S + R_L). Raising R_L above R_S improves efficiency but delivers less power.
There are two more limits. First, X_S usually changes with frequency, so -X_S can only be matched exactly at one frequency. A real matching network approximates the conjugate over a band, as discussed in RF Matching Networks: L, Pi and T. Second, the derivation assumes a linear source with a fixed Thevenin impedance. When a two-port sits between source and load, both ports have to be matched together. That case is covered in Simultaneous Conjugate Match for Maximum Amplifier Gain.
In plain words · for students and beginners
Maximum power transfer in plain words
An everyday analogy, and where it breaks down
Think of a bicycle with gears. In a very low gear your legs spin fast but the bike barely moves. In a very high gear you push hard but the pedals hardly turn. The most useful power comes from a gear in between. Resistance works the same way. A load resistance that is too small lets lots of current flow but takes little power from it. A load resistance that is too large chokes the current. The best value equals the source's own resistance.
Reactance is a different problem. Reactance is the part of an impedance that comes from inductors or capacitors. It stores energy and hands it back each cycle without using any. If the source has some, the load should have an equal amount of the opposite kind so the two cancel. It is like pushing a swing in time with its motion: you stop fighting the swing.
The analogy breaks down in one place. On a bicycle, the best gear depends on the rider. In the circuit, there is an exact rule, and at that rule half the power turns into heat inside the source itself.
The worked example in simple steps
Our source makes 2 V peak behind 50 Ω of resistance plus some reactance, +jX_S.
- Cancel the reactance. Give the load -jX_S, so the two reactances add to zero.
- Match the resistance. Give the load 50 Ω, the same as the source.
- Find the current. 2 V divided by 100 Ω total is 20 mA peak.
- Find the load power. Half of current squared times 50 Ω gives 10 mW.
- Find the heat in the source. The same 20 mA flows through the source's own 50 Ω, so another 10 mW becomes heat there.
So 20 mW leaves the source, 10 mW reaches the load, and the efficiency is 50 percent. "Conjugate" just means the same resistance with the sign of the reactance flipped.
Mistakes beginners make
- Setting Z_L equal to Z_S instead of its conjugate. That doubles the reactance and cuts the current.
- Thinking a conjugate match is efficient. It gives the most power to the load, but half of the total is lost in the source.
- Mixing peak and RMS voltages. With RMS voltage the formula is |V|^2/(4 R_S), not /(8 R_S).
- Worrying about a small resistance error. A load at twice or half the right resistance still gets about 89 percent of the best power.
Remember
- Cancel the reactance first: X_L = -X_S.
- Then make the resistances equal: R_L = R_S.
- The most power you can get is |V_S|^2/(8 R_S), which is 10 mW for 2 V peak behind 50 Ω, at 50 percent efficiency.
Related pages on RF View
Frequently asked questions
Why is the maximum power transfer condition Z_L = Z_S* and not Z_L = Z_S?
The load reactance has to cancel the source reactance so that the total loop reactance is zero. That requires X_L = -X_S, which is the complex conjugate. Setting Z_L = Z_S would double the loop reactance and reduce the current.
What is the available power of a source?
It is the largest average power the source can deliver to any passive load. You get it with a conjugate-matched load, and it equals |V_S|^2/(8 R_S) for a peak-amplitude phasor V_S.
Is a conjugate match 100 percent efficient?
No. At conjugate match the source's internal resistance dissipates the same power as the load, so the efficiency in this model is 50 percent. Conjugate matching maximizes power delivered to the load, not efficiency.
How much power is lost if the load resistance is double the source resistance?
With the reactances cancelled, the delivered fraction is 4 R_S R_L/(R_S + R_L)^2. For R_L = 2 R_S that gives 8/9, about 0.889 of the available power. R_L = R_S/2 gives the same result.