How to calculate an L-network for impedance matching
Compute Q = √(Rhigh ÷ Rlow − 1), then put a series reactance of Q × Rlow next to the lower resistance and a shunt reactance of Rhigh ÷ Q across the higher one. Matching 50 Ω to 10 Ω gives Q = 2, a 20 Ω series reactance and a 25 Ω shunt reactance: at 100 MHz that is 31.8 nH and 63.7 pF. A reactive load changes these values, and the calculator includes it.
Impedance matching formula for an L-network
An L-section has two degrees of freedom, enough to set both the real and imaginary parts of the input impedance. Which arrangement works, and so which impedance matching equation applies, depends on where the load sits relative to the source resistance.
When RL is above Rs (or the load’s magnitude is large enough), a shunt element goes across the load and a series element goes toward the source, as in Pozar’s equations 5.3:
B = [XL ± √(RL / Rs) × √(RL² + XL² − Rs × RL)] ÷ (RL² + XL²) X = 1/B + XL × Rs / RL − Rs / (B × RL)
- Rs
- source resistance in Ω
- RL, XL
- load resistance and reactance in Ω
- B
- shunt susceptance in S (positive: capacitor; negative: inductor)
- X
- series reactance in Ω (positive: inductor; negative: capacitor)
When RL is below Rs, a series element goes next to the load and a shunt element goes across the source (equations 5.6):
X = ±√(RL × (Rs − RL)) − XL B = ±√((Rs − RL) / RL) ÷ Rs
Each reactance becomes a part at the design frequency:
Series: L = X / ω or C = −1 / (ω X) Shunt: C = B / ω or L = −1 / (ω B)
For a resistive load, the Q of the network is set by the ratio alone. The loaded Q is half of it, a relation IIIT-Delhi’s RF circuit notes state for any L-type matching network, and it follows from the source and load each contributing equal resistance at the junction when matched:
Qn = √(Rhigh / Rlow − 1) QL = Qn / 2 Bandwidth ≈ f0 / QL
L-matching network design examples
Example 1: 200 − j100 Ω to a 100 Ω line at 500 MHz (Pozar Example 5.1)
RL is above the 100 Ω reference, so the shunt element goes at the load. The first solution has B = 2.9 mS and X = 122 Ω: Series L 39 nH, shunt C 0.923 pF. The second has B = -6.9 mS and X = -122 Ω: Series C 2.6 pF, shunt L 46.1 nH. Both give exactly 100 Ω at 500 MHz, and they differ in bandwidth away from it.
Example 2: 10 Ω to 50 Ω at 100 MHz
The load is below the source, so the series element goes next to the load. Q = √(50/10 − 1) = 2, the series reactance is 20 Ω and the shunt reactance is 25 Ω. The low-pass form is Series L 31.8 nH, shunt C 63.7 pF; the high-pass form is Series C 79.6 pF, shunt L 39.8 nH. The low-pass form holds 10 dB of return loss from 79.2 MHz to 117 MHz.
Example 3: 20 − j30 Ω to 50 Ω at 915 MHz, four solutions
RL is below Rs, and |Z|² = 1300 is above Rs × RL = 1000, so both arrangements are available:
- High-pass: Series C 6.35 pF, shunt L 5.52 nH (series at source, shunt across load)
- Series L, shunt L: Series L 0.958 nH, shunt L 7.1 nH (shunt across source, series to load)
- Series L, shunt L: Series L 4.76 nH, shunt L 11.9 nH (series at source, shunt across load)
- Low-pass: Series L 9.48 nH, shunt C 4.26 pF (shunt across source, series to load)
Q and reactances for common 50 Ω transformations
For a resistive load driven from 50 Ω. The element next to the lower resistance is the series part, and the larger the resistance step, the higher the Q and the narrower the match.
| RL | Series part next to | Node Q | Loaded Q | Series |X| | Shunt |X| |
|---|
| 5 Ω | Load | 3 | 1.5 | 15 Ω | 16.7 Ω |
| 10 Ω | Load | 2 | 1 | 20 Ω | 25 Ω |
| 12.5 Ω | Load | 1.73 | 0.866 | 21.7 Ω | 28.9 Ω |
| 25 Ω | Load | 1 | 0.5 | 25 Ω | 50 Ω |
| 75 Ω | Source | 0.707 | 0.354 | 35.4 Ω | 106 Ω |
| 100 Ω | Source | 1 | 0.5 | 50 Ω | 100 Ω |
| 200 Ω | Source | 1.73 | 0.866 | 86.6 Ω | 115 Ω |
| 300 Ω | Source | 2.24 | 1.12 | 112 Ω | 134 Ω |
| 500 Ω | Source | 3 | 1.5 | 150 Ω | 167 Ω |
| 1 kΩ | Source | 4.36 | 2.18 | 218 Ω | 229 Ω |
Low-pass L-match parts from 50 Ω
Series inductor and shunt capacitor for a resistive load, at common frequencies. Round to standard values and retune for parasitics.
| Load | 13.56 MHz | 100 MHz | 433 MHz | 915 MHz | 2.45 GHz |
|---|
| 10 Ω | 235 nH, 469 pF | 31.8 nH, 63.7 pF | 7.35 nH, 14.7 pF | 3.48 nH, 6.96 pF | 1.3 nH, 2.6 pF |
| 25 Ω | 293 nH, 235 pF | 39.8 nH, 31.8 pF | 9.19 nH, 7.35 pF | 4.35 nH, 3.48 pF | 1.62 nH, 1.3 pF |
| 100 Ω | 587 nH, 117 pF | 79.6 nH, 15.9 pF | 18.4 nH, 3.68 pF | 8.7 nH, 1.74 pF | 3.25 nH, 0.65 pF |
| 200 Ω | 1.02 µH, 102 pF | 138 nH, 13.8 pF | 31.8 nH, 3.18 pF | 15.1 nH, 1.51 pF | 5.63 nH, 0.563 pF |
What goes wrong in real matching networks
Parts are not ideal. Every inductor has a self-resonant frequency and a Q of its own, so a high-node-Q match loses more of the power in its windings: keep the node Q low unless you need the selectivity. The shunt part connects through a pad and a via, and at high frequency that adds several tenths of a nanohenry in series with it. Part tolerance matters too: on a narrow match, a 5 % part can move the center frequency out of the band. Measure the load in place if you can, since a load quoted at a different frequency or package may differ substantially, and allow a footprint for both a series and a shunt part at each end so that the network can be tuned on the board.
Glossary
- L-section
- A matching network of one series and one shunt reactance, forming an L shape.
- Reactance (X)
- The imaginary part of an impedance, in ohms: positive for an inductor, negative for a capacitor.
- Susceptance (B)
- The imaginary part of an admittance, in siemens: positive for a shunt capacitor, negative for a shunt inductor.
- Node Q
- The ratio of reactance to resistance at the junction of the matching parts.
- Loaded Q
- The Q of the matched circuit including the source, half the node Q for an L-section.
- Return loss
- How far the reflected power is below the incident power, in dB. 10 dB means 10 % of the power is reflected.
- Self-resonant frequency
- The frequency where a part’s parasitics cancel its intended reactance and it stops behaving as an L or C.
Frequently asked questions
How many L-network solutions are there for one load?
Usually two or four. With the load resistance above the source, only the shunt-across-load arrangement can work, giving two sets of values; below the source, the series-to-load arrangement gives two and the shunt-across-load arrangement adds two more when |Z|² ≥ Rs × RL. A solution can lose a part or coincide with another for special loads.
What is the Q of an L-match?
Q = √(Rhigh / Rlow − 1) when the source and load are resistive. Matching 50 Ω to 10 Ω gives Q = 2. Q is fixed by the ratio, so an L-section cannot trade match for bandwidth: a pi or T network, or two L-sections in cascade, adds that freedom.
What is the difference between node Q and loaded Q?
Node Q is the reactance-to-resistance ratio at the junction between the two parts, and loaded Q is half of it when the network is matched, because source and load contribute equal resistance at that point. The half-power bandwidth is about f0 / QL: within about 15 % for a loaded Q above 2.5 and closer as Q rises, while a lower-Q match is wider and may not fall 3 dB on one side at all. A strongly reactive load stores energy the node Q does not count, so for those read the computed band instead.
Should I choose the low-pass or high-pass L-match?
Low-pass (series L, shunt C) attenuates harmonics, which suits amplifier outputs. High-pass (series C, shunt L) blocks DC in the signal path and gives a DC path to ground through the inductor, useful for feeding bias or for ESD on an antenna input. The other two combinations are also valid when the load reactance permits them.
How does the L-network calculator handle a reactive load?
The load reactance is included in the design, so the values you get are the parts to install. They already absorb or resonate that reactance at the design frequency, which is why a load can need a much smaller or larger part than its resistance alone suggests.
Can I match a load given as a resistor in parallel with a capacitor?
Yes. Choose the R ∥ C tab and enter the parallel resistance and capacitance from the datasheet. The tool converts it to R + jX at your frequency, and sweeps the parallel form for the bandwidth.
Can I use this as an antenna impedance matching calculator?
Yes, as long as the source side is resistive. Enter the antenna’s R + jX at your frequency as the load and the resistance of the line or radio port as Rs. Measure the antenna in place if you can, since a value quoted at a different frequency or on a different board may differ substantially.
Does this matching network calculator handle pi networks?
No. It solves L-sections, with one series and one shunt part. With a resistive source and load, the Q of an L-section is fixed by the resistance ratio, so when you also need to set the bandwidth, a pi or T network, or two L-sections in cascade, adds that freedom.
How accurate are the L and C values at high frequency?
They are exact for ideal parts at the design frequency. Real inductors and capacitors have self-resonance, pad capacitance and via inductance, so treat the values as a starting point above a few hundred MHz and tune on a board or in an S-parameter simulation with vendor models.
More free PCB tools
See all free electronics tools →
- Pi attenuator calculator (and T pad)Resistor values for Pi and T attenuator pads for any attenuation and impedance, with nearest standard values and their real loss.
- EIRP and effective radiated power calculatorEIRP and ERP from transmitter power, antenna gain and cable loss, in dBm and watts, and the conducted power a limit allows.
- Coaxial cable impedance calculatorCharacteristic impedance, velocity factor, capacitance and inductance per meter of a coaxial line from its dimensions and dielectric.
- RF link budget calculatorReceived power and fade margin from transmit power, antenna gains, cable losses, path loss and receiver sensitivity, plus maximum range.
- Antenna length calculatorWavelength, half-wave dipole and quarter-wave monopole lengths for any frequency, with an end-effect factor.
- Free-space path loss calculatorFree-space path loss between two antennas from distance and frequency, in dB, with the distance for a given loss.
Sources
- D. M. Pozar, Microwave Engineering, 4th ed., Wiley, section 5.1 (book reference: the publisher’s page does not show the equations; the Whites notes below reproduce eqs. 5.3 and 5.6, and the JBNU notes Example 5.1)
- Jeonbuk National University, Chapter 5 lecture notes: Impedance Matching and Tuning (L-section equations and Example 5.1)
- K. Whites, EE 481/581 Lecture 7: Transmission Line Matching Using Lumped L Networks
- University of Utah ECE 6130 notes: Pi-matching networks built from L-sections, after Bowick, RF Circuits
- IIIT-Delhi ECE321/521 RF Circuit Design, Lecture 14 (2014), slide 10: nodal Q and QL = Qn / 2 for any L-type matching network
Design the whole board, not one number
ProtoFlow is free AI PCB design software: describe a circuit and get a schematic with real parts, PCB layout with an autorouter, SPICE simulation and Gerber files. Native KiCad, Altium and EAGLE import and export. Free on macOS, Windows and Linux.