Pi, L and T Networks
Impedance matching is one of the most important skills in high-frequency and power electronics. A signal source or power stage is designed to drive a particular load — often 50 Ω. The load you actually need to drive might present 200 Ω, or 20 Ω, or 600 + j300 Ω. Simply connecting them together means maximum power is not transferred — some reflects back, reducing efficiency and potentially stressing the source's output stage. The solution is an impedance matching network: a circuit of reactive elements that transforms the load impedance into the impedance the source expects.
Pi, L, and T networks are the three fundamental impedance matching topologies. They use only inductors and capacitors — no resistors — so they introduce no dissipative losses. They transform impedances while simultaneously filtering harmonics, making them doubly useful in power output stages. Understanding how these networks work, and how to calculate their component values, is the culmination of everything you have learned in this module about AC circuits, reactance, impedance, and resonance.
The three fundamental impedance matching networks. The L network uses two reactive elements. The Pi network uses three elements (named for its resemblance to the Greek letter π). The T network also uses three elements (named for its T-shape). All three can transform impedances without dissipating power.
View LargerWhy Impedance Matching Is Necessary
The maximum power transfer theorem states that maximum power is transferred from a source to a load when the load impedance equals the complex conjugate of the source impedance. For a purely resistive source with RS = 50 Ω, maximum power transfer occurs when Rload = 50 Ω.
When impedances are mismatched, some power is reflected from the load back toward the source. The ratio of reflected to forward power is determined by the reflection coefficient Γ:
Γ = (Zload − Zsource) / (Zload + Zsource)
VSWR = (1 + |Γ|) / (1 − |Γ|)
Power reflected = |Γ|² × forward power
Example: 200 Ω load on 50 Ω source: Γ = 150/250 = 0.6, VSWR = 4:1, reflected power = 36%
A 4:1 VSWR means 36% of the source's power never reaches the load — it reflects back into the source. Many solid-state sources automatically reduce power when they detect a large reflection to protect their transistors, so the actual output power drops further. A good impedance matching network between the source and the mismatched load eliminates the reflected power, presenting the source with 50 Ω regardless of what the load actually presents.
Impedance matching is not about making the source and load impedances physically equal — it is about transforming the load impedance so it appears equal to the source impedance at the source terminals. The load still has its natural impedance; the matching network creates the illusion of 50 Ω for the source.
The L Network
The L network is the simplest impedance matching network, consisting of just two reactive elements arranged in an L shape: one in series with the signal path and one in shunt (parallel) with either the source or the load. It can match any two resistive impedances with a specific Q factor determined by the impedance ratio.
For matching a higher impedance RH (the "high" side) to a lower impedance RL (the "low" side), the Q factor of the L network is:
Q = √(RH/RL − 1)
The component values at frequency f are:
Xshunt = RL × (RH/RL − 1) / Q = RH / Q
Wait — simplified: Xshunt = RH / Q and Xseries = RL × Q
The shunt element goes on the high-impedance side; the series element goes on the low-impedance side.
The L network has a fixed Q determined entirely by the impedance ratio. You cannot choose Q independently — it comes with the territory. This is a key limitation: for an impedance ratio of 4:1 (50 Ω to 200 Ω), Q = √(4 − 1) = √3 = 1.73. For a ratio of 100:1, Q = √99 ≈ 9.95. Higher impedance ratios automatically give higher Q (narrower bandwidth), which is useful for harmonic rejection but means the network must be retuned as you move across a band.
The L network has two configurations — shunt element on the input side (low-pass L) or shunt element on the output side (high-pass L). The low-pass L is more common in power output stages because it inherently attenuates harmonics.
Design a low-pass L network to match a 50 Ω source to a 200 Ω load at 7.1 MHz.
Step 1 — Q:
Q = √(RH/RL − 1) = √(200/50 − 1) = √(4 − 1) = √3 = 1.732
Step 2 — Shunt element (on high-impedance / load side):
Xshunt = RH / Q = 200 / 1.732 = 115.5 Ω
This is a capacitor in shunt on the load side (for low-pass L).
Cshunt = 1 / (2π × 7.1×10⁶ × 115.5) = 1 / (5,161,880) = 193.7 pF ≈ 200 pF
Step 3 — Series element (in series on source / low-impedance side):
Xseries = RL × Q = 50 × 1.732 = 86.6 Ω
This is an inductor in series.
Lseries = 86.6 / (2π × 7.1×10⁶) = 86.6 / 44,611,890 = 1.94 µH ≈ 2 µH
Result: 2 µH series inductor (source side) + 200 pF shunt capacitor (load side) matches 50 Ω to 200 Ω at 7.1 MHz with Q = 1.73. Bandwidth = f0/Q = 7.1/1.73 = 4.1 MHz — a very wide match, because a low-ratio L network has a low Q.
L Network Calculator
L Network Impedance Matching Calculator
Enter the source impedance (low side), load impedance (high side), and frequency. The calculator outputs the series and shunt reactances with component values.
The Pi Network
The Pi network is a popular impedance matching network for power output stages and RF generators. It consists of three reactive elements: a shunt capacitor on the input side, a series inductor in the middle, and a shunt capacitor on the output side. The circuit resembles the Greek letter π — hence the name. The same topology, with small values, is the standard matching network between a Wi-Fi or Bluetooth chip and its antenna.
The Pi network's key advantage over the L network is that the designer can freely choose the Q factor, independent of the impedance ratio. This is possible because the Pi network has an extra degree of freedom (three elements instead of two). Higher Q means narrower bandwidth and better harmonic suppression; lower Q means broader bandwidth and easier tuning. For power output stages, Q values of 10–15 are typical — high enough for useful harmonic rejection, low enough for easy tuning and tolerance to component variation.
Given: Rin (source/input impedance), Rout (load/output impedance), Q (designer's choice), frequency f
Input shunt capacitor: XC1 = Rin / Q → C1 = Q / (2π f Rin)
Series inductor: XL = Q × Rin × (1 + Rin/(Q² × Rout)) → L = XL / (2πf)
Actually, the standard design procedure uses Q referred to the high-impedance (source) side:
QS = √(RS/RL × (Q² + 1) − 1)
XCin = RS / Q, XCout = RL / QS, XL = Q × RS + RL × QS
Wait — for clarity, the simplified standard formulas are:
C1 (input shunt): XC1 = RS / Q
C2 (output shunt): XC2 = RL / √(RS/(Q² + 1) × 1/RL − 1)
L (series): XL = Q × RS − XC2 × RS/RL
These formulas are handled by the calculator below — entering Q, RS, RL, and frequency gives all three component values.
The Pi network can also step impedances up or down — it is not limited to stepping down like an L network. If Rout > Rin, the same topology still works with the appropriate component values from the calculator. Adjustable matching units often use a variant of the Pi network for this reason — it can handle a wide range of load impedances, both above and below 50 Ω.
A 13.56 MHz RF generator stage (the frequency used by RFID readers and industrial heating equipment) needs a 500 Ω load for full output. It must drive a 50 Ω output. Design a Pi network with Q = 5.
Step 1 — Input shunt capacitor C1 (source side, RS = 500 Ω):
XC1 = RS / Q = 500 / 5 = 100 Ω
C1 = 1 / (2π × 13.56×10⁶ × 100) = 117 pF
Step 2 — Output shunt capacitor C2 (load side, RL = 50 Ω):
XC2 = RL × √((RS/RL) / ((Q² + 1) − RS/RL)) = 50 × √(10 / (26 − 10)) = 50 × 0.791 = 39.5 Ω
C2 = 1 / (2π × 13.56×10⁶ × 39.5) = 297 pF
Step 3 — Series inductor L:
XL = (Q × RS + RS × RL / XC2) / (Q² + 1) = (2,500 + 25,000 / 39.5) / 26 = (2,500 + 632.5) / 26 = 120.5 Ω
L = 120.5 / (2π × 13.56×10⁶) = 1.41 µH
Check: With C1 = 117 pF, L = 1.41 µH and C2 = 297 pF, the network presents 500 + j0 Ω to the power stage when loaded with 50 Ω. The Pi network works best for impedance ratios up to about 10:1; for much larger ratios a transformer or a two-stage network is used.
Pi Network Calculator
Pi Network Impedance Matching Calculator
Enter source impedance, load impedance, desired Q factor, and frequency. The calculator returns all three component values. Works for both step-up and step-down matching.
The T Network
The T network is the dual of the Pi network — where the Pi has shunt-series-shunt topology, the T has series-shunt-series topology. It consists of two series elements (typically inductors or capacitors) with one shunt element in the middle. Its shape resembles the letter T.
The T network is commonly used in low-impedance-to-high-impedance matching and where the Pi network cannot achieve certain impedance combinations. It offers the same designer freedom of choosing Q, independently of the impedance ratio.
One important difference: the T network with two series inductors and a shunt capacitor forms a high-pass topology — it attenuates low frequencies but passes high frequencies more readily. This contrasts with the Pi network (two shunt capacitors and a series inductor) which is inherently low-pass. For power output applications where harmonic suppression is critical, the Pi (low-pass) is generally preferred. However, the high-pass T network may have advantages where a very low or very reactive load impedance must be matched.
L vs. Pi vs. T: Choosing the Right Network
| Property | L Network | Pi Network | T Network |
|---|---|---|---|
| Number of elements | 2 | 3 | 3 |
| Q control | Fixed by impedance ratio | Designer selects Q ≥ Qmin | Designer selects Q ≥ Qmin |
| Typical Q range | 1–5 (low-ratio matching) | 5–20 (power output stages) | 5–20 (wide-range matching) |
| Inherent filter type | Low-pass or high-pass | Low-pass | High-pass (with series L + shunt C + series L) |
| Harmonic suppression | Moderate (low-pass L) | Good | Poor (high-pass T amplifies harmonics) |
| Typical application | Simple impedance matching, chip-to-antenna matching | Power output stages, harmonic filtering, EMI filters | Wide-range adjustable matching |
| Impedance ratio limit | Any ratio works | Best for ratios up to ~25:1 | Best for ratios up to ~25:1 |
| Minimum Q | Qmin = √(RH/RL − 1) | Qmin = √(RH/RL − 1) | Qmin = √(RH/RL − 1) |
Matching Networks in Practice
L, Pi and T networks turn up wherever power or signals must move efficiently between parts with different impedances, and the same topologies double as filters.
Wireless Modules
Look closely at a Wi-Fi or Bluetooth module and you will usually find two or three tiny components between the chip and the antenna. They form an L or Pi matching network that transforms the antenna's impedance to the 50 Ω the chip expects. Manufacturers give suggested values, but the final values are tuned on each board design because nearby metal and plastic change the antenna's impedance.
Pi Filters for Power Lines
A capacitor–inductor–capacitor Pi network is also the classic power-line EMI filter: the shunt capacitors and series inductor block switching noise from travelling between a power supply and the rest of the circuit in either direction. Here the network is designed as a low pass filter rather than an impedance transformer, but the topology and calculations are the same.
Ultrasonic and Induction Drivers
Ultrasonic cleaners, ultrasonic welders and induction heaters drive loads whose impedance is far from the driver's ideal. An L network — often combined with the transducer's own capacitance or the work coil's inductance — transforms the load to the impedance the power stage needs and tunes out its reactance.
Standard Impedances
Matching is easiest when everything shares a standard impedance. Test equipment and high-frequency cable use 50 Ω, video uses 75 Ω, and many data links specify 90–120 Ω differential pairs. A matching network is the tool for connecting anything that does not already fit the standard.
Frequently Asked Questions
Why must the Q of a Pi or L network be at least as large as √(Rhi/Rlo − 1)?
This minimum Q is a mathematical constraint arising from the requirement that the shunt capacitor on the high-impedance side must have positive reactance (capacitive). If you attempt to design a network with Q below Qmin, the equations produce imaginary (impossible) component values. You cannot build a lossless two-reactive-element network with a Q lower than this minimum. A cascade of two networks can achieve any Q above zero for any impedance ratio, but a single-section network has this inescapable constraint.
Can an impedance matching network work for reactive loads, not just resistive ones?
Yes — this is the real-world case for most loads. A transducer, a coil or an antenna typically presents a complex impedance like 35 + j120 Ω (resistive plus inductive). The matching network must cancel the reactive part and transform the resistive part. In practice, this means the network's component values are adjusted from the purely resistive design to compensate for the load's reactance. With an adjustable network, tuning the L and C values for minimum reflected power finds the values that accomplish both reactive cancellation and resistive transformation simultaneously.
Why is the T network said to be high-pass? Doesn't it still match impedances at all frequencies?
A T network with two series inductors and a shunt capacitor has a high-pass frequency response — signals above the operating frequency are passed with relatively low insertion loss, while signals below it are attenuated. The impedance matching function works at the design frequency, but the harmonic behavior is unfavorable: harmonics (which are above the fundamental) see a relatively low-loss path through the network to the load. This is why Pi networks (low-pass) are strongly preferred in power output stages where harmonic emissions must be kept low.
Test Your Knowledge
Answer the questions below to check your understanding of this lesson. Every answer can be found in the lesson above.