Pi-L Networks
When an RF power amplifier delivers power to a 50 Ω load — an antenna connector, a filter, an induction-heating coil or a test load — the amplifier is rarely designed to work directly into 50 Ω. Transistor power amplifiers in class AB, C or E operation present an output impedance of perhaps 5–25 Ω when running at full power; a wireless chip's output may want to see 20 or 100 Ω; valve (tube) amplifiers present 1,500–5,000 Ω. The network between the amplifier and the 50 Ω load — the output matching network — must transform the impedance to 50 Ω, provide some harmonic suppression, and ideally do both while introducing minimal insertion loss.
The Pi network and L network are the workhorses of output matching. They have been used since the early days of electronics, they are simple to design and build, they can be made adjustable to accommodate load variations, and they combine impedance matching with built-in harmonic filtering. The Pi-L network combines both topologies in series for even better harmonic suppression. Anyone who designs, builds or repairs RF circuits will meet these networks — on wireless module boards they are the two or three tiny components between the chip and the antenna connector — and calculating their component values is an essential skill.
Why Impedance Matching Matters
The maximum power transfer theorem states that a source delivers maximum power to a load when the load impedance equals the complex conjugate of the source impedance. For a purely resistive source with source resistance RS, maximum power is delivered when the load resistance RL = RS.
If the impedances are mismatched, power is reflected rather than transferred. For a source impedance of RS = 1500 Ω trying to drive RL = 50 Ω directly, the mismatch ratio is 30:1. The power delivered to the load compared to maximum possible is:
For RS = 1500 Ω, RL = 50 Ω: efficiency = 4 × 1500 × 50 / (1550)² = 300,000 / 2,402,500 = 12.5%
Driving a 50 Ω load directly from a 1500 Ω source delivers only 12.5% of the available power — an 87.5% loss, or about 9 dB. The matching network recovers this lost power by presenting the load impedance (50 Ω) as the correct impedance to the source, and presenting the source impedance (1500 Ω) as the correct impedance to the load. The network is passive — it contains no power amplification — but it transforms the impedances so both source and load "see" their optimal impedance, maximizing power transfer.
The L Network
The L network is the simplest practical matching network, consisting of exactly two reactive elements: one in series and one in shunt (parallel). The two possible configurations are a low-pass L (shunt capacitor on the high-impedance side, series inductor on the low-impedance side) and a high-pass L (shunt inductor on the high-impedance side, series capacitor on the low-impedance side). The most common for amplifier output matching is the low-pass L because it provides inherent high-frequency attenuation.
L network (2 components), Pi network (3 components), and Pi-L network (4 components). All three transform impedance from source to load. The Pi and Pi-L networks have an additional degree of freedom (adjustable Q) that allows harmonic suppression to be optimized independently of the impedance ratio.
View LargerL Network Design
For a low-pass L network matching Rhigh to Rlow (where Rhigh > Rlow):
Series inductor reactance: XL = Q × Rlow
Shunt capacitor reactance: XC = Rhigh / Q
Component values:
L = XL / (2π × f) C = 1 / (XC × 2π × f)
Worked example: L network matching a 50 Ω source to a 200 Ω load at 14.2 MHz.
Rhigh = 200 Ω, Rlow = 50 Ω
Q = √(200/50 − 1) = √(4 − 1) = √3 = 1.73
XL = Q × Rlow = 1.73 × 50 = 86.5 Ω → L = 86.5 / (2π × 14.2×10⁶) = 970 nH ≈ 0.97 µH
XC = Rhigh / Q = 200 / 1.73 = 115.6 Ω → C = 1 / (115.6 × 2π × 14.2×10⁶) = 97 pF
The series inductor (0.97 µH) goes between the 50 Ω source and the shunt capacitor. The shunt capacitor (97 pF) connects from the junction to ground on the 200 Ω side. Note that the Q of 1.73 is fixed by the impedance ratio — you cannot change the Q without changing the impedance ratio.
The L network's main limitation is that its Q is completely determined by the impedance ratio. For matching 1500 Ω to 50 Ω, Q = √(1500/50 − 1) = √29 = 5.39. This relatively low Q means moderate harmonic suppression. Where little harmonic suppression is required — or where the impedance ratio is small, as when matching a wireless chip to 50 Ω — the L network is adequate. When higher Q (and thus better harmonic suppression) is needed without changing the impedance ratio, the Pi network is used.
The Pi Network
The Pi network consists of three reactive elements: two shunt capacitors and one series inductor (or two shunt inductors and one series capacitor for a high-pass Pi). The topology resembles the Greek letter π — two vertical shunt elements on the left and right, and one horizontal series element across the top. This gives the Pi network one additional degree of freedom compared to the L network: the loaded Q can be set independently of the impedance ratio, within certain limits.
Why the Pi Network Dominates RF Output Circuits
The Pi network is the standard output circuit for RF power amplifiers, both transistor and valve, for several reasons:
- Variable Q: Unlike the L network, the Pi network allows Q to be selected independently. Higher Q means more harmonic suppression. Q = 10–15 is typical for amplifier output Pi networks, providing 40–60 dB of harmonic attenuation at the second harmonic.
- Harmonic suppression: The shunt capacitors present a low impedance path to ground for harmonics, while the series inductor presents high impedance. This low-pass filtering action is integral to the matching function.
- Wide impedance range: A single Pi network can match a high source impedance (1500–5000 Ω) or a low one to a 50 Ω load, and with variable components it can be retuned over a wide frequency range, which is why adjustable matching units use it.
- Practical tunability: With C1 and C2 as variable capacitors and L switched or continuously variable, the network can be adjusted for best output while watching the amplifier's supply current, a well-established technique described below.
Pi Network Design Formulas
For a Pi network matching RS (source, typically the higher impedance) to RL (load, typically 50 Ω), with loaded Q specified at the source side:
XC1 = RS / Q → C1 = Q / (2π × f × RS)
Loaded Q at load side:
QL = √(RS/RL × (1 + Q²) − 1)
Shunt capacitor at load (C2):
XC2 = RL / QL → C2 = QL / (2π × f × RL)
Series inductor:
XL = Q × RS + RL / QL
Wait — correct formula: XL = Q × RS − XC1_reactance_correction
Simplified: XL = (Q × RS + QL × RL) / (not directly additive)
Most accurate direct formula for L:
XL = Q × RS + RL × QL
(This is actually: X_L = Q*Rs from source side + Q_L*R_L from load side; these are the reactances "seen" by L from each side, and they add because L is in series between the two shunt elements)
L = XL / (2π × f)
Worked example: Pi network at 7.1 MHz, matching RS = 1500 Ω (a high-impedance amplifier output) to RL = 50 Ω, Q = 12.
The easiest way to design a Pi network is to treat it as two L networks back to back, meeting at a low “virtual” resistance Rv:
Rv = RS / (1 + Q²) = 1500 / 145 = 10.34 Ω
C1: XC1 = RS / Q = 1500 / 12 = 125 Ω → C1 = 1 / (2π × 7.1×10⁶ × 125) = 179 pF
Load-side Q: QL = √(RL/Rv − 1) = √(50/10.34 − 1) = √3.836 = 1.959
C2: XC2 = RL / QL = 50 / 1.959 = 25.5 Ω → C2 = 1 / (2π × 7.1×10⁶ × 25.5) = 879 pF
L: XL = Rv × (Q + QL) = 10.34 × 13.96 = 144.3 Ω → L = 144.3 / (2π × 7.1×10⁶) = 3.24 µH
The calculator below uses exactly these formulas. Notice that the network only works if Rv is smaller than both RS and RL; that sets the minimum usable Q, √(RS/RL − 1) = 5.39 here.
Q Factor and Harmonic Suppression
The harmonic suppression provided by a Pi network increases with Q. At the second harmonic (2 × f), the shunt capacitors present half the reactance (more effective bypassing) while the series inductor presents twice the reactance (more effective series blocking). The net result is that harmonic suppression approximately doubles for each unit increase in Q above about 5. Typical design guidelines:
| Loaded Q | Approx. 2nd harmonic attenuation | Relative insertion loss | Component values |
|---|---|---|---|
| 5 | 25–30 dB | Low | Smaller; easier to achieve |
| 10 | 40–45 dB | Moderate (0.1–0.3 dB) | Typical amplifier output network |
| 15 | 50–55 dB | Higher (0.3–0.6 dB) | High-performance design |
| 20 | 60–65 dB | Significant (>0.6 dB) | Large reactive values |
Q = 10–12 is the standard engineering compromise for most amplifier Pi networks: enough harmonic suppression to meet typical limits on harmonic emissions (often 40–60 dB below the fundamental) with low insertion loss and practical component values. Note that a Pi network designed for one frequency has a different Q at other frequencies, so an amplifier that must cover a wide frequency range needs its network switched or retuned.
The Pi-L Network
The Pi-L network connects a Pi network and an L network in cascade, with the L network on the output side. The combined structure has four reactive elements: two shunt capacitors, one series inductor (from the Pi section), and one additional series inductor (from the L section). The result is a network with substantially better harmonic suppression than a Pi network alone, at the cost of one additional component.
The Pi-L is the output network of choice where maximum harmonic suppression is required from a high-impedance amplifier. The additional L section improves the harmonic attenuation by an additional 15–20 dB compared to a Pi alone at the same Q, and the two sections together create a steeper rolloff. It is a common choice in high-power RF amplifiers for industrial, scientific and broadcast use.
Adjusting a Matching Network in Practice
When an amplifier has an adjustable Pi network — common on valve amplifiers and on RF test and industrial amplifiers — the classic adjustment procedure uses the amplifier's DC supply current and an output power meter:
- Set the input (tune) capacitor C1 to maximum capacitance, and the output (load) capacitor C2 to about mid-range.
- Apply reduced drive and observe the amplifier's supply current.
- Slowly adjust C1. The supply current dips to a minimum at resonance — the “dip” shows that the network is resonant at the operating frequency and the amplifier is delivering power efficiently.
- Adjust C2 (the load capacitor) to increase the output power read on the power meter. The supply current will rise slightly as C2 is adjusted.
- Alternate between small adjustments of C1 (to maintain the dip) and C2 (to peak output power). The correct operating point is reached when the supply current is at a moderate dip and output power is maximized.
- Increase the drive to full and repeat the final adjustment to confirm the operating point at full power.
A mistuned network causes excessive dissipation in the amplifier device (the current does not dip, and the transistor or valve overheats), poor harmonic suppression (the network is not at its design Q), and reduced output power. On a circuit board with fixed components, the same adjustment is done once at the design stage with a network analyzer, by trimming component values until the amplifier sees the impedance it needs.
Pi Network Calculator
Pi Network Impedance Matching Calculator
Calculates component values for a Pi network matching source to load impedance. Enter Q = 10–12 for a typical amplifier output network. RS should be the higher impedance.
L Network Calculator
L Network Impedance Matching Calculator
Calculates L and C for a two-element L network. Network Q is determined by the impedance ratio and cannot be independently set. Choose the low-pass topology (shunt C on the high-impedance side) for most amplifier outputs, because it also suppresses harmonics.
Frequently Asked Questions
What is the "dip" when adjusting an amplifier's output network, and why does it matter?
The dip is the minimum in the amplifier's DC supply current that occurs when the output network is resonant at the operating frequency. At resonance, the network presents a purely resistive load of the right value to the amplifier device, so the device converts the most DC power into RF output and wastes the least as heat. Above and below resonance, the network presents a reactive load: current and voltage are out of phase, the device dissipates more power, and the supply current rises. Tuning to the dip ensures the network is at resonance and the amplifier is operating at maximum efficiency. Running far off resonance makes the transistor or valve dissipate excessive power as heat instead of delivering it to the load, and can destroy it.
Can I use the Pi network calculator to design an adjustable matching unit?
Yes. An adjustable matching unit is essentially a variable Pi or T network that matches whatever load impedance is presented to the 50 Ω output of an amplifier or signal source. The calculator gives you the component values needed to match specific impedances at specific frequencies, which is useful for designing a fixed matching network for a known load, or for working out the range your variable components must cover. For loads whose impedance is unknown or changes, the practical approach is to use variable components (roller inductors, variable capacitors) and adjust them while watching a reflected-power or VSWR meter.
Test Your Knowledge
Answer the questions below to check your understanding of this lesson. Every answer can be found in the lesson above.