Matched segment reflection containment

When a matching network is inserted between a Transmission line and a mismatched load, the feedline sees a perfect match — but reflections still occur between the matching plane and the load. The matching network doesn’t eliminate reflections; it confines them to a short segment beyond the matching plane.

This is a subtle property of impedance-matched circuits that’s easy to miss. The matching network gives the source-facing side what it wants (zero reflection coefficient looking into the network), while the load-facing side may still have huge $|\Gamma |$.

The geometry

Two distinct planes in the circuit:

• Source side, on the feedline. Looking into the matching network sees $Z_0$ → no reflections back toward the source. ${\Gamma }_{\text{source}}=0$.
• Load side, between the matching network and the actual load $Z_L$. The waveform here is determined by ${\Gamma }_L=(Z_L-Z_0)/(Z_L+Z_0)$, which may be large (e.g., $|{\Gamma }_L|=0.9$ if $Z_L$ is far from $Z_0$).

The matching network presents one impedance to one side and a different impedance to the other side. Reflections in the short segment between matching plane and load do bounce back and forth — they just can’t escape past the matching plane into the feedline.

Why this works

The matching network is, in essence, a transformer between two impedance environments. From the source side it looks like $Z_0$; from the load side it looks like $Z_L^{\ast }$ (in the case of conjugate matching, for maximum power transfer). The standing wave that exists between the matching plane and the load is bounded geometry — confined to that section.

In a single-stub or quarter-wave-transformer match:

• Between the source and the matching point, only a forward wave exists. No standing wave. $|\Gamma |=0$ at every point.
• Between the matching point and the load, forward and reflected waves coexist. A standing wave pattern exists. $|\Gamma |$ depends on position, varying between 0 (at the matching plane!) and $|{\Gamma }_L|$ (at the load).

The matching plane is where ${\Gamma }_d$ (the position-dependent reflection coefficient) transitions from “nonzero behind me” to “zero ahead of me.” It’s a one-way mirror for reflections.

Why this matters

Source amplifier protection. A high-power RF transmitter is designed to deliver power into a known impedance (typically 50 Ω). If the antenna or load is mismatched and the feedline has no matching network, reflected power returns to the source and may damage the amplifier. A matching network at the load end makes the feedline see 50 Ω regardless of the actual load — protecting the source. The reflections still exist (between matching plane and load), but they’re contained in the matching network section.

Bandwidth considerations. The matching is exact only at the design frequency. Off-frequency, $|\Gamma |$ on the feedline rises from 0 — reflections leak past the matching plane. The amount of leakage determines the matching bandwidth.

Hot spots in the matched section. Between matching plane and load, the standing wave pattern has voltage maxima and minima. A current maximum in a lossy line dissipates more $I^{2}R$ power locally. A voltage maximum across a lossy dielectric dissipates more $V^{2}G$. In high-power applications, these standing-wave maxima can produce localized heating or breakdown in the matching section — even though the feedline sees no SWR. The matching section needs to be rated for the local conditions, not the feedline conditions.

Why a matching network can still be lossless. Reactive components (inductors, capacitors, transmission-line stubs) absorb zero real power. The reflections inside the matching section are a redistribution of phases, not a dissipation. The net real power flowing through the matching plane equals the real power delivered to the load — all of it, with no resistive loss in the (ideal) matching network.

Worked picture: quarter-wave transformer

A 50 Ω line feeds a 50 → 100 Ω quarter-wave transformer (so $Z_T=\sqrt{50\cdot 100}=70.7$ Ω), which feeds a $Z_L=100$ Ω load.

On the 50 Ω feedline: looking into the transformer sees the transformed impedance $Z_T^2/Z_L=70.7^2/100=50$ Ω. Matched. $\Gamma =0$. Only a forward wave on the feedline.

On the 70.7 Ω transformer section: looking from the start of the transformer toward the load sees the actual load (after going through the $\lambda /4$ line). Reflection coefficient on the transformer section at the load end is $(100-70.7)/(100+70.7)=0.172$. A standing wave exists with SWR $\approx 1.4$. After traveling $\lambda /4$ to the input of the transformer, the standing-wave pattern has rotated 90° on the Smith chart — and a different value of $\Gamma$ is seen there, but the overall input impedance to the transformer (and thus to the feedline) lands on 50 Ω.

The reflections live between the load (at $z=0$) and the input of the transformer (at $z=-\lambda /4$). They never escape past the transformer’s input plane onto the feedline.

In single-stub matching

Same story. The stub is placed at a distance $d$ from the load where the line admittance has the right real part. The stub adds reactance to cancel the imaginary part, achieving a match at the stub plane.

• Source-side of the stub: $\Gamma =0$, single forward wave.
• Load-side of the stub: standing wave, $|\Gamma |$ varying with position, settling to $|{\Gamma }_L|$ at the load.

The stub itself is open- or short-circuited — fully reflective. But the bounce between stub and load combines with the bounce between load and the line, creating the right superposition for matched-input behavior.

When the matched section becomes important

In low-power applications, the standing wave in the matched section is harmless — it just sits there cycling energy back and forth, the only loss being the small dissipation in real components.

In high-power applications:

• Voltage maxima can exceed dielectric breakdown.
• Current maxima can overheat conductors.
• Mismatch loss in the matched section itself becomes significant.

Designers therefore worry about the internal SWR of the matching network, not just the input-side match. Good matching circuits minimize the length of the section with high SWR.

Takeaway

“Matched” doesn’t mean “no reflections anywhere.” It means “no reflections reaching the source.” Reflections inside the matching segment are a fundamental and unavoidable feature of any non-trivial matching network — they’re how the matching works. Recognizing this distinction matters for high-power design, for bandwidth analysis, and for understanding why matching networks need to be physically and thermally sized for the worst-case standing-wave conditions in their interior.