PCB

Impedance matching and controlled impedance in PCB

By CourseTron Team ·  7 min read ·  Updated Oct 2026 💬 𝕏 in f
Table of Contents
    🖨  Download / Print this guide (PDF)

    In short: Impedance matching in PCB means designing a trace so its characteristic impedance equals the impedance the driver and receiver expect, so a fast signal travels down the line without reflecting back. You control it through trace width, dielectric thickness and the dielectric constant of the laminate, and you enforce it with termination resistors at the ends of the line.

    What Is impedance matching in pcb?

    Impedance matching in PCB is the practice of making a signal path present a single, consistent impedance from the moment the driver launches an edge to the moment the receiver samples it. A copper trace over a reference plane is not a simple wire at high speed — it behaves as a transmission line with a defined characteristic impedance (Z₀). When the source impedance, the trace impedance and the load impedance all agree, energy flows forward cleanly. When they disagree, part of the signal bounces back as a reflection, corrupting the waveform.

    The reason this matters is the speed of the edge, not the clock frequency. The moment a signal's rise time is comparable to the time it takes to travel the trace, the trace must be treated as a controlled-impedance line. A rough rule: if the physical trace length exceeds about one-sixth of the rising edge's electrical length, reflections become significant and matching is mandatory.

    Key fact: For a typical FR-4 board the signal travels at roughly 15 cm per nanosecond. A 1 ns rise time therefore has a critical length near 2.5 cm — shorter edges turn ordinary short traces into transmission lines that demand impedance control.

    Characteristic impedance and where it comes from

    Characteristic impedance is set entirely by the trace geometry and the surrounding dielectric, not by the trace length. For a lossless line it is Z₀ = √(L/C), where L and C are the inductance and capacitance per unit length. In practical terms, on a PCB, Z₀ depends on four things:

    • Trace width (w) — wider traces have more capacitance to the plane, lowering Z₀.
    • Dielectric height (h) — the distance from the trace to its reference plane; more height raises Z₀.
    • Dielectric constant (εr) — the laminate's permittivity; higher εr lowers Z₀ and slows the signal.
    • Copper thickness (t) — a secondary effect that slightly lowers Z₀.

    The two dominant geometries are microstrip (a trace on an outer layer over one plane) and stripline (a trace buried between two planes). Microstrip is faster and easier to route; stripline is better shielded and more immune to crosstalk. Designers rarely compute Z₀ by hand — a field solver or the fab's stackup calculator returns the exact width for a target impedance. Common targets are 50Ω single-ended and 100Ω or 90Ω differential.

    Microstrip geometry sets Z₀ Dielectric (εr) GND reference plane signal trace width w height h Z₀ rises with h · Z₀ falls as w and εr riseCourseTron · coursetron.com
    How trace width, dielectric height and εr together determine the characteristic impedance of a microstrip line.

    Reflections: what goes wrong without a match

    When a signal meets a change in impedance, a fraction of it reflects. The reflection coefficient Γ = (Z_L − Z₀) / (Z_L + Z₀) tells you how much. An open end (Z_L = ∞) reflects the full wave positively; a short (Z_L = 0) reflects it inverted; a perfect match (Z_L = Z₀) gives Γ = 0 and no reflection at all.

    These reflections do real damage at the receiver:

    • Overshoot and undershoot that can forward-bias protection diodes or exceed absolute-maximum ratings.
    • Ringing — a decaying oscillation that blurs the voltage levels the receiver must distinguish.
    • False or double clocking when ringing crosses the logic threshold more than once on a single edge.
    • Timing jitter, because the settled level arrives later than the ideal edge.

    The animation below shows a forward edge launched into a mismatched line, hitting an unterminated open end and returning as a reflected wave — the mechanism behind ringing.

    A reflection on a mismatched line Driver openend forward edge → ← reflected wave (Γ = +1 for open)CourseTron · coursetron.com
    With no termination the edge reaches the open end and returns toward the driver — superimposing on the line as ringing.

    Termination strategies that enforce the match

    Termination absorbs the signal's energy so nothing reflects. The common schemes are:

    • Series (source) termination — a resistor at the driver sized so R_series + R_driver = Z₀. Simplest and lowest-power; ideal for point-to-point nets with a single load.
    • Parallel termination — a resistor equal to Z₀ at the receiver to ground. Excellent matching but draws steady current.
    • Thevenin termination — two resistors forming a divider to both rails; common on buses, at the cost of DC power.
    • AC (RC) termination — a resistor in series with a capacitor, matching the line to AC while blocking DC current.
    • Differential termination — a single resistor across the pair (or a split with a center capacitor) for 90Ω/100Ω links such as USB, LVDS and PCIe.

    Beyond discrete resistors, good matching depends on keeping the reference plane continuous under the trace, avoiding splits and stubs, and keeping vias and connectors — which are themselves impedance discontinuities — short and well referenced.

    Series (source) termination DriverR_driver R_series trace, Z₀ Receiver R_series + R_driver = Z₀ Reflection from the open receiver is absorbed at the sourceCourseTron · coursetron.com
    Series termination places a resistor at the driver so that R_series plus the driver output resistance equals Z₀.

    A practical workflow

    1. Decide target impedance from the interface spec (e.g. 50Ω single-ended, 100Ω differential).
    2. Give the stackup and target to your fabricator and get back exact trace widths and spacings.
    3. Route with continuous reference planes and matched lengths for differential pairs.
    4. Place termination at the correct end and simulate the net if edges are very fast.
    5. Specify controlled-impedance testing on the fab drawing so the board house verifies it with a TDR.
    Learn this hands-on

    Real projects and industry tool flows, with certification.

    Explore CourseTron Courses →

    impedance matching in pcb — FAQ

    Why is 50 ohms the standard single-ended impedance?

    50Ω is a practical compromise. For coaxial lines, minimum loss sits near 77Ω and maximum power handling near 30Ω, and 50Ω balances the two while being easy to realize with manufacturable trace widths on common stackups. Most driver and connector ecosystems standardized around it.

    Do low-speed designs need impedance matching?

    Usually not. What matters is edge rate, not clock frequency. A slow signal with a very fast edge can still require matching, while a modest-edge signal on a short trace does not. Compare the trace length to the critical length derived from the rise time to decide.

    What is the difference between microstrip and stripline?

    Microstrip runs on an outer layer over a single reference plane — faster and easy to probe, but more exposed to radiation and crosstalk. Stripline is buried between two planes — better shielded and slightly slower because it is fully surrounded by dielectric.

    How is controlled impedance verified on a finished board?

    Fabricators use a Time-Domain Reflectometer (TDR) on a dedicated coupon built into the panel. The TDR launches a fast edge and reads the reflected profile to confirm the trace impedance is within tolerance, typically ±10%.

    Related: signal integrity in PCB · Glossary · all courses.

    Advertisement