Understanding Characteristic Resistance in High-Speed PCB Traces and Coax

Characteristic resistance is the apparent opposition a transmission line presents to a high-frequency signal propagating down it, determined solely by the line's distributed inductance and capacitance rather than its DC wire resistance. When you send a fast digital edge down a PCB trace or a coaxial cable, the conductor doesn't just act as a perfect short circuit; to the rising edge of the signal, it acts like a resistor. If your source and load impedances don't match this value, your signal bounces back, causing ringing, overshoot, and phantom clock edges that can completely derail a high-speed digital bus or RF link.

In radio frequency (RF) and high-speed digital design, this property is most commonly called characteristic impedance ($Z_0$). However, for a lossless transmission line (where the DC resistance of the wire and the dielectric conductance are negligible), the imaginary part of the impedance drops to zero. The value becomes purely real, meaning it behaves exactly like a resistor to a traveling wave. Hence, the term characteristic resistance ($R_0$ or $Z_0$) is technically precise for the real-world behavior of high-quality cables and PCB traces at high frequencies.

The Physics: Why a Wire Acts Like a Resistor to AC

To understand characteristic resistance, you have to stop looking at a wire as a single lumped component and start viewing it as a distributed network. Every inch of a PCB trace or coaxial cable possesses a tiny amount of series inductance ($L$) due to the magnetic field generated by current flow, and a tiny amount of shunt capacitance ($C$) due to the electric field between the signal conductor and the ground plane.

Notice what is missing from that formula: length. The characteristic resistance of a 50-ohm microstrip trace is 50 ohms whether the trace is 2 millimeters long or 2 meters long. It is an intrinsic property of the cross-sectional geometry and the dielectric material, not the total length of the wire.

Worked Numeric Example: Calculating 50-Ohm Coax

Let's look at a real-world benchmark: standard RG-58C/U coaxial cable, the ubiquitous 50-ohm cable used for bench oscilloscope probes, ham radio antennas, and SDR (Software Defined Radio) setups.

According to the All About Circuits transmission line textbook, the physical geometry of RG-58 (a 20 AWG stranded copper center conductor, solid polyethylene dielectric, and braided copper shield) yields specific distributed values:

• Distributed Inductance ($L$): ~253 nH/m ($253 \times 10^{-9}$ H/m)
• Distributed Capacitance ($C$): ~101 pF/m ($101 \times 10^{-12}$ F/m)

Plugging these into our formula:

$Z_0 = \sqrt{\frac{253 \times 10^{-9}}{101 \times 10^{-12}}}$

$Z_0 = \sqrt{2504.95}$

$Z_0 \approx 50.05 \Omega$

This 50-ohm value is not arbitrary; it represents the optimal geometric compromise between maximum power handling (which peaks around 30 ohms) and minimum signal attenuation (which peaks around 77 ohms) for air-dielectric coaxial lines. When you terminate this cable with a 50-ohm load resistor, the traveling wave sees no change in environment, and 100% of the energy is absorbed. Zero reflection.

Where You Meet Characteristic Resistance in Practice

You will encounter characteristic resistance whenever the physical length of your interconnect becomes a significant fraction of the signal's wavelength, or more practically in digital design, when the trace propagation delay exceeds half the signal's rise time. Here is where you will find specific $Z_0$ targets in modern electronics:

For PCB designers, achieving these targets requires using a stackup calculator provided by your fab house, as the exact copper thickness and prepreg resin content will shift your capacitance and alter the final characteristic resistance.

Real-World Scenario Walkthrough: The Un-Terminated SPI Bus

Theory is clean; the workbench is messy. Here is a classic signal integrity failure involving characteristic resistance that plagues hobbyists and junior engineers alike.

The Setup

You are building a data logger using an ESP32-WROOM-32 dev board and an external 18-bit ADC (like the ADS8861) communicating via SPI. The clock line (SCK) is routed as a 12cm (4.7 inch) microstrip trace on a standard 1.6mm FR4 PCB. Because you are pushing the SPI clock to 20 MHz to get fast sample rates, the signal edges are sharp—roughly 3 nanoseconds rise time.

The Numbers

• Trace Characteristic Resistance ($Z_0$): 60 $\Omega$ (based on your trace width and dielectric).
• ESP32 GPIO Output Impedance ($R_{source}$): ~25 $\Omega$ (typical for 3.3V CMOS outputs).
• ADC Input Impedance ($R_{load}$): >1 M$\Omega$ (essentially an open circuit to high-frequency edges).

The Outcome

You hook up your oscilloscope and probe the ADC's SCK pin. Instead of a clean 3.3V square wave, you see massive overshoot ringing up to 5.5V, followed by a negative undershoot dipping below 0V. Worse, the ADC is returning garbage data, triggering on the phantom secondary edges caused by the ringing.

What Went Wrong

When the ESP32 drives the pin high, the 25-ohm source and the 60-ohm trace form a voltage divider. The initial wave launched down the trace is only $3.3V \times (60 / (25 + 60)) = 2.32V$. When this 2.32V wave hits the 1 M$\Omega$ ADC input, it sees an open circuit. The reflection coefficient ($\Gamma$) is practically +1. The wave reflects entirely, doubling the voltage at the receiver to 4.64V. This reflection travels back to the ESP32, bounces again, and sets up a standing wave of ringing.

The Fix: Series Termination

To solve this, we must match the source to the characteristic resistance of the trace. We use a technique called source-series termination.

1. Calculate the required resistor: Target $Z_0$ (60 $\Omega$) minus ESP32 internal $R_{source}$ (25 $\Omega$) = 35 $\Omega$.
2. Select the component: Choose a standard 33 $\Omega$ or 36 $\Omega$ 0402 SMD resistor (1% tolerance).
3. Placement is critical: Solder this resistor as close to the ESP32 GPIO pin as physically possible (within 2mm). If you place it near the ADC, it becomes a useless stub.
4. Verify: Probe the ADC pin again. The initial edge will rise to exactly half-voltage (~1.65V), and when the reflection returns from the open-circuit ADC, it will step up to a perfectly flat 3.3V with zero ringing.

What People Commonly Confuse It With

The biggest trap for beginners is conflating characteristic resistance with DC resistance. If you take a 10-meter spool of 50-ohm RG-58 coax and measure it with your Fluke multimeter on the resistance setting, you will read roughly 0.3 ohms. The multimeter applies a DC voltage, meaning the inductors act as shorts and the capacitors act as opens. You are only measuring the physical copper resistance. Characteristic resistance only exists for changing signals (AC or digital edges) where the distributed L and C network is actively charging and discharging.

Another common confusion is mixing up characteristic resistance (a property of the cable/trace geometry) with load resistance (the physical resistor placed at the end of the line). You can have a 50-ohm cable terminated with a 100-ohm load. The cable's characteristic resistance remains 50 ohms, but because the load doesn't match it, you will get signal reflections.