Yes. A wire starts as a resistor, but it does not stop there. At low speed, the simplest model is its series resistance, given by R=L/A, so longer and thinner wire causes more voltage drop and less current. As signals get faster, the wire’s geometry and return path matter: current changes create voltage via V=L di/dt, electric fields store charge as capacitance, and a long enough cable behaves as a transmission line with characteristic impedance Z0L′/C′ and propagation velocity v1/L′C′. A practical rule from Texas Instruments is that if a signal’s rise time is more than four times the cable’s one-way delay, the cable can often be treated as not being a transmission line for that signal; otherwise reflections matter. For sinusoidal signals, the same idea appears as electrical length: once wire length is no longer small compared with wavelength, the line matters. [1]
For the three beginner tests, the dominant effect changes. A slow LED circuit mostly notices resistance, so the symptom is dimming or a small current reduction. A fast digital pulse notices inductance, capacitance, and reflections, so the symptom is ringing, overshoot, undershoot, and timing shifts. A long cable driving a load can look like a big capacitor when the source is slow, but like a transmission line when the source edge is fast; then termination, cable impedance, and propagation delay become central. Skin effect is a brief but important footnote: AC current crowds toward the conductor surface at higher frequency, which increases effective resistance and cable loss. [2]
The examples below assume ordinary copper conductors and safe, low-voltage hobby setups because the wire type, insulation, and operating voltage were unspecified. Do not try any of these experiments on mains wiring. Core source links are embedded in the citations throughout this report, with the main foundations coming from OpenStax/LibreTexts, MIT OpenCourseWare, Texas Instruments, NXP, CommScope, and Times Microwave. [3]
A practical way to decide how to model a wire
A good beginner workflow is: start with resistance, then add lumped inductance and capacitance, and only move to full transmission-line thinking when the wire’s delay is no longer tiny compared with the signal’s rise time or wavelength. For low-loss lines, Z0L′/C′, vp1/L′C′, and =vp/f. For digital edges, a rise-time criterion is often more useful than wavelength, because even a “slow” clock can have very fast edges. [4]
This progression also matches the standard distributed RLGC transmission-line model: real wires have finite resistance, and the spacing between conductors creates distributed capacitance and inductance per unit length. Steer notes that lumped subsections work well when each subsection is much shorter than a wavelength, about /20 or less. [5]
Start with resistance and series voltage drop
For a wire of length L, cross-sectional area A, and resistivity , the resistance is R=L/A. Copper at 20C has resistivity about 1.6810−8 m, and copper resistance increases with temperature. In a two-wire circuit, remember to count the loop length: current must go out and come back, so a 20 m run usually means about 40 m of conductor in the current path. [6]
In a slow LED circuit, that series resistance simply adds to your resistor. Suppose a 5 V supply drives a red LED at about 2 V through a 220 resistor. With no cable resistance, the current is I5−2/220=13.6 mA. If the LED is 100 m away using thin copper wire roughly like 24 AWG, the 200 m loop resistance is about 16.4 , so the current becomes about 12.7 mA, about 7% lower. At only 20 m one-way, the loop resistance is about 3.28 , and the current change is only about 1.5%. That is why short hookup wires usually seem “invisible” in beginner DC circuits, while very long or very thin wires do not. [7]
A safe experiment is simple: build two identical LED circuits from 5 V, one with short leads and one with a long spool of thin wire in series. Measure current with a multimeter and, if possible, compare brightness side by side. The expected result is not ringing or oscillation; it is plain current reduction from added series resistance. [8]
Add inductance, capacitance, and frequency dependence
Once current changes quickly, resistance is no longer the only story. NXP’s grounding note states the key relation directly: a change in current through inductance creates a voltage, V=L di/dt. It also emphasizes that return current matters, because loop size and return-path inductance strongly affect noise. Short, wide traces, close return paths, and twisted conductors reduce inductance; large loops make it worse. [9]
The same wire pair also has capacitance. Texas Instruments models real lines with distributed resistance, inductance, capacitance, and shunt conductance, while MIT and TI both give the classic low-loss relationships Z0L′/C′ and v1/L′C′. Real cable data make these values concrete. One CommScope Cat 6 cable lists Z0=100 , mutual capacitance 5.6 nF/100 m or 56 pF/m, nominal velocity factor 68%, and propagation delay 536 ns/100 m, about 5.36 ns/m. From L′=Z02C′, that implies L′0.56 H/m. A Times Microwave LMR-195 coax lists Z0=50 , C′=83.3 pF/m, L′=0.21 H/m, and time delay 4.17 ns/m. The same wire pair also has capacitance. Texas Instruments models real transmission lines with distributed resistance, inductance, capacitance, and shunt conductance, while both Texas Instruments and MIT derive the classic low-loss relationships Z0 = √(L′/C′) and v = 1/√(L′C′). 3, 4
Real cable specifications make these values concrete. One CommScope Category 6 cable lists Z0 = 100 Ω, mutual capacitance = 5.6 nF/100 m (56 pF/m), a nominal velocity factor of 68%, and a propagation delay of 536 ns/100 m (5.36 ns/m). 9 From L′ = Z02C′, this implies L′ ≈ 0.56 µH/m.
A Times Microwave LMR-195 coaxial cable lists Z0 = 50 Ω, C′ = 83.3 pF/m, L′ = 0.21 µH/m, and a propagation delay of 4.17 ns/m. 10
Skin effect adds one more layer at higher frequency. For good conductors, skin depth is approximately 1/f. Using copper conductivity from OpenStax, the skin depth is about 8.4 mm at 60 Hz, 65 m at 1 MHz, and 6.5 m at 100 MHz. That means skin effect is usually negligible for short DC hobby wiring, but it becomes important in RF and in very fast, long interconnects because effective resistance and attenuation rise with frequency. The LMR-195 datasheet shows this trend directly: attenuation rises from 6.5 dB/100 m at 30 MHz to 55.4 dB/100 m at 2 GHz. [11]
When the wire becomes a transmission line
For a sine wave, transmission-line behavior matters when the wire is not electrically short. Steer gives =vp/f and electrical length ℓ. In a cable with 68% velocity factor, the wavelength at 100 MHz is about 2.04 m, so a 20 cm wire is already about 0.1, no longer “nothing.” [12]
For digital signals, the rise time criterion is more practical. TI’s rule says that if rise time is greater than four times the one-way propagation delay, the cable is no longer considered a transmission line for that signal. Rearranged, a rough critical length is ℓcritv tr/4. For cable velocity around 0.68c, that is about 5.1 cm for a 1 ns edge, 25 cm for a 5 ns edge, and 51 cm for a 10 ns edge. Very fast outputs therefore make even “short” wires electrically long. [13]
Once you are in that regime, reflections are described by the voltage reflection coefficient
Γ = ZL − Z0 ZL + Z0
If ZL=Z0, then =0 and there is no reflection. If the load is open, +1, so the voltage reflection is positive; if it is shorted, =−1. This is the mathematical reason a wire is sometimes “not just a connection”: it can store and launch traveling waves whose behavior depends on impedance matching, not only on Ohm’s law. [14]
What the three beginner tests should show
Slow LED circuit
Expected behavior: mostly resistive. If you blink an LED slowly through a long wire, the wire mainly acts like added series resistance. The human eye does not care about nanosecond cable delay in a 1 Hz blink test, and the LED current is set mainly by the resistor plus the wire’s DC resistance. In the 5 V, 220 , red-LED example above, 100 m one-way of thin copper drops current from 13.6 mA to 12.7 mA; 20 m one-way barely changes it. [15]
A simple safe experiment is to run the same LED circuit over short leads and then over a long spool of wire, keeping the supply at 5 V from USB or a bench supply. If you want a larger visible effect, use thinner wire or longer distance. If you want less effect, use thicker wire or place the current-limiting resistor near the load so the long run carries less dynamic current. [6]
Fast digital pulse
Expected behavior: ringing, overshoot, undershoot, and delay. Take a 2 m twisted pair with Cat 6-like parameters. The one-way delay is about 10.7 ns, so a 5 ns edge is already in transmission-line territory by TI’s rule. If the receiving input is high impedance, L+1, so a positive reflection returns from the far end, and without source termination the waveform can bounce for several round trips before settling. [16]
Inductance also matters even before you think in full traveling-wave terms. The same pair has about L′0.56 H/m, so 2 m is about 1.1 H. If a current changes by only 10 to 20 mA in 5 ns, the order-of-magnitude voltage from L di/dt is a few volts. In practice the exact value depends on the real return path, source impedance, and current path, but the lesson is robust: fast edges make wiring inductance visible. Twisting the pair or keeping the return path close lowers loop inductance and reduces trouble. [17]
A beginner experiment is to drive a long twisted pair from a logic buffer or microcontroller output and watch the far end on an oscilloscope. Compare three cases: no termination, a series resistor near the source, and a matched load at the far end. The repetition rate can be low; what matters is the edge rate, not the clock frequency. [18]
Long cable driving a load
Expected behavior: either capacitive slowing or transmission-line behavior, depending on the source. A 50 m Cat 6 pair has Ctotal2.8 nF and delay about 280 ns. If it is driven through a 1 k source into a high-impedance load, the cable acts like a lumped capacitor with =RC2.8 s. A 1 s pulse would only rise to about 1−e−1/2.830% of final value, so pulses get rounded and may never cross logic thresholds reliably. TI explicitly notes that the familiar i=C dv/dt treatment is reasonable when rise time is long compared with line delay. [19]
If the same cable is driven with a 5 ns edge, it is no longer “just a capacitor”; now it is unquestionably a transmission line. With an open load, reflections dominate. With a matched 100 termination, reflections are suppressed, but the driver must supply the necessary current. On 10 m of LMR-195 coax, the delay is 41.7 ns and total capacitance is about 833 pF; driven slowly through 1 k, that looks like 0.83 s, but driven fast it must be treated as a 50 line. [20]
A safe experiment is to send a pulse down 5–10 m of Ethernet cable or coax and compare the far-end waveform with the load open and then terminated near Z0. Beginners with only a multimeter can still explore the slower capacitive case by charging the cable through a large resistor and timing the voltage rise. [21]
Troubleshooting, mitigation, and a quick comparison table
The fixes follow directly from the physics. To reduce resistive effects, shorten the wire or use thicker copper. To reduce inductive noise, keep the return path close and make the loop area small; twisted pair helps for exactly this reason. To control capacitance and impedance, use cable intended for signaling rather than arbitrary loose wires. To suppress reflections, terminate close to Z0, either at the load or with source-series termination where appropriate. To keep digital power rails quiet, place decoupling capacitors close to the IC so the cable and supply wiring do not have to deliver the fastest current spikes. Shielding helps mainly with external coupling and emissions, while series resistors near the source often tame ringing by slowing edges and improving source match. [22]
The core idea is simple enough for beginners and rigorous enough for design work: a wire is sometimes “just a connection,” but only when the circuit is slow enough and short enough that its resistance, inductance, capacitance, and propagation delay do not matter. As soon as those quantities are comparable to the rest of the circuit, the wire becomes part of the circuit’s behavior. [31]
References
- Resistivity and Resistance – Physics LibreTexts
- AN1259 – NXP
- TI Application Note SNLA026
- MIT OpenCourseWare – Electromagnetics and Applications, Chapter 7
- Skin Depth – Physics LibreTexts
- Transmission Line Theory – Engineering LibreTexts
- AN-903: A Comparison of Differential Termination Techniques (Rev. B)
- Voltage Reflection Coefficient – Engineering LibreTexts
- CommScope CS30CM BLU C6 4/24 U/UTP RIB 305M Datasheet
- LMR-195 Datasheet