The Telegrapher Equations in Plain Terms
The telegrapher equations link voltage and current along a line through R, L, G and C per unit length. Their solution is one forward and one backward wave. On a lossless line with vp = 2.0e8 m/s, a pulse crosses 1 m in 5 ns.
Engineer’s note: The telegrapher equations look heavy, but the core idea is simple: a signal travels as a wave, not as a voltage that appears everywhere at once. Fast digital links like PCIe and USB4 now behave like RF for exactly this reason. Understand how a wave reflects at an impedance discontinuity and you're most of the way to solving signal integrity problems.
For practitioners
Why a line needs its own equations
Circuit theory assumes a wire has the same voltage along its whole length at any instant. That only holds when the wire is short compared with a wavelength, or compared with the distance a signal edge travels during its rise time. A PCB trace, a coax cable or an antenna feed often breaks that assumption. Voltage and current then depend on both position z and time t.
The telegrapher equations handle this. They do not treat the line as one component. They cut it into sections so short that each one does behave like a lumped circuit, then let the section length go to zero. The result is a pair of partial differential equations that link v(z,t) and i(z,t).
One short section of line
Take a piece of line of length Δz. Its two conductors have some series resistance and some series inductance. Between them there is some capacitance and some leakage conductance. All four scale with length, so we write them per unit length: R in Ω/m, L in H/m, G in S/m and C in F/m. The section then holds R·Δz and L·Δz in series, followed by C·Δz in parallel with G·Δz across the line.
Apply Kirchhoff's voltage law around the section. The voltage at the output equals the input voltage minus the drop across the series elements:
Apply Kirchhoff's current law at the output node. The current that leaves equals the current that enters minus what flows through the shunt elements:
Divide both by Δz and let Δz go to zero. The differences become derivatives and you get the telegrapher equations:
Read them like this. The first equation says voltage falls along the line because current flows through series R and L. The second says current falls along the line because some of it leaks away through shunt G and charges shunt C. Each equation contains both v and i, so neither can be solved on its own.
From time domain to phasors
For a single frequency ω, write v(z,t) = Re[V(z)·e^(jωt)] and do the same for current. Each time derivative becomes a multiplication by jω, and the partial derivatives in z become ordinary ones:
Here R + jωL is the series impedance per unit length and G + jωC is the shunt admittance per unit length. Most RF line analysis starts from this phasor form.
The wave equation and its two solutions
To uncouple the equations, differentiate the first one with respect to z and substitute the second:
This is a second-order equation with constant coefficients. Its general solution has two exponential terms:
Write γ = α + jβ. The term V0+·e^(-γz) gets smaller by e^(-αz) and lags in phase by βz as z increases. Put back the e^(jωt) factor and this is a wave moving toward +z. The term V0-·e^(+γz) is the same kind of wave moving toward -z. α is the attenuation constant in Np/m and β is the phase constant in rad/m.
Putting V(z) back into dV/dz = -(R + jωL)·I gives the current:
The minus sign in front of the backward term matters. For each wave on its own, the ratio of voltage to current is Z0. Their sum is not, which is why the impedance you see changes along a mismatched line. The two constants V0+ and V0- are set by the boundary conditions: the source at one end and the load at the other. If the load equals Z0, then V0- = 0 and only the forward wave exists.
The lossless case: a pulse on a 1 m line
Set R = 0 and G = 0. Then γ = jω·sqrt(LC), so α = 0 and β = ω·sqrt(LC). The phase velocity is
It does not depend on frequency. In the time domain, the lossless equations combine into ∂²v/∂z² = LC·∂²v/∂t². Any waveform of the form f(t - z/vp) solves it. Every frequency component of a pulse travels at the same speed and none is attenuated, so the pulse arrives with the same shape it started with.
Worked example. Take a 1 m lossless line with vp = 2.0e8 m/s and launch a pulse at z = 0 at t = 0. Position is z = vp·t:
- t = 1.25 ns: z = 2.0e8 × 1.25e-9 = 0.25 m
- t = 2.5 ns: z = 0.5 m
- t = 3.75 ns: z = 0.75 m
- arrival at z = 1 m: t = 1 m / 2.0e8 m/s = 5 ns
At each of these instants you see the same pulse, moved along by the same step. This is the forward wave V0+·e^(-γz). Whether a backward wave V0-·e^(+γz) then starts back from z = 1 m depends only on the load. A matched load absorbs the pulse. Any other load sends part of it back toward the source, and that part takes another 5 ns to get there.
On a real line, R and G are not zero. α then shrinks the pulse as it travels. Because γ is no longer a simple multiple of ω, its frequency components also spread apart slightly. The lossless result is still the right first estimate for delay.
In plain words · for students and beginners
The telegrapher equations in plain words
An everyday picture
Picture a long rope tied to a far wall. Flick your end once and a bump runs along the rope. The rope as a whole doesn't move toward the wall, only the bump does. A signal on a cable behaves the same way: a bump of voltage travels down the line at a fixed speed.
Where the picture breaks down: the rope carries only one thing, the sideways bump. A cable always carries two things together, voltage (the electrical "push") and current (the flow of charge). Each one drives the other. That is why there are two telegrapher equations instead of one. Also, the bump on a rope always bounces off the wall. On a cable, a perfectly matched end soaks up the bump completely, so nothing comes back.
The same example, step by step
We use a 1 m cable with no loss ("lossless" means nothing is turned into heat). The signal speed is vp = 2.0e8 m/s, which is two hundred million metres per second. That speed comes from the cable's inductance L (how much it resists changes in current) and capacitance C (how much charge it stores between the two conductors): vp = 1/sqrt(LC).
- Start the pulse at z = 0 at t = 0. Here z means position along the cable.
- After 1.25 ns (a nanosecond is one billionth of a second), the pulse is at 0.25 m.
- After 2.5 ns it is at 0.5 m. After 3.75 ns it is at 0.75 m.
- To cover the full metre: 1 m / 2.0e8 m/s = 5 ns.
- The pulse keeps exactly the same shape the whole way, because the cable has no loss.
The arrow labelled "forward wave" is the V0+ part of the solution, the part moving away from the source. If the far end isn't matched, a second wave (V0-) comes back the other way.
Mistakes beginners make
- Using total values instead of per-metre values. R, L, G and C are given per metre of cable. Totals give the wrong speed.
- Forgetting the backward wave. The maths always allows two waves. The far end decides whether the backward one shows up.
- Expecting voltage to be the same everywhere at once. On a long line, one end can be at a peak while the other end is still at zero.
- Thinking the charge itself moves at vp. What travels at vp is the signal, not individual electrons.
Remember
- The two equations say: voltage drops because of series R and L, and current drops because of shunt G and C.
- The solution is always one forward wave plus one backward wave: V(z) = V0+·e^(-γz) + V0-·e^(+γz).
- On a lossless line the pulse moves at vp = 1/sqrt(LC) and keeps its shape. At 2.0e8 m/s it crosses 1 m in 5 ns.
Related pages on RF View
Frequently asked questions
What are the telegrapher equations?
They are two coupled equations for voltage and current on a transmission line: ∂v/∂z = -R·i - L·∂i/∂t and ∂i/∂z = -G·v - C·∂v/∂t. R, L, G and C are the line's resistance, inductance, conductance and capacitance per unit length.
How do the telegrapher equations lead to forward and backward waves?
Combining them gives the wave equation d²V/dz² - γ²V = 0. Its general solution is V(z) = V0+·e^(-γz) + V0-·e^(+γz), which is one wave travelling toward +z and one toward -z.
What is the propagation constant γ?
γ = sqrt((R + jωL)(G + jωC)). Its real part α sets the attenuation per unit length, and its imaginary part β sets the phase change per unit length.
How fast does a pulse travel on a lossless line?
It travels at vp = 1/sqrt(LC) and keeps its shape. With vp = 2.0e8 m/s, a pulse takes 5 ns to cross a 1 m line.