EC-4.5 The Laplace Transform
The standard Laplace transform treatment as taught in engineering mathematics courses, September 2026
What this is and why it exists
The Fourier transform has two problems for engineering. It does not converge for signals that grow. And it has no natural way to carry initial conditions, because it integrates over all time and engineering problems start at a definite instant.
The Laplace transform fixes both. Multiply the signal by a decaying exponential before transforming and things that would not converge now do. Start the integral at zero rather than at minus infinity and the initial conditions come along for free.
What you get in exchange is the most useful single move in applied mathematics. Differentiation becomes multiplication. A differential equation turns into an algebraic one, and algebra is something everybody can do.
The vocabulary
- Transform variable — the complex variable, usually written s, that replaces time.
- Region of convergence — the part of the complex plane where the defining integral converges.
- Transfer function — the transform of the output divided by the transform of the input.
- Pole — a value of s where the transform grows without bound.
- Zero — a value of s where the transform is zero.
- Partial fractions — splitting a rational transform into simple terms that can be inverted from a table.
- Initial value result — a limit of the transform giving the signal's value at the start.
- Final value result — a limit of the transform giving the signal's settled value, when it settles.
The mental model
The defining move is worth stating plainly. Transforming a derivative gives the transform of the function multiplied by s, minus the function's initial value. Every derivative brings down one factor of s and hands back one initial condition.
So a differential equation with constant coefficients becomes a polynomial equation. Solve it by ordinary algebra, then invert. The initial conditions are inside the algebra rather than applied at the end, which is the practical difference from the classical method.
The procedure has three steps and never changes. Transform every term. Solve for the transform of the unknown. Invert.
One piece is routinely dropped in engineering courses and is kept here because dropping it is a real error. The transform is only defined where its integral converges, which is a half plane. The same algebraic expression with two different regions of convergence describes two different signals, one that decays forward in time and one that grows backward. Engineering texts assume the forward case silently. Knowing that they have assumed it is what stops you from being confused when a text on stability suddenly cares.
Poles are where the transform grows without bound, and they carry the behaviour. Each pole contributes one exponential term to the answer, with the pole's value as the exponent. Real negative poles give decay. Complex pairs give decaying oscillation. This is the same reading of a complex number as in the second-order topic, arriving from a different direction.
Inverting is done by partial fractions in almost every practical case. Split the rational transform into simple terms, look each one up in a short table, and add. Repeated poles and complex pairs each need their own handling. Both are worth working through once, rather than being met for the first time in an examination. The table itself is the residue theorem computed in advance, which the contour integration topic explains.
Two shortcuts round it out. The initial value result gives the signal's starting value from a limit of the transform, without inverting anything. The final value result gives its settled value the same way, but only when the signal actually settles. Applying it to something that oscillates forever or grows gives a confident wrong number, and that is a common mistake worth naming.
What you should now be able to explain or do
- Say why the Laplace transform exists when the Fourier transform already does.
- Transform a differential equation and explain where the initial conditions went.
- Say what the region of convergence is and why quoting a transform without it is incomplete.
- Invert a rational transform by partial fractions, including a repeated pole.
- State the final value result and say when it must not be used.
Check yourself
What happens to a derivative under the Laplace transform, and why does that help?
It becomes the transform multiplied by s, minus the initial value. A differential equation turns into an algebraic one, with the initial conditions already inside it.
Why is a transform quoted without its region of convergence incomplete?
Because the same algebraic expression can come from two different signals, distinguished only by where the integral converges.
A transform has a pole at minus three and a complex pair at minus one plus and minus four i. Describe the response.
A term decaying with time constant one third, plus an oscillation at four radians per second inside an envelope decaying with time constant one.
When must the final value result not be used?
When the signal does not settle. Applied to something oscillating or growing it returns a number that means nothing.
Go deeper
We haven't checked most of these for screen reader use yet.