Laplace transforms in practice: table, partial fractions, inverse
Transform the equation, split the result into table-sized pieces, and read the time response off the table.
THE SHORT ANSWER
The inverse Laplace transform recovers a time signal f(t) from F(s). In practice you expand F(s) in partial fractions and read each term from a transform table: a real pole gives an exponential, a repeated pole adds t times an exponential, and a complex pair gives a damped sinusoid. The initial- and final-value theorems check the result, the latter only when every pole of sF(s) lies in the open left half-plane.
Solve a linear differential equation with initial conditions by hand: transform it, expand in partial fractions, invert, and check the result with the initial- and final-value theorems.
The previous chapter turned a derivative into multiplication by s and produced a transfer function. It left the return trip open: given Y(s), what is y(t)? This chapter supplies the table, the rules and the partial-fraction technique that take you back to the time domain.
Review the previous lesson: Why use the Laplace transform?First, picture it
Nobody evaluates the Laplace integral in the middle of a control problem. You work like a translator with a small dictionary: about eight signal–transform pairs and a handful of rules. A differential equation goes in, algebra gives you Y(s), and Y(s) comes out as one large fraction that is not in the dictionary. Partial-fraction expansion breaks it into small fractions that are. Each small fraction belongs to one pole, and each pole contributes one term of the time response: a real pole gives an exponential, a repeated pole gives t times an exponential, and a complex pair gives a damped sinusoid.
Build the idea step by step
The method has four steps. First, transform every term of the differential equation; the derivative rule brings the initial conditions in as extra terms. Second, solve the resulting algebraic equation for Y(s). Third, factor the denominator to find the poles and expand Y(s) in partial fractions: distinct real poles, repeated poles and complex-conjugate poles each have their own form. Fourth, invert term by term with the table. Finish with two quick checks. The initial-value theorem must reproduce y(0+), and the final-value theorem gives y(∞) when every pole of sY(s) lies in the open left half-plane. If a check fails, the mistake is almost always in a partial-fraction coefficient.
Read the equations
This integral defines the one-sided Laplace transform. Only t ≥ 0 enters, so f(t) is treated as zero before t = 0. You rarely evaluate it; the table below holds the results you need.
Table, part 1: impulse, unit step, ramp and powers of t. Read ↔ from left to right as 'transforms to' and from right to left as 'inverts to'.
Table, part 2: exponential, and t times an exponential. A factor (s + a) in a denominator is a real pole at −a and gives a term e^(−at); a squared factor is a repeated pole and gives a term te^(−at).
Table, part 3: sine and cosine. Poles at ±jω on the imaginary axis give an oscillation that never decays.
Table, part 4: damped sine and cosine. The denominator (s + a)² + ω² means poles at −a ± jω: a sets the decay rate and ω the oscillation frequency. These two rows invert every complex-pole term.
Linearity: transform a sum term by term and pull constants out. This rule lets you invert a partial-fraction expansion one fraction at a time.
Derivative rule: each derivative multiplies by s and subtracts initial conditions. This is where y(0) and ẏ(0) enter a problem. With zero initial conditions the rule reduces to multiplication by s and s².
Integration rule: integrating in time divides by s. An integrator in a block diagram is therefore the block 1/s.
Shift rules. Multiplying by e^(−at) shifts s to s + a, which is how the damped rows follow from the undamped ones. Delaying a signal by T seconds multiplies its transform by e^(−sT).
Final-value theorem. It is valid only when every pole s_k of sF(s) has a negative real part. Counter-example: for F(s) = 1/[s(s − 1)] the limit gives −1, but f(t) = e^t − 1 diverges.
Initial-value theorem, valid when F(s) is strictly proper. Use it to check that an inverse transform starts at the right value.
Distinct real poles: one constant per pole. To get A_i, multiply F(s) by (s + p_i) and substitute s = −p_i. Each term inverts to A_i e^(−p_i t).
Repeated pole: a double pole needs two terms. B_2 comes from the usual substitution; B_1 needs one derivative with respect to s before substituting. The pair inverts to B_2 te^(−pt) + B_1 e^(−pt).
Complex-conjugate poles: keep the quadratic, complete the square, and split the numerator into an (s + a) part and an ω part. The first part is the damped cosine row and the second is the damped sine row.
What the symbols mean
- time signal, taken as zero for t < 0
- Laplace transform of f(t)
- complex variable s = σ + jω, in s⁻¹
- Laplace transform operator
- unit step: 0 for t < 0 and 1 for t ≥ 0
- unit impulse at t = 0, in s⁻¹
- decay rate of an exponential, in s⁻¹; the pole is at s = −a
- oscillation frequency, in rad/s
- time delay, in s
- integration variable (a past time), in s
- non-negative integer power of t
- constants
- initial value and initial rate of change of f
- value of f just after t = 0
- poles of sF(s), in s⁻¹
- pole magnitudes: the i-th pole is at s = −p_i, in s⁻¹
- partial-fraction coefficient of the pole at −p_i
- coefficients of a double pole at s = −p
- numerator coefficients over a quadratic factor
Every transform here is one-sided: f(t) is taken as zero for t < 0, so each table entry is implicitly multiplied by the unit step 1(t). Expand in partial fractions only when F(s) is strictly proper, meaning the numerator degree is lower than the denominator degree; otherwise divide the polynomials first. The final-value theorem holds only if every pole of sF(s) lies in the open left half-plane.
Step response of a second-order system, from equation to y(t)
A system obeys the equation below and starts at rest. The input is a unit step. Input and output are in the same normalized units; time is in seconds.
Transform every term. The initial conditions are zero and the step transforms to 1/s. Solve for Y(s).
Find the poles. The quadratic has complex roots, so keep it in one piece and complete the square. The expansion needs one constant for the pole at 0 and a first-order numerator over the quadratic.
Get A by multiplying by s and setting s = 0. Then multiply both sides by the full denominator and match powers of s to get B and C.
Split the numerator s + 2 into (s + 1) + 1. The table needs ω = 2 in the sine numerator, so write the leftover 1 as one half times 2.
Invert term by term: a step, a damped cosine and a damped sine.
Check both ends. The poles of sY(s) are −1 ± j2, in the open left half-plane, so the final-value theorem applies. Both limits agree with the time function: it starts at 0 and settles at 1.
Read the response. Differentiating gives a slope that first returns to zero at t = π/2, so the output peaks there at about 1.21, roughly 21% above its final value. The chapter on second-order specifications turns this calculation into a formula.
Reference tables
Read the table from right to left to invert: expand F(s) into partial fractions until every term matches one row.
| Property | Time domain | s-domain |
|---|---|---|
| Linearity | ||
| Derivative | ||
| Second derivative | ||
| Integral | ||
| Shift in s | ||
| Time delay | ||
| Final value (only if every pole of sF(s) has a negative real part) | ||
| Initial value (F(s) strictly proper) |
More detail and model assumptions
Where the table rows come from
Every table row comes from the defining integral, and one row does most of the work. For f(t) = e^(−at), the integral of e^(−at)e^(−st) from 0 to ∞ equals 1/(s + a), provided Re(s) > −a so that the integrand decays. Setting a = 0 gives the unit step, 1/s. Setting a = ∓jω and combining the two results with Euler's formula gives the sine and cosine rows.
The s-shift rule produces the rest. Multiplying a signal by e^(−at) replaces s by s + a in its transform. The ramp row 1/s² therefore becomes te^(−at) ↔ 1/(s + a)², and the sine and cosine rows become the two damped rows. In practice you memorise the step, the ramp, sine and cosine, and let the shift rule generate the others.
Read the table in both directions, and read denominators as pole patterns. A factor (s + a) is a real pole at −a and inverts to e^(−at). A factor (s + a)² + ω² is a complex pair at −a ± jω: it inverts to e^(−at) times a sine and a cosine of ωt. The lower limit 0⁻ means that an impulse at t = 0 is included, and that the derivative rule uses the value just before the input acts.
Three pole types, three expansions
Distinct real poles get one constant each. To find the constant over (s + p), multiply F(s) by (s + p) and set s = −p; in effect you cover that factor up and evaluate what remains. For F(s) = 6/[s(s + 3)]: covering s gives 6/3 = 2, covering (s + 3) gives 6/(−3) = −2, so f(t) = 2 − 2e^(−3t).
A pole repeated twice needs two terms, one over (s + p)² and one over (s + p). The cover-up step still gives the coefficient of the highest power. Differentiate (s + p)²F(s) once before substituting s = −p to get the other. For F(s) = 1/[s(s + 1)²], (s + 1)²F = 1/s: its value at s = −1 is −1 and its derivative −1/s² is also −1 there, while the pole at 0 has coefficient 1. So F(s) = 1/s − 1/(s + 1) − 1/(s + 1)² and f(t) = 1 − e^(−t) − te^(−t).
For a complex pair, keep the quadratic factor in one piece with a numerator Bs + C. Find B and C by multiplying out and matching coefficients of equal powers of s. Then complete the square in the denominator and split the numerator into a multiple of (s + a) and a multiple of ω. Those two pieces are exactly the damped cosine and damped sine rows. Complex residues give the same answer, but the quadratic route needs only real arithmetic.
A differential equation with initial conditions, start to finish
Take ÿ + 3ẏ + 2y = 4·1(t) with y(0) = 0 and ẏ(0) = 1. Transform each term: [s²Y − s·0 − 1] + 3[sY − 0] + 2Y = 4/s. Collect Y: (s² + 3s + 2)Y = 4/s + 1 = (s + 4)/s, so Y(s) = (s + 4)/[s(s + 1)(s + 2)].
The poles 0, −1 and −2 are distinct, so use the cover-up rule three times. At s = 0: 4/(1·2) = 2. At s = −1: 3/[(−1)(1)] = −3. At s = −2: 2/[(−2)(−1)] = 1. Therefore y(t) = 2 − 3e^(−t) + e^(−2t) for t ≥ 0.
Check the answer. y(0) = 2 − 3 + 1 = 0 and ẏ(0) = 3 − 2 = 1, as given. The poles of sY(s) are −1 and −2, so the final-value theorem applies: y(∞) = 4/2 = 2, which agrees with the equation at rest, 2y = 4. The solution also splits into two parts: the zero-state response 2 − 4e^(−t) + 2e^(−2t) from the input alone, and the zero-input response e^(−t) − e^(−2t) from the initial velocity alone. The transfer function 1/(s² + 3s + 2) describes only the first part.
When the final-value theorem applies
The final-value theorem states that f(t) approaches the limit of sF(s) as s → 0, on one condition: every pole of sF(s) lies in the open left half-plane. Equivalently, F(s) may have a single pole at s = 0 and all its other poles must have negative real parts. The partial-fraction expansion shows why: each left-half-plane pole contributes a decaying term, and the single pole at the origin contributes the constant that remains.
Three examples show what goes wrong outside the condition. F(s) = 1/[s(s − 1)] has a pole at +1: the formula returns −1, yet f(t) = e^t − 1 grows without bound. F(s) = 3/(s² + 9) has poles at ±j3 on the imaginary axis: the formula returns 0, yet f(t) = sin 3t oscillates forever. F(s) = 1/s² has a double pole at the origin: f(t) = t is a ramp and has no final value.
In control work the theorem is applied to the error E(s) to get the steady-state error, so confirm that the closed loop is stable before you use it. The initial-value theorem, f(0⁺) = lim sF(s) as s → ∞, needs only a strictly proper F(s). Use it as a check: your inverse transform evaluated at t = 0 must equal that limit.
Connect the example to the lab
Open the mass–spring–damper lab's release experiment: m = 1 kg, c = 1.6 N·s/m, k = 16 N/m, x(0) = 0.25 m, no applied force. Transforming ẍ + 1.6ẋ + 16x = 0 with these initial conditions gives X(s) = 0.25(s + 1.6)/(s² + 1.6s + 16), with poles at −0.8 ± j3.92. Completing the square yields x(t) = 0.25e^(−0.8t)[cos 3.92t + 0.204 sin 3.92t] m. This predicts the first trough at t ≈ 0.80 s with x ≈ −0.132 m. Drag the replay slider to 0.80 s and compare the displacement reading.
Check an inverse transform against a released massMake a prediction before moving a control.
- PREDICTBefore running anything, solve the release problem by Laplace transform: m = 1 kg, c = 1.6 N·s/m, k = 16 N/m, x(0) = 0.25 m, ẋ(0) = 0. Where are the poles, and what is x at t = 0.80 s?
- CHANGE ONE THINGRun the baseline and pin it. Then change only the damping to c = 8 N·s/m and redo the inverse transform: X(s) = 0.25(s + 8)/(s + 4)² now has a repeated pole at −4.
- OBSERVERead the displacement at 0.80 s for c = 1.6 and at 0.50 s for c = 8. The formulas predict −0.132 m and 0.102 m. Note that the c = 8 trace never crosses zero.
- EXPLAINExplain both traces from their poles. The pair −0.8 ± j3.92 inverts to a damped cosine and sine; the double pole at −4 inverts to 0.25(1 + 4t)e^(−4t), a polynomial times an exponential with no oscillation.
The final-value theorem always returns a number, even when no final value exists. For Y(s) = 1/[s(s − 1)] it gives −1, yet y(t) = e^t − 1 grows without bound. Find the poles of sY(s) first; use the limit only when all of them are in the open left half-plane.
Check your understanding+
F(s) = 10/[s(s + 2)(s + 5)] and G(s) = 3/(s² + 9). For which one may you use the final-value theorem, and what is the final value?
Only F. The poles of sF(s) are −2 and −5, so f(∞) = 10/(2·5) = 1. The poles of sG(s) are ±j3 on the imaginary axis: g(t) = sin 3t oscillates forever and has no final value, although the formula would return 0.Solve these before opening the solutions
Work each problem on paper first. Open the answer to check yours, then the full solution if you need the steps, and mark how it went.
Use the table and linearity to find the Laplace transform of f(t). Then list its poles.
Show answer
AnswerThree table rows, one per term. The poles are 0, −2 and −1 ± j3.
Show full solution
Solution
Transform the constant and the exponential with the step and exponential rows.
The last term is a damped sine with a = 1 and ω = 3.
Add the three results. Each denominator gives its poles directly.
Find f(t) by partial fractions, and confirm f(0+) with the initial-value theorem.
Show answer
AnswerThree distinct real poles give three exponentials. The response starts at 0.
Show full solution
Solution
The poles −1, −2 and −3 are distinct, so write one constant over each factor.
Cover each factor in turn and substitute its pole.
Invert each term with the exponential row.
Check with the initial-value theorem. The numerator of sF(s) has degree 2 and the denominator degree 3, so the limit is 0.
F(s) has a repeated pole. Find f(t), then find the final value and justify the method you use.
Show answer
AnswerThe double pole at −2 gives a te^(−2t) term. The final value is 2.
Show full solution
Solution
Write one term for the pole at 0 and two terms for the double pole at −2.
A and B_2 come from direct substitution.
B_1 needs one derivative of (s + 2)²F(s) = 8/s before substituting.
Invert with the step, te^(−at) and e^(−at) rows.
sF(s) has only the double pole at −2, in the open left half-plane, so the final-value theorem applies.
Solve the differential equation for t ≥ 0. Both the input and the initial conditions are nonzero.
Show answer
AnswerThe poles are 0 and −2 ± j3. The response starts at 2 and settles at 1.
Show full solution
Solution
Transform each term with the derivative rule, keeping the initial conditions.
Collect Y(s) on the left and put the right side over a common denominator.
Expand with a constant over s and Bs + C over the quadratic. Cover-up gives A; matching powers of s gives B and C.
Complete the square: the poles are −2 ± j3. Split s + 5 into (s + 2) + 3.
Invert, then check the initial conditions: y(0) = 1 + 1 = 2 and ẏ(0) = −2 + 3 = 1.
A unit ramp r(t) = t drives the system G(s) from rest. (a) Find y(t). (b) Find the steady-state error of e(t) = r(t) − y(t) with the final-value theorem and justify its use. (c) Explain why the theorem cannot be applied to Y(s) itself.
Show answer
AnswerThe output follows the ramp with a constant lag of 4/3. The theorem is valid for E(s) because sE(s) has poles only at −1 and −3; sY(s) has a pole at the origin.
Show full solution
Solution
Form Y(s) = G(s)R(s). It has a double pole at 0 and simple poles at −1 and −3.
For the double pole, substitute into s²Y(s) to get A_2 and differentiate it once to get A_1.
The simple poles use the cover-up rule.
Invert. Check: y(0) = −4/3 + 3/2 − 1/6 = 0.
Form E(s) = R(s) − Y(s). The poles of sE(s) are −1 and −3, so the final-value theorem applies and gives 4/3. sY(s) = 3/[s(s + 1)(s + 3)] still has a pole at s = 0: y(t) grows like t and has no final value.
Mark each exercise after you check your answer. Marks stay in this browser.
Terms this lecture teaches
- Laplace transform라플라스 변환
- final value theorem최종값 정리
- partial-fraction expansion부분분수 전개
- initial value theorem초기값 정리
- unit step단위 계단
- mass–spring–damper질량–스프링–댐퍼
- inverse Laplace transform라플라스 역변환
- impulse response임펄스 응답
- mode모드
Frequently asked questions
How do I find partial-fraction coefficients quickly?
For a distinct real pole at −p, multiply F(s) by (s + p) and substitute s = −p; this is the cover-up rule. For F(s) = 6/[s(s + 3)] it gives 2 and −2, so f(t) = 2 − 2e^(−3t). Repeated and complex poles need the extra steps for those cases.
Why does the final-value theorem sometimes give a wrong answer?
The limit of sF(s) as s → 0 always returns a number, even when f(t) has no final value. The theorem is valid only if every pole of sF(s) has a negative real part. For F(s) = 1/[s(s − 1)] the limit is −1, yet f(t) = e^t − 1 grows without bound.
How do I invert a term with complex-conjugate poles?
Keep the quadratic factor whole, complete the square as (s + a)² + ω², and split the numerator into a multiple of (s + a) and a multiple of ω. The two pieces invert to e^(−at) cos ωt and e^(−at) sin ωt, so only real arithmetic is needed.
What if the numerator degree is not lower than the denominator degree?
Divide the polynomials first. Partial-fraction expansion applies only to a strictly proper F(s); the polynomial quotient inverts to an impulse and its derivatives, and only the proper remainder is expanded.
Why the next lesson follows
You can now take any rational Y(s) back to y(t). The next chapter studies the ratio Y(s)/U(s) itself, the transfer function, and shows how much of the response its poles and zeros reveal before you invert anything.
Further reading
Finished the reading and the exercises? Mark the lecture as done to tick it on the syllabus. Progress stays in this browser.
EIGENROOM · https://eigenroom.com/theory/laplace-inverse · Updated 2026-11-15