L1.1 Write L{y′′−4y′+3y} in terms of Y=Y(s) with general initial conditions y(0)=y0, y′(0)=y1.
Recall Solution L1.1
Transform term by term. What we do: replace each derivative by its rule. Why: the transform is linear, so it passes through + and constants.
L{y′′}=s2Y−sy0−y1,L{−4y′}=−4(sY−y0),L{3y}=3Y.
Add:
L{y′′−4y′+3y}=(s2−4s+3)Y−sy0−y1+4y0.
L1.2 State L{u(t−3)} and L{δ(t−3)}.
Recall Solution L1.2
Special case g=1 of the shift rule: L{u(t−3)}=se−3s.
The impulse: L{δ(t−3)}=e−3s.
Why the difference? The step is "on forever after t=3" (a lasting s1), the impulse is "one instant only" (no s1).
L1.3 Invert s2+92s and s2+94.
Recall Solution L1.3
Match s2+ω2s↔cosωt with ω=3: s2+92s→2cos3t.
Match s2+ω2ω↔sinωt: we need ω=3 on top, but we have 4. Write s2+94=34⋅s2+93→34sin3t.
Transform:sY+2Y=se−s, so Y=s(s+2)e−s.
Invert the un-delayed core first. Let H(s)=s(s+2)1. Cover-up: at s=0 residue 21, at s=−2 residue −21. So H=s1/2−s+21/2 and h(t)=21−21e−2t.
Apply the delay tag e−s: delay h by 1 AND gate it:
y(t)=u(t−1)[21−21e−2(t−1)].
Meaning: nothing until t=1, then y rises toward 21. See the figure.
L3.2 Find L{tu(t−2)}.
Recall Solution L3.2
The trick: the shift rule needs the forcing written as g(t−2), a function of (t−2). But we have plain t. Rewrite:t=(t−2)+2. So g(t−2)=(t−2)+2, i.e. g(τ)=τ+2.
G(s)=L{τ+2}=s21+s2. Multiply by the delay tag:
L{tu(t−2)}=e−2s(s21+s2).
L3.3 A rectangular pulse: solve y′+y=u(t−1)−u(t−3), y(0)=0. (Input is ON only between t=1 and t=3.)
Recall Solution L3.3
Transform:sY+Y=se−s−e−3s, so Y=s(s+1)e−s−e−3s.
Core inverse:H(s)=s(s+1)1=s1−s+11⇒h(t)=1−e−t.
Two delay tags → two gated copies (subtract the second):y(t)=u(t−1)[1−e−(t−1)]−u(t−3)[1−e−(t−3)].
Meaning: flat 0 until t=1; charges up toward 1; at t=3 the input switches off and the second term makes it relax back toward 0. See the figure.
L4.1 Solve y′′+3y′+2y=u(t−1), y(0)=0, y′(0)=0, then evaluate y(2).
Recall Solution L4.1
Transform:(s2+3s+2)Y=se−s, so Y=s(s+1)(s+2)e−s.
Core inverse:H(s)=s(s+1)(s+2)1. Cover-up: at s=0: (1)(2)1=21; at s=−1: (−1)(1)1=−1; at s=−2: (−2)(−1)1=21.
H=s1/2−s+11+s+21/2⇒h(t)=21−e−t+21e−2t.Delay + gate:y(t)=u(t−1)[21−e−(t−1)+21e−2(t−1)].
Evaluate at t=2 (so t−1=1): y(2)=21−e−1+21e−2≈0.5−0.367879+0.067668=0.199789.
L4.2 Solve y′′+y=sint, y(0)=0, y′(0)=0 (resonance — the forcing matches the natural frequency).
Recall Solution L4.2
Transform:(s2+1)Y=s2+11, so Y=(s2+1)21.
Why is this special? The forcing sint has the same frequency as the natural oscillation, so its transform shares the same denominator — squaring it. A repeated quadratic factor signals resonance, growth in t.
Invert using the known pair (s2+ω2)21↔2ω31(sinωt−ωtcosωt), with ω=1:
y(t)=21(sint−tcost).
The −21tcost term grows without bound — the amplitude climbs linearly, the fingerprint of resonance.
Check: y(0)=0; y′(t)=21(tsint) (after simplification) so y′(0)=0 ✔.
Transform:(s2+4)Y=se−πs, so Y=s(s2+4)e−πs.
Core inverse:H(s)=s(s2+4)1=sA+s2+4Bs+C. Multiply out: 1=A(s2+4)+(Bs+C)s. At s=0: A=41. Coeff of s2: A+B=0⇒B=−41. Coeff of s: C=0.
H=s1/4−41⋅s2+4s⇒h(t)=41−41cos2t.Delay + gate:y(t)=u(t−π)[41−41cos2(t−π)].
At t=23π:t−π=2π, cos2⋅2π=cosπ=−1. So y=41−41(−1)=21.
L5.1 Solve y′′+2y′+y=δ(t−2), y(0)=0, y′(0)=0. (A single hammer blow on a critically-damped system.)
Recall Solution L5.1
Transform:(s2+2s+1)Y=e−2s. Note s2+2s+1=(s+1)2. So Y=(s+1)2e−2s.
Core inverse:(s+1)21↔te−t (from (s−a)n+1n!↔tneat with n=1,a=−1). So h(t)=te−t.
Delay + gate:y(t)=u(t−2)(t−2)e−(t−2).
Meaning: nothing until the kick at t=2, then a bump that rises and decays back to 0 — the system's impulse response, delayed. Sanity: at t=3, y=1⋅e−1≈0.367879.
L5.2 Degenerate check: solve y′=δ(t), y(0)=0. What does the answer say physically, and why is the tiny t=0 subtlety worth noticing?
Recall Solution L5.2
Transform:sY−0=e−0⋅s=1, so Y=s1 and y(t)=1 for t>0.
Physical meaning: a unit impulse at t=0 instantly deposits one unit; afterward y is stuck at 1 (no restoring term). The impulse acts as an instantaneous "step-in" of the initial condition.
The subtlety: the impulse sits exactly at the lower limit t=0. With the standard one-sided transform the kick is fully captured, giving y=1 for all t>0; right at t=0 the value is undefined (it is jumping). This is the boundary/degenerate case where "just after 0" and "just before 0" differ — always ask which side you mean.
L5.3 Reasoning (no full inversion needed): for y′′+5y′+6y=f(t) with zero initial conditions, the transform factors as Y(s)=(s+2)(s+3)F(s). Predict the long-time behaviour when f(t)=u(t) (a constant switch-on), using only the roots.
Recall Solution L5.3
Read the poles. The system's own decay modes come from the roots s=−2 and s=−3, both negative → both give terms e−2t,e−3t that die out. With f=u(t), F=s1 adds a pole at s=0, whose residue gives a constant term.
Compute the steady state directly by the value the constant term takes: as t→∞ the decaying modes vanish, leaving the s=0 residue of s(s+2)(s+3)1, which is (2)(3)1=61.
Conclusion:y(t)→61. You never had to fully invert — the poles told the story. (Cross-check with the ODE: at steady state y′′=y′=0, so 6y=1⇒y=61 ✔.)
How do you invert e−asH(s)? ::: Delay AND gate: u(t−a)h(t−a).
Why rewrite t=(t−2)+2 before transforming tu(t−2)? ::: The shift rule needs the forcing as a function of (t−2); then G(s)=s21+s2 and you tag with e−2s.
What denominator signals resonance? ::: A repeated quadratic like (s2+ω2)2; it yields a tcosωt growing term.
L{δ(t−a)} vs L{u(t−a)}? ::: e−as vs se−as — impulse has no s1.
Steady state of s(s+2)(s+3)1 as t→∞? ::: The s=0 residue 61 (decaying modes vanish).