Second Order Overdamped Response

When the damping ratio ζ exceeds 1, a second-order system enters the overdamped regime. The step response no longer oscillates. Instead, it rises monotonically and approaches its final value as a sum of two decaying exponentials. While overdamped systems are slower than the critically damped case, they are used in applications where any overshoot is strictly prohibited. Understanding their mathematical structure and response shape is essential for GATE analysis.

Core Concept: Why Overdamped Systems Do Not Oscillate

In a standard second-order system G(s) = ωn²/(s²+2ζωns+ωn²), the poles are given by s = -ζωn ± ωn√(ζ²-1). When ζ > 1, the term ζ²-1 is positive, so the square root is real. This means both poles are real and negative. Without an imaginary component, there is no oscillatory term in the time response. The step response is therefore a sum of two decaying exponentials, and it approaches the final value smoothly from below without ever crossing or overshooting it.

Let the two poles be s₁ = -ζωn + ωn√(ζ²-1) and s₂ = -ζωn - ωn√(ζ²-1). Since ζ > 1, we have s₂ more negative than s₁. This means the pole s₁ is closer to the imaginary axis. The pole closer to the origin, s₁, has a larger time constant τ₁ = 1/|s₁| and contributes the dominant slow term to the response. The pole s₂ is farther from the origin and decays much faster, contributing only in the early transient.

Mathematical Expression: Sum of Exponentials

For a unit step input R(s) = 1/s, the output in the Laplace domain is C(s) = ωn²/[s(s-s₁)(s-s₂)]. Performing partial fraction expansion gives C(s) = 1/s + A₁/(s-s₁) + A₂/(s-s₂), where A₁ and A₂ are residues. In the time domain: c(t) = 1 + A₁·e^(s₁t) + A₂·e^(s₂t). Since both s₁ and s₂ are negative, both exponentials decay to zero, and c(t) asymptotically approaches 1. The coefficients A₁ and A₂ are negative for a step input, meaning the response starts at zero and rises monotonically.

The long-term behavior is dominated by the term with the smaller magnitude of exponent, which is A₁·e^(s₁t). For large t, this term decays as e^(-t/τ₁) where τ₁ = 1/|s₁|. Therefore, the effective settling time of an overdamped system is approximately Ts ≈ 4τ₁ = 4/|s₁|. In design, this means it is always the slower pole that governs how long the system takes to settle, regardless of how fast the second pole is.

Dominant Pole Approximation

When the two poles are well separated, meaning |s₂| is significantly larger than |s₁| (typically by a factor of 5 or more), the fast pole decays so quickly that it effectively vanishes early in the transient. The system then behaves approximately like a first-order system with time constant τ₁ = 1/|s₁|. This is the dominant pole approximation and it is frequently used in GATE problems to simplify analysis. The approximation is valid when one pole is at least five times further from the origin than the other.

Practical Understanding

Overdamped systems are used in applications where zero overshoot is critical. Hard disk drive actuator control, precision optical lens positioning, and some biomedical devices such as infusion pump flow controllers require strictly monotonic responses. In these cases, designers intentionally choose ζ slightly above 1 (often ζ = 1.2 to 1.5) to ensure no overshoot while not making the system excessively slow. Beyond ζ = 2 to 3, the response becomes very sluggish and is usually unacceptable for real-time control.

Compared to the critically damped case (ζ = 1), the overdamped response is always slower. This is because the critically damped case has a repeated pole at -ωn, which is the optimal placement for the fastest non-oscillatory response. As ζ increases beyond 1, the dominant pole s₁ moves toward the origin while s₂ moves further away, and the dominant time constant τ₁ grows larger, making settling take longer.

Given:
System T(s) = 16 / (s² + 10s + 16)
Input: Unit step

Why this formula applies:
Find poles to verify overdamped behavior, then compute time constants

Formula:
s = (-b ± √(b²-4ac)) / 2a for s²+10s+16=0
Time constants τ₁=1/|s₁|, τ₂=1/|s₂|
Settling time Ts ≈ 4τ₁

Substitution:
Discriminant = 10²- 4×16 = 100 - 64 = 36 > 0  (overdamped confirmed)
s₁ = (-10+6)/2 = -2
s₂ = (-10-6)/2 = -8

Calculation:
τ₁ = 1/2 = 0.5 s  (dominant, slower pole)
τ₂ = 1/8 = 0.125 s (faster pole)
ωn = √16 = 4 rad/s
ζ = 10/(2×4) = 1.25  (confirms ζ > 1, overdamped)
Ts ≈ 4 × 0.5 = 2 s

Final Answer:
Poles at s=-2 and s=-8, ζ=1.25, Dominant τ=0.5 s, Settling time ≈ 2 s, No overshoot

Exam Tip: In GATE, if the characteristic equation has two distinct real negative roots, the system is overdamped. Compute both roots, identify the dominant pole as the one closer to the origin (smaller magnitude), and estimate settling time as 4 divided by the magnitude of that dominant pole. Do not use the %OS or Tp formulas — those apply only to underdamped systems.

Comparison with Other Damping Cases

• Overdamped (ζ > 1): Two distinct real negative poles. Response is sum of two decaying exponentials. No overshoot.
• Critically damped (ζ = 1): Repeated real poles at -ωn. Fastest monotonic response. Still no overshoot.
• Underdamped (0 < ζ < 1): Complex conjugate poles. Oscillatory decaying response with overshoot.
• Dominant pole approximation: Valid when |s₂| ≥ 5|s₁|. System then behaves like first-order with τ = 1/|s₁|.
• Response speed ranking for same ωn: Critically damped > Overdamped (increases with ζ) > Very overdamped.

Quick Revision

• Overdamped condition: ζ > 1. Discriminant of characteristic equation is positive. Two distinct real negative poles.
• Step response: c(t) = 1 + A₁·e^(s₁t) + A₂·e^(s₂t). Both exponentials decay. No oscillation, no overshoot.
• Dominant pole: closer to origin, larger time constant τ₁ = 1/|s₁|. Governs settling time.
• Settling time estimate: Ts ≈ 4/|s₁| (2% criterion based on dominant pole).
• As ζ increases beyond 1, system gets slower. Critically damped (ζ=1) is always faster than any overdamped case at same ωn.
• Trap: Never apply %OS = exp(-πζ/√(1-ζ²)) to overdamped systems. The formula is invalid for ζ ≥ 1.
• Trap: Dominant pole assumption requires |s₂| to be at least 5 times larger than |s₁|. Otherwise both poles contribute significantly.