Complex Poles and Real Time-Domain Signals
A worked explanation of complex poles and real time-domain signals, including the key limitation and a checkable example.
Complex Poles and Real Time-Domain Signals
A student solves for the step response of a second-order system and gets back an expression full of complex numbers. That expression should be a voltage or a position, not an algebra exercise. The reason it becomes real is that conjugate poles, when combined through the inverse Laplace transform, produce real sine and cosine terms. Exactly one worked example you can recompute yourself shows how complex poles and real time-domain signals connect.
Why the Method Works
When a real-coefficient rational function F(s) has complex poles, they always occur as a conjugate pair: p = a + jb and p* = a − jb. The partial-fraction decomposition for that pair takes the form A/(s − p) + A*/(s − p*), where A and A* are themselves complex conjugates. This structure is guaranteed when F(s) has real coefficients, which every physically meaningful transfer function does.
The inverse Laplace transform of that sum is not two separate complex exponentials. It combines into a single real expression: 2eat[Re(A) cos(bt) − Im(A) sin(bt)]. The exponential factor eat carries the decay or growth rate from the real part of the pole. The sine and cosine terms carry the oscillation frequency from the imaginary part. The coefficients Re(A) and Im(A) come from the residue of the complex pole.
This formula is derived from Euler's formula and the frequency-shift theorem. It is the standard result found in Kreyszig 'Advanced Engineering Mathematics' and Boyce & DiPrima 'Elementary Differential Equations'.
A Worked Example
Work through F(s) = (s+1)/(s² + 2s + 5). The poles are the roots of the denominator: s = [−2 ± √(4 − 20)]/2 = −1 ± 2j. So a = −1 s⁻¹ and b = 2 rad/s.
Set up the partial-fraction form: (s+1)/(s²+2s+5) = A/(s+1−2j) + A*/(s+1+2j). Solve for A by multiplying both sides by (s+1−2j) and evaluating at s = −1+2j:
A = (s+1)/(s+1+2j) evaluated at s = −1+2j = (2j)/(4j) = 0.5.
The coefficient is real, so Re(A) = 0.5 and Im(A) = 0. Apply the combination formula: ℒ⁻¹{F(s)} = 2e−t[0.5 cos(2t) − 0 sin(2t)] = e−t cos(2t).
The time-domain signal is a cosine wave decaying at a rate determined by the real part of the pole. No complex numbers remain. You can verify this result using a complex number calculator to check the residue, then confirm against any standard Laplace transform table.
Calculation Steps
To perform the calculation on any conjugate pair, follow these steps:
- Factor the denominator to find a and b.
- Write the partial-fraction form with conjugate coefficients.
- Solve for A using the Heaviside cover-up method for complex poles: multiply by (s−p) and evaluate at s = p.
- Extract Re(A) and Im(A).
- Plug into 2eat[Re(A) cos(bt) − Im(A) sin(bt)].
For the worked example, the arithmetic is clean because the numerator and denominator degree differ by one. If the numerator degree equals the denominator degree, perform polynomial division first. If the poles are repeated, the partial-fraction form requires additional terms with higher powers of 1/(s−p).
| Quantity | Symbol | Unit | Role |
|---|---|---|---|
| Real part of pole | a | s⁻¹ | Exponential decay rate |
| Imaginary part of pole | b | rad/s | Oscillation frequency |
| Residue magnitude | |A| | Depends on F(s) | Amplitude of sine/cosine terms |
| Phase angle of residue | arg(A) | rad | Determines mix of sine and cosine |
| Exponential factor | e<sup>at</sup> | dimensionless | Decay or growth envelope |
Failure Case: Non-Conjugate Coefficients
The most common failure in this calculation is an arithmetic error that produces coefficients that are not complex conjugates. Suppose you solve for A and get 0.5 + 0.2j, but for A* you get 0.5 − 0.3j. The imaginary parts differ. Plugging these into the combination formula gives a time-domain term with a leftover imaginary component, which cannot represent a physical voltage, position, or current.
This happens when you misapply the Heaviside cover-up method or when the partial-fraction form is wrong. Always verify that the imaginary parts of A and A* are equal in magnitude and opposite in sign. If they are not, go back and re-solve the system of equations. The combination formula only works when the coefficients are true conjugates.
Another limit: the method applies only to distinct complex conjugate poles. If a pole pair is repeated, the partial-fraction expansion requires terms with denominators (s−p)² and (s−p*)², and the inverse produces terms like t·eatcos(bt). The formula changes and requires a separate derivation from the frequency-shift and derivative theorems.
Common Questions
Why do the coefficients have to be complex conjugates?
Because F(s) has real coefficients. The partial-fraction sum must be real when s is real. That forces A and A* to be conjugates. If they are not, the sum produces a complex time-domain result, which is physically meaningless.
What if the imaginary part b is zero?
Then the poles are real and distinct. No sine or cosine terms appear. The inverse is a sum of real exponentials, handled with ordinary partial fractions. The complex-pole method only applies when b ≠ 0.
Do I need to compute the Bromwich integral?
No. The combination formula replaces the contour integral. Engineers and students use the formula and the partial-fraction decomposition. The Bromwich integral is the theoretical definition but almost never computed in practice.
What if the residue A is pure imaginary?
Then Re(A) = 0 and the result is a pure sine term multiplied by e<sup>at</sup>. This happens when the numerator of F(s) is an odd function of s about the pole's real part. The formula still produces a real signal.