A topic in the Open Knowledge Graph — a free, open map of 15,290 topics and the order to learn them in.
Poles in the left-half plane (LHP) contribute stable exponentially decaying terms; right-half plane (RHP) poles are unstable. Pole location (real part controls decay rate; imaginary part controls frequency) directly determines time response; frequency response magnitude has peaks near poles and nulls near zeros.
Every pole in a transfer function corresponds to a natural mode of the system — a way the system "wants" to behave on its own, without being driven. You already know from pole-zero plots that a pole at s = σ + jω generates a time-domain term of the form eσtcos(ωt). The sign of σ is everything: if σ < 0 (left-half plane), the exponential decays and the mode dies out — the system is stable. If σ > 0 (right-half plane), the mode grows without bound — the system is unstable. This is the geometric interpretation of stability: LHP poles are stable, RHP poles are unstable, and poles on the imaginary axis produce sustained oscillations.
The real part σ of a pole controls *how fast* the associated mode decays. A pole at s = −10 decays ten times faster than one at s = −1; equivalently, the time constant τ = 1/|σ| is ten times shorter. The imaginary part jω controls the *oscillation frequency* of the mode. A complex conjugate pole pair at s = −1 ± j5 produces a damped sinusoid oscillating at 5 rad/s that fades away with time constant 1 second. Poles close to the imaginary axis — small |σ| — are sluggish and poorly damped; they produce slow, ringing responses. Poles far into the left half plane decay so quickly they barely register in the transient.
Zeros play the complementary role in frequency response. Near a zero, the numerator of H(jω) goes toward zero, so the output is suppressed at that frequency regardless of input magnitude. Near a pole, the denominator of H(jω) is small, so the output is amplified — you see a peak in the Bode magnitude plot. A system with a pole at s = −ω₀ on the negative real axis has its frequency response peak at DC and rolls off toward ω₀, the corner frequency. RHP zeros are especially important: they cause the phase to drop (non-minimum phase behavior) even as the magnitude may appear normal, which creates fundamental limits on how aggressively a feedback controller can act.
The power of the pole-zero picture is that it lets you read off qualitative system behavior at a glance, without computing the full inverse Laplace transform. Count poles in the RHP for instability. Look at pole proximity to the imaginary axis for damping. Look at the real part magnitude for speed of response. Identify zero locations for frequency-response nulls. When you proceed to root locus and Nyquist methods, you will be manipulating these pole and zero locations deliberately — using feedback to move poles from unstable or poorly damped positions into the LHP where you want them.