Formula sheet

Control Systems Formula Sheet

When tuning a PID controller for a DC motor speed loop or analyzing gain margin in a GATE 2024 frequency-response question, every formula from characteristic equation roots to phase margin calculation matters immediately. These formulas cover transfer function modeling, time-domain response, stability criteria (Routh-Hurwitz, Nyquist, Bode), root locus, state-space representation, and controller design — the full scope tested in GATE EEE/ECE, ESE, and BITS Pilani.

EEE, ECE

Transfer Function and Block Diagram Reduction

Closed-Loop Transfer Function

T(s) = \frac{G(s)}{1 + G(s)H(s)}

SymbolDescriptionUnit
T(s)Closed-loop transfer functiondimensionless
G(s)Forward path transfer functiondimensionless
H(s)Feedback transfer functiondimensionless

Worked example

Find T(s) for G(s) = 10/(s+2), H(s) = 1 (unity feedback).

Given: G(s) = 10/(s+2), H(s) = 1

  1. T(s) = G(s) / (1 + G(s)·H(s))
  2. = [10/(s+2)] / [1 + 10/(s+2)]
  3. = [10/(s+2)] / [(s+2+10)/(s+2)]
  4. = 10 / (s+12)

Answer: T(s) = 10/(s+12)

Characteristic Equation

1 + G(s)H(s) = 0

Worked example

Write characteristic equation for G(s) = K/(s(s+4)), H(s) = 1.

Given: G(s)H(s) = K/(s(s+4))

  1. 1 + K/[s(s+4)] = 0
  2. s(s+4) + K = 0
  3. s² + 4s + K = 0

Answer: Characteristic equation: s² + 4s + K = 0

Signal Flow Graph — Mason's Gain Formula

T = \frac{\sum_k P_k \Delta_k}{\Delta}

SymbolDescriptionUnit
P_kk-th forward path gaindimensionless
\DeltaGraph determinantdimensionless
\Delta_kCofactor of k-th pathdimensionless

Worked example

SFG with one forward path P1 = G1·G2·G3, one loop L1 = G2·H1. Find T.

Given: P1 = G1G2G3, L1 = G2H1

  1. Δ = 1 - L1 = 1 - G2H1
  2. Δ1 = 1 (loop L1 touches forward path)
  3. T = P1·Δ1 / Δ = G1G2G3 / (1 - G2H1)

Answer: T = G1G2G3 / (1 - G2H1)

Time-Domain Response Specifications

Second-Order System Standard Form

T(s) = \frac{\omega_n^2}{s^2 + 2\zeta\omega_n s + \omega_n^2}

SymbolDescriptionUnit
\omega_nNatural frequencyrad/s
\zetaDamping ratiodimensionless

Worked example

T(s) = 25/(s²+6s+25). Find ωn and ζ.

Given: T(s) = 25/(s²+6s+25)

  1. ωn² = 25 → ωn = 5 rad/s
  2. 2ζωn = 6 → ζ = 6/(2×5) = 0.6

Answer: ωn = 5 rad/s, ζ = 0.6 (underdamped)

Peak Overshoot

M_p = e^{-\pi\zeta / \sqrt{1-\zeta^2}} \times 100\%

SymbolDescriptionUnit
M_pPercent peak overshoot%

Worked example

Find peak overshoot for ζ = 0.6.

Given: ζ = 0.6

  1. √(1-ζ²) = √(1-0.36) = √0.64 = 0.8
  2. Exponent = -π × 0.6 / 0.8 = -1.8850 / 0.8 = -2.356
  3. M_p = e^{-2.356} × 100 = 0.0948 × 100

Answer: M_p ≈ 9.48%

Settling Time (2% criterion)

t_s \approx \frac{4}{\zeta \omega_n}

SymbolDescriptionUnit
t_sSettling times

Worked example

Find settling time for ζ = 0.6, ωn = 5 rad/s.

Given: ζ = 0.6, ωn = 5 rad/s

  1. t_s = 4 / (ζ × ωn)
  2. = 4 / (0.6 × 5)
  3. = 4 / 3
  4. = 1.33 s

Answer: t_s ≈ 1.33 s

Rise Time (0% to 100%)

t_r = \frac{\pi - \phi}{\omega_d}, \quad \phi = \cos^{-1}\zeta, \quad \omega_d = \omega_n\sqrt{1-\zeta^2}

SymbolDescriptionUnit
t_rRise times
\omega_dDamped natural frequencyrad/s
\phiPhase anglerad

Worked example

Find rise time for ζ = 0.6, ωn = 5 rad/s.

Given: ζ = 0.6, ωn = 5

  1. ωd = 5 × √(1-0.36) = 5 × 0.8 = 4 rad/s
  2. φ = cos⁻¹(0.6) = 0.9273 rad
  3. t_r = (π - 0.9273) / 4
  4. = 2.2143 / 4 = 0.5536 s

Answer: t_r ≈ 0.55 s

Stability — Routh-Hurwitz Criterion

Routh Array Construction

\text{All roots in LHP} \Leftrightarrow \text{all elements of first column} > 0

Worked example

Check stability of s³ + 6s² + 11s + 6 = 0.

Given: Coefficients: a3=1, a2=6, a1=11, a0=6

  1. Row s³: 1 11
  2. Row s²: 6 6
  3. Row s¹: b1 = (6×11 - 1×6)/6 = (66-6)/6 = 10
  4. Row s⁰: c1 = (10×6 - 6×0)/10 = 6
  5. First column: 1, 6, 10, 6 — all positive

Answer: System is stable (all first-column elements positive)

Routh Marginal Stability (Row of Zeros)

\text{If row of zeros appears, form auxiliary polynomial from row above}

Worked example

s⁴+s³+3s²+3s+2=0. Identify if purely imaginary roots exist.

Given: Row s²: 3·1 - 1·3 / 1 ... (check for zero row)

  1. Row s⁴: 1 3 2
  2. Row s³: 1 3
  3. Row s²: (3×1 - 1×3)/1 = 0 → auxiliary from s³ row
  4. Auxiliary: s³ + 3s = 0 → s(s²+3) = 0 → s = ±j√3
  5. Replace s² row using d/ds(s³+3s) = 3s²+3 → row: 3 3

Answer: System has purely imaginary roots at s = ±j√3 (marginally stable)

Frequency Response — Bode Plots

Gain Margin

GM = -20\log|G(j\omega_{pc})H(j\omega_{pc})| \text{ dB}

SymbolDescriptionUnit
GMGain margindB
\omega_{pc}Phase crossover frequency (∠GH = -180°)rad/s

Worked example

At ω_pc = 2 rad/s, |G(j2)H(j2)| = 0.5. Find gain margin.

Given: |GH| at ω_pc = 0.5

  1. GM = -20 log(0.5)
  2. = -20 × (-0.301)
  3. = 6.02 dB

Answer: GM = +6 dB (stable)

Phase Margin

PM = 180° + \angle G(j\omega_{gc})H(j\omega_{gc})

SymbolDescriptionUnit
PMPhase margindegrees
\omega_{gc}Gain crossover frequency (|GH| = 1)rad/s

Worked example

At ω_gc, ∠GH = -135°. Find phase margin.

Given: ∠GH at ω_gc = -135°

  1. PM = 180° + (-135°)
  2. PM = 45°

Answer: PM = 45° (well-damped, stable)

Bode Magnitude of First-Order Factor

|1 + j\omega/\omega_c|_{dB} \approx 0 \text{ dB for } \omega \ll \omega_c;\quad +20\log(\omega/\omega_c) \text{ for } \omega \gg \omega_c

SymbolDescriptionUnit
\omega_cCorner frequency = 1/τrad/s

Worked example

G(s) = 1/(s+10). Find Bode magnitude at ω = 1, 10, 100 rad/s.

Given: G(jω) = 1/(10+jω), corner at ωc = 10 rad/s

  1. At ω = 1: |G| = 1/√(100+1) ≈ 1/10.05 → -20.04 dB ≈ -20 dB
  2. At ω = 10 (corner): |G| = 1/√200 = 0.0707 → -23 dB
  3. At ω = 100: |G| = 1/√(100²+10²) ≈ 1/100 → -40 dB (slope -20 dB/dec)

Answer: Bode plot: flat at -20 dB below corner; rolls off at -20 dB/decade above ωc = 10

Root Locus

Angle Condition (Root Locus)

\angle G(s)H(s) = (2k+1) \times 180°, \quad k = 0, \pm1, \pm2, \ldots

Worked example

Check if s = -1+j1 is on root locus of G(s) = K/[s(s+2)].

Given: s = -1+j, G(s) = K/[s(s+2)]

  1. ∠(s) = angle of (-1+j) from origin = 135°
  2. ∠(s+2) = angle of (1+j) from -2 = 45°
  3. ∠G(s) = 0 - 135° - 45° = -180°
  4. -180° = (2×0+1)×(-180°) ✓

Answer: s = -1+j is on the root locus

Centroid of Asymptotes

\sigma_a = \frac{\sum \text{poles} - \sum \text{zeros}}{n - m}

SymbolDescriptionUnit
\sigma_aCentroid on real axisdimensionless
nNumber of polesdimensionless
mNumber of zerosdimensionless

Worked example

G(s) = K/[s(s+2)(s+4)]. Find centroid of asymptotes.

Given: Poles: 0, -2, -4; no zeros

  1. Σ poles = 0 + (-2) + (-4) = -6
  2. Σ zeros = 0
  3. n - m = 3 - 0 = 3
  4. σ_a = (-6 - 0) / 3 = -2

Answer: Centroid σ_a = -2

Angle of Asymptotes

\theta_a = \frac{(2k+1) \times 180°}{n - m}

Worked example

Find asymptote angles for G(s) = K/[s(s+2)(s+4)].

Given: n = 3, m = 0

  1. k=0: θ = 1×180°/3 = 60°
  2. k=1: θ = 3×180°/3 = 180°
  3. k=2: θ = 5×180°/3 = 300° (= -60°)

Answer: Asymptote angles: 60°, 180°, 300°

Steady-State Error

Steady-State Error (Unity Feedback)

e_{ss} = \frac{1}{1 + K_p} \text{ (step)};\quad \frac{1}{K_v} \text{ (ramp)};\quad \frac{1}{K_a} \text{ (parabolic)}

SymbolDescriptionUnit
K_pPosition error constant = lim_{s→0} G(s)dimensionless
K_vVelocity error constant = lim_{s→0} sG(s)1/s
K_aAcceleration constant = lim_{s→0} s²G(s)1/s²

Worked example

G(s) = 10/[s(s+2)]. Find steady-state error to unit ramp.

Given: G(s) = 10/[s(s+2)]

  1. K_v = lim_{s→0} s·G(s) = lim_{s→0} s·10/[s(s+2)]
  2. = lim_{s→0} 10/(s+2)
  3. = 10/2 = 5
  4. e_ss = 1/K_v = 1/5 = 0.2

Answer: e_ss = 0.2 for unit ramp input

System Type Number

\text{Type} = \text{number of pure integrators (poles at } s=0\text{) in } G(s)H(s)

Worked example

Identify type of G(s) = 5(s+3)/[s²(s+10)].

Given: G(s) = 5(s+3)/[s²(s+10)]

  1. Count poles at s = 0: s² contributes 2 poles at origin
  2. Type = 2

Answer: Type 2 system — zero steady-state error to ramp, finite error to parabolic

State-Space Representation

State-Space Equations

\dot{\mathbf{x}} = A\mathbf{x} + B\mathbf{u}, \quad \mathbf{y} = C\mathbf{x} + D\mathbf{u}

SymbolDescriptionUnit
ASystem matrix (n×n)dimensionless
BInput matrix (n×1)dimensionless
COutput matrix (1×n)dimensionless
DFeed-forward matrixdimensionless

Worked example

Write state-space for G(s) = 1/(s²+3s+2).

Given: Transfer function: 1/(s²+3s+2)

  1. Choose states x1 = y, x2 = dy/dt
  2. ẋ1 = x2
  3. ẋ2 = -2x1 - 3x2 + u
  4. A = [[0,1],[-2,-3]], B = [[0],[1]], C = [1,0], D = 0

Answer: A = [0 1; -2 -3], B = [0; 1], C = [1 0], D = 0

Characteristic Equation from State Matrix

\det(sI - A) = 0

Worked example

Find eigenvalues of A = [[0,1],[-2,-3]].

Given: A = [0 1; -2 -3]

  1. det(sI - A) = det([[s,-1],[2,s+3]])
  2. = s(s+3) - (-1)(2) = s² + 3s + 2
  3. s² + 3s + 2 = 0 → (s+1)(s+2) = 0

Answer: Eigenvalues: s = -1 and s = -2 (system is stable)

Quick reference

FormulaExpression
Closed-Loop TFT(s) = G(s)/(1+G(s)H(s))
Characteristic Equation1 + G(s)H(s) = 0
Peak OvershootMp = exp(-πζ/√(1-ζ²)) × 100%
Settling Time (2%)t_s = 4/(ζωn)
Rise Timet_r = (π - φ)/ωd
Gain MarginGM = -20 log|GH(jω_pc)| dB
Phase MarginPM = 180° + ∠GH(jω_gc)
Velocity Error ConstantKv = lim_{s→0} sG(s)
Centroid of Asymptotesσ_a = (Σpoles - Σzeros)/(n-m)
Asymptote Anglesθ_a = (2k+1)×180°/(n-m)
State Equationẋ = Ax + Bu
Eigenvaluesdet(sI - A) = 0
ωn from TFωn² = constant term in denominator
Damping Ratio from TFζ = (middle coeff)/(2ωn)
Mason's GainT = Σ(Pk Δk)/Δ

Exam tips

  • GATE examiners test Routh-Hurwitz by asking how many roots are in the RHP — count sign changes in the first column, not just whether all elements are positive.
  • For Bode plot gain margin, find the phase crossover frequency first (where phase = -180°), then read the magnitude — do not confuse ω_gc with ω_pc.
  • Phase margin of 45°–60° corresponds roughly to ζ = 0.45–0.6 — examiners often ask you to infer damping from frequency-domain specs.
  • In root locus, the breakaway point between two poles on the real axis satisfies dK/ds = 0 — always solve this before sketching the locus.
  • Steady-state error requires identifying system type first — Type 0 has infinite error to ramp, Type 1 has zero error to step, finite error to ramp.
  • For state-space controllability, verify rank of [B AB A²B ... A^{n-1}B] = n; for observability check rank of [C; CA; CA² ...; CA^{n-1}] = n.