Routh Array Special Case Zero Row

A second special case in Routh array construction occurs when an entire row of the array becomes all zeros. This situation cannot be resolved by the epsilon method. Instead, it signals a specific algebraic structure in the characteristic polynomial: the existence of poles that are symmetrically placed with respect to the origin in the s-plane. The auxiliary polynomial method is used to handle this case, and it provides not only stability information but also the exact location of the special poles causing the zero row.

Core Concept Explanation

An entire row of zeros in the Routh array appears because the characteristic polynomial D(s) has a factor that is an even polynomial, meaning it contains only even powers of s. Such a polynomial has roots that come in pairs symmetric about the origin of the s-plane. The possible symmetric root configurations are: purely imaginary pairs ±jω (poles on the imaginary axis), purely real symmetric pairs ±σ (one stable, one unstable), or complex quadrantal pairs σ ± jω and -σ ± jω.

When such a symmetric factor is present, the row corresponding to the order of that factor in the Routh array will become all zeros because the symmetric structure causes the determinant computations to cancel exactly. The Routh construction cannot continue from an all-zero row, so the auxiliary polynomial is formed from the row just above the zero row, and its derivative is used to replace the zero row and allow continuation.

Auxiliary Polynomial Method

The auxiliary polynomial A(s) is constructed from the row immediately above the zero row. If this row corresponds to the sᵏ row and has elements α, β, γ, then the auxiliary polynomial is A(s) = α·sᵏ + β·sᵏ⁻² + γ·sᵏ⁻⁴ + .... Note that the powers of s decrease by 2 in each term, consistent with an even or odd polynomial structure. The roots of A(s) are exactly the symmetric poles causing the zero row.

To replace the zero row, differentiate the auxiliary polynomial with respect to s: dA(s)/ds = k·α·sᵏ⁻¹ + (k-2)·β·sᵏ⁻³ + .... The coefficients of this derivative polynomial are used to fill the zero row. Normal Routh array construction then continues from that point. The stability analysis of the remaining rows above and below the filled row proceeds as usual, with sign changes in the first column indicating RHP poles.

Mathematical Expression and Imaginary Axis Poles

The roots of the auxiliary polynomial A(s) are always the poles that caused the zero row. If A(s) = αs² + β, setting A(s) = 0 gives s² = -β/α. If β/α > 0, the roots are s = ±j·sqrt(β/α), which are purely imaginary. These poles lie exactly on the imaginary axis, making the system marginally stable (assuming no other RHP poles exist from the rest of the array). The frequency of oscillation of such a marginally stable system is ω = sqrt(β/α) rad/s.

For design problems where a variable gain K is present, the zero row condition identifies the critical gain Kc at which the system transitions from stable to marginally stable. At Kc, the closed-loop system has imaginary axis poles, and A(s) gives their exact frequencies. This is frequently tested in GATE problems involving finding the range of K for stability.

Practical Understanding

In practical control design, a zero row in the Routh array at a particular gain value is an important signal. It indicates that the system is at the boundary between stable and unstable operation. For the system to be useful, the operating gain must be kept below the critical value where the zero row occurs. Above this gain, the closed-loop poles cross into the right half plane and the system becomes unstable.

The auxiliary polynomial also provides useful information for frequency-domain design. The imaginary axis poles correspond to sustained oscillations in the time domain. The frequency found from A(s) = 0 is the natural oscillation frequency at the stability boundary, which can be correlated with gain margin and phase margin in frequency response methods.

Given:
Characteristic polynomial D(s) = s⁵ + 2s⁴ + 24s³ + 48s² - 25s - 50
Find: stability and imaginary axis pole locations if any.

Why this formula applies:
Expecting a zero row. Apply auxiliary polynomial method.

Routh Array:
  s⁵ |  1      24     -25
  s⁴ |  2      48     -50
  s³ |  b1     b2

b1 = (2×24 - 1×48)/2 = (48-48)/2 = 0   ← Entire row check needed
b2 = (2×(-25) - 1×(-50))/2 = (-50+50)/2 = 0

Entire s³ row = [0, 0, 0] ← Zero row!

Form Auxiliary Polynomial from s⁴ row:
A(s) = 2s⁴ + 48s² - 50

Derivative: dA/ds = 8s³ + 96s
Coefficients of dA/ds: 8, 96 → replace s³ row

  s⁵ |  1      24     -25
  s⁴ |  2      48     -50
  s³ |  8      96     (from derivative)
  s² |  c1     c2

c1 = (8×48 - 2×96)/8 = (384-192)/8 = 24
c2 = (8×(-50) - 2×0)/8 = -400/8 = -50
  s¹ |  d1
d1 = (24×96 - 8×(-50))/24 = (2304+400)/24 = 112.7 (positive)
  s⁰ | -50

First column: 1, 2, 8, 24, 112.7, -50
Sign changes: 112.7 → -50 = 1 sign change

Roots from A(s) = 2s⁴ + 48s² - 50 = 0:
Let u = s²: 2u² + 48u - 50 = 0 → u = (-48 ± sqrt(2304+400))/4
u = (-48 ± 52)/4 → u = 1 or u = -25
s² = 1 → s = ±1 (real axis, symmetric)
s² = -25 → s = ±j5 (imaginary axis poles)

Final Answer:
1 sign change → 1 RHP pole (s = +1).
Imaginary axis poles at s = ±j5.
System is UNSTABLE with oscillatory modes at ω = 5 rad/s.

Exam Tip: When a zero row appears in the Routh array, the roots of the auxiliary polynomial A(s) are the symmetric poles. If A(s) has purely imaginary roots (s = ±jω), those are imaginary axis poles. The frequency ω found from A(s) is the oscillation frequency of the marginally stable modes. This frequency is often directly asked in GATE problems.

• An all-zero row in the Routh array indicates a symmetric factor in D(s) with roots symmetric about the s-plane origin.
• Form the auxiliary polynomial A(s) from the row immediately above the zero row using powers decreasing by 2.
• Replace the zero row with the coefficients of dA(s)/ds and continue the Routh array construction normally.
• Roots of A(s) are the special symmetric poles. Purely imaginary roots give imaginary axis poles and the oscillation frequency.
• Count sign changes in the first column of the completed array to determine total number of RHP poles.
• The zero row case and the single-zero-first-column case are completely different and require different resolution methods.

Quick Revision

• Zero row (all elements zero): caused by a symmetric even polynomial factor in D(s).
• Resolution: form A(s) from the row above, differentiate, use dA/ds coefficients to replace zero row.
• Roots of A(s): imaginary axis poles (±jω), real symmetric poles (±σ), or quadrantal pairs.
• Marginal stability: only when all roots of A(s) are purely imaginary AND no RHP poles from rest of array.
• Oscillation frequency at marginal stability: ω = sqrt(β/α) from A(s) = αs² + β = 0.
• GATE trap: zero row indicates imaginary axis poles, but system could still be unstable if the full array has sign changes.
• Critical gain Kc is the value of K at which a zero row first appears, found by setting the first-column element of a row to zero and solving for K.