Skip to content
Tutorials Center
My learning

Topics

Lesson 16 of 37 ยท Stability

The Routh-Hurwitz Criterion

Build the Routh array step by step, count right-half-plane roots from sign changes, and implement the Routh-Hurwitz test in Python to check against np.roots.

Finding the roots of a high-degree polynomial used to be hard work, and even today you often want to know stability without solving for roots, especially when a gain K is still a symbol. The Routh-Hurwitz criterion answers the question from the coefficients alone. It tells you how many roots of a polynomial lie in the right half-plane, using only simple arithmetic on a table.

A necessary condition

For the polynomial a_n s^n + a_(n-1) s^(n-1) + ... + a_0 with a_n > 0, all roots can be in the left half-plane only if every coefficient is positive (no missing powers, no sign changes). This is a quick first test:

  • s^3 + 2 s^2 + s + 1: all positive, so it might be stable; check further.
  • s^3 + 2 s^2 + 1: the s term is missing, so it is not stable.
  • s^3 - s^2 + 4 s + 2: a negative coefficient, so it is not stable.

The condition is not sufficient. For a polynomial of degree 3 or more, you still need the full test.

Building the Routh array

For a_4 s^4 + a_3 s^3 + a_2 s^2 + a_1 s + a_0, write the coefficients alternately in the first two rows:

Text
s^4 | a4  a2  a0
s^3 | a3  a1  0
s^2 | b1  b2
s^1 | c1  c2
s^0 | d1

Each later entry is a negative determinant divided by the entry in the pivot position:

Text
b1 = (a3 a2 - a4 a1) / a3
b2 = (a3 a0 - a4 * 0) / a3
c1 = (b1 a1 - a3 b2) / b1
d1 = (c1 b2 - b1 * 0) / c1

The theorem: the number of sign changes in the first column equals the number of roots in the right half-plane. All roots are in the left half-plane exactly when the first column has no sign changes (all entries have the same sign).

Routh array recipe

Fill two rows Alternate coefficients, highest power first.
Cross-multiply Each entry is a determinant over the pivot above.
Look at column one Count the sign changes.

Sign changes equal roots in the right half-plane.

1 / 2

A worked example

Characteristic polynomial: s^4 + 2 s^3 + 3 s^2 + 4 s + 5.

Text
s^4 | 1    3    5
s^3 | 2    4    0
s^2 | (2*3 - 1*4)/2 = 1    (2*5 - 1*0)/2 = 5
s^1 | (1*4 - 2*5)/1 = -6   0
s^0 | (-6*5 - 1*0)/(-6) = 5

The first column is 1, 2, 1, -6, 5. There are two sign changes (from 1 to -6, and from -6 to 5), so the polynomial has two roots in the right half-plane: unstable.

Implementing it

A short function builds the array and counts sign changes. We then compare the verdict with np.roots.

Python A Routh array function

Press Run (or Ctrl + Enter) to see the result here.

Program input ยท what you'd type while the program runs

Type the answers before you press Run, one per line. Each time the program reads a line (with input()), it gets the next line from this box.

For every case the two counts agree. The first example matches our hand calculation, with two right-half-plane roots.

Using Routh to find stable gain ranges

The real power appears when a parameter is symbolic. For a loop with characteristic polynomial s^3 + 6 s^2 + 8 s + K:

Text
s^3 | 1          8
s^2 | 6          K
s^1 | (48 - K)/6   0
s^0 | K

Stability needs every first-column entry positive: 6 > 0, (48 - K)/6 > 0 so K < 48, and K > 0. The stable range is 0 < K < 48. At K = 48 the s^1 row is zero (we cover that special case in Routh Special Cases) and the polynomial has a pair of imaginary-axis roots.

Python Checking the gain range found by Routh

Press Run (or Ctrl + Enter) to see the result here.

Program input ยท what you'd type while the program runs

Type the answers before you press Run, one per line. Each time the program reads a line (with input()), it gets the next line from this box.

The sign of the largest real part flips exactly at K = 48, with roots at +/- 2.83 j (that is +/- sqrt(8) j), found by solving the s^2 row: 6 s^2 + 48 = 0. This auxiliary equation gives the frequency of oscillation at the stability limit.

Shortcuts for low orders

Degree Polynomial Stable if
2 a2 s^2 + a1 s + a0 all coefficients same sign
3 a3 s^3 + a2 s^2 + a1 s + a0 all positive and a2 a1 > a3 a0
4 a4 s^4 + ... + a0 all positive, a3 a2 > a4 a1, and a3 a2 a1 > a4 a1^2 + a3^2 a0

The degree-3 rule explains the earlier example s^3 + 2s^2 + s + K: 2 x 1 > K, so K < 2, matching the scan in Stability and Pole Locations.

Limits of the method

Routh-Hurwitz works for polynomials, so a time delay (which is e^(-sT), not a polynomial) needs an approximation. It also tells you how many roots are in the right half-plane, not where. For relative stability, shift the axis (s = z - a) and test again, or use the tools of later chapters.

Check your understanding

Quick check ยท 4 questions

of 4 answered You scored / 4 ยท perfect!

  1. 1. The polynomial s^3 + 2 s^2 + 1 is missing an s term. Is it stable?

    A necessary condition for stability is that all coefficients are present and have the same sign.
  2. 2. The first column of a Routh array is 1, 2, 1, -6, 5. How many right-half-plane roots are there?

    The sign changes from 1 to -6 and from -6 to 5, so two roots are in the right half-plane.
  3. 3. For s^3 + 6 s^2 + 8 s + K, what is the stable range of K?

    The s^1 entry is (48 - K)/6 which must be positive, and the s^0 entry K must be positive.
  4. 4. For a cubic a3 s^3 + a2 s^2 + a1 s + a0 with positive coefficients, the stability condition is

    The s^1 row of the Routh array is (a2 a1 - a3 a0)/a2, which must be positive.

Summary

  • Routh-Hurwitz counts right-half-plane roots from sign changes in the first column of the Routh array, without solving for roots.
  • First check that all coefficients are present and positive.
  • With a symbolic gain, the first-column conditions give the stable range of K.
  • The auxiliary equation at the boundary gives the oscillation frequency.
  • It handles polynomials only; delays need approximation.

Next: Routh Special Cases