Thread Subject: Bug? in root locus plot using "rlocus" from Control System Toolbox

Hi All,

Please try the below two similar lines:
>> rlocus([1],[1 4 5.33 0])
and
>> rlocus([1],[1 4 5.34 0])

They produce different results of plot. Can anyone tell me why? Or, there is a bug in the "rlocus" function.

Thank you.
H.C. Liaw

"Hwee Choo Liaw" <hwee-choo.liaw@eng.monash.edu.au> wrote in message
news:21613842.1194860796364.JavaMail.jakarta@nitrogen.mathforum.org...
> Hi All,
>
> Please try the below two similar lines:
>>> rlocus([1],[1 4 5.33 0])
> and
>>> rlocus([1],[1 4 5.34 0])
>
> They produce different results of plot. Can anyone tell me why? Or, there
> is a bug in the "rlocus" function.
>
> Thank you.
> H.C. Liaw

I do not think this is a bug.

Remember that root locus is the plot of the locations of the closed loop
poles as gain is changed.

Location of closed poles are roots of the characteristic equation, but root
locations (solution of the char. equation) could change in greater amount by
just making a small change in the coefficients of the equation itself. ie.
it is possible for a small change to an equation coefficient to cause a
larger change in its roots location.

I could be wrong since its been a while.

Nasser

Hwee Choo Liaw <hwee-choo.liaw@eng.monash.edu.au> wrote in
message
<21613842.1194860796364.JavaMail.jakarta@nitrogen.mathforum.org>...
> Hi All,
>
> Please try the below two similar lines:
> >> rlocus([1],[1 4 5.33 0])
> and
> >> rlocus([1],[1 4 5.34 0])
>
> They produce different results of plot. Can anyone tell me
why? Or, there is a bug in the "rlocus" function.
>
> Thank you.
> H.C. Liaw

There is no bug here. These are two different systems. They
have slightly different root loci as should be expected.

Duane

Dear Nasser,

Thank you, you are right!

H.C. Liaw