Dominant poles

Isn't your textbook any good? Do you know what "dominant pole" means? Tell me what you think it means and we'll discuss it some more.

Jerry

Hi Jerry,

Well my books says, in the higher order system if the ratio among real parts is more than 5 then there is a dominant pole.

Is it right that in this case there exist no dominant pole?

Well for their effect to the system no idea?

What is the boundary in the left half s plane on real axis for which we say, this is the region beyond poles are near to the imajnary axis?In my case one ploe lies on the real axis at 35 other at 75 and third at

Zia

All of that. Dominant poles determine the general shape of the frequency response in the region of importance. If you plot the response several times, omitting a pole (or complex pole pair) from all but one of the plots, you will see that some poles don't change the overall shape much. The poles whose absence make a large difference are the dominant ones.

How much is much is a matter of judgment. Poles far from the imaginary axis tend to have weak influence. Usually, poles much higher in frequency than most others have little influence in the region that matters. After all, what happens to a response curve after it falls below unity gain doesn't greatly affect loop stability.

Jerry

A Polish national wearing shiny black leather and carrying a cat-o-nine-tails?

...

I have to add that to my list. :-)

Jerry

Consider the real value of the poles (sigma) - the response that a negative sigma is responsible for in the system is one of exponential decay

the further to the left i.e. the greater negative value of sigma the faster the time constant for this particular pole.

This is very important for tuning your design i.e.select a frequency range which to model your system and reduce the order of the model to remove time constants that are extremely fast (I think 5 - 10 times the dominant pole is acceptable)

Say that you modelled a temperature measuring system and the system was first order with a slow time constant say now that you would like to measure this and include the measurement in the model. - remember the sensor will have its own dynamics.

The point is that: as long as the temperature measurement has a sensor with a fast response (small time constant) then there is little need to include its dynamic effect in your overall model.

Maybe you should try simulating a controller for a simple first order system and then vary the dynamics of the measurement to see where the dynamics of the sensor become crucial.

Hope this helps

Kieran

Never looked at the date on that post - how old?????