Different results in MATLAB and Simulink

Hi guys,

I have the same system in both MATLAB and Simulink, and each system has a different step response. One is stable and the other is not. How can I solve this problem? I have attached the systems and responses.

Thanks.

Answer 1

[Stephan]

Hi,

your two systems differ in the second number in the numerator of your transfer function by about a factor of 10:

here you have:

.

0.0001

.

and in the second System you got:

.

which is rounded: 0.001

.

So the numbers are different by factor:

>> 0.0009955 / 0.0001

ans =

                       9.955

I think if you fix this you will get the same results.

Best regards

Stephan

[Hussein Harb]

Oops! My bad! Thanks anyways.

[Hussein Harb]

I just tried it with a different system, and I got different results again with the exact same numbers.

[Stephan]

Did it work with the first system when you corrected the error? Or do we have to search for a further problem for system 1?

Please provide what you have found in the other system.

[Hussein Harb]

The problem was that those systems are results of using (c2d) on continuous systems. So when I try to enter them in discrete time directly I get instability. And I couldn't find the older systems.

Anyways, the problem is apparently from approximations. The systems I'm getting have poles close to 1. So the results of the c2d command must have stable roots, while when entering the system manually, I must be getting different roots due to the approximations.

Do you know how I can increase the precision of MATLAB's calculations.

Thanks!

[Stephan]

To be honest, I do not know that, so I recommend you ask this in a new question.Since the answer to the current question has already been accepted, I suspect that you will find relatively little response to this new question. But there are several very informed people in the forum who can help you, if you turn theit attention to a new unanswered question.

Best regards

Stephan