The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Modeling >> Behavioral Models >> First order GmC integrator
https://designers-guide.org/forum/YaBB.pl?num=1319360550

Message started by NTR on Oct 23^rd, 2011, 2:02am

Title: First order GmC integrator
Post by NTR on Oct 23^rd, 2011, 2:02am

Hi,
I am implementing integrator with tf for ex "tf=1/s+10".I know that in simulink it is easy to implement but I want to do it matlab code. So I am applying this tf to input signal with the help of filter function available in matlab like this "out=filter(tf,input_signal)"; my output should be integrated signal. But its behaving like low pass filter. Please correct me if I am doing wrong in applying integrator transfer function to input.

Thanks for help.

Title: Re: First order GmC integrator
Post by boe on Oct 24^th, 2011, 1:08am

NTR,
the MATLAB function filter expects a z-domain transfer function, not an s-domain TF. You can use the MATLAB function bilinear to convert from s-domain to z-domain.
Does this help?
- B O E

Title: Re: First order GmC integrator
Post by NTR on Oct 24^th, 2011, 2:08am

Hi BOE,
Thank you for your answer and your valuable time . You are correct. But for time being I have used "lsim(tf, input,time)" to get tf response to arbitrary input. It seems working fine.
May I ask you one more question?
What are the key aspects system designer must look at for behavioral modeling first order integrator with (unity gain bandwidth)UGBW 2GHz, DC gain should as high as possible and cutoff frequency must be very low.
I am solving this by using tf=G_dc*Wp/s+Wp where Wp is cutoff freq and G_dc --dc gain.
Here How can i specify UGBW in the transfer function.
thank you once again.

Title: Re: First order GmC integrator
Post by boe on Oct 24^th, 2011, 2:34am

NTR,
UGBW, cut-off frequency and gain are not independent. You have only two degrees of freedom in a first-order system (without zeros, of course).
- B O E

Title: Re: First order GmC integrator
Post by NTR on Oct 25^th, 2011, 3:00pm

HI BOE,
Yes. I think I have to read more about integrator to understand better its design issues at system level.

How to decide saturation limits of integrator based on its transfer function?
/NTR

Title: Re: First order GmC integrator
Post by boe on Oct 27^th, 2011, 3:46am

Hi NTR,

NTR wrote on Oct 25^th, 2011, 3:00pm:

How to decide saturation limits of integrator based on its transfer function?

The transfer function describes the integrator in the linear region, while saturation occurs when you leave the linear region.
Thus, the saturation limits depend on you integrator implementation and your accuracy requirements.
- B O E