[Scilab-users] Define numerator and denominator of transfer function as variables in Xcos (DLR block)

[Nikolay Strelkov]

Dear all!

I'm playing with simple digital filtering in Xcos 5.5.2.
I have two filters - low-pass and high-pass, connected in series.

Xcos does not have convolution block.
So I write FIR transfer function
<https://ccrma.stanford.edu/%7Ejos/fp/FIR_Transfer_Function.html> manually
from impulse response function.
I get impulse response functions from wfir function.

Let's assume that we have two simple filters with impulse responses:
h1 = [0.3741957    0.3741957]; // it's low-pass filter from h1 =
wfir("lp",2,[.2 0],"re",[0 0])
and
h2 = [- 0.1870979    0.8  - 0.1870979]; // it's high-pass filter h2 =
wfir("hp",3,[.1 0],"re",[0 0])

I convert them to DLR <https://help.scilab.org/docs/5.5.2/en_US/DLR.html>
SISOs and get:
H1(z) = 0.3741957*z^-1 + 0.3741957*z^-2 = ( 0.3741957*z + 0.3741957 ) /
(z^2);
H2(z) = -0.1870979*z^-1 + 0.8*z^-2  - 0.1870979*z^-3 = ( -0.1870979*z^2 +
0.8*z  - 0.1870979*z ) / (z^3);
and place their numerator and denominator to the corresponding DLR blocks.
Model is in attachment. It's an illustration, not real world example.

I have a question. How I can automate the aforementioned process
programmatically?
I tried to create numerator and denominator with poly function and as
strings, but Xcos does not support these types in DLR settings.

--




*With best regards,Ph.D., assistant professor at MPEI
<http://mpei.ru/lang/en/Pages/default.aspx>,IEEE member,maintainer of
Mathieu functions toolbox for Scilab
<http://atoms.scilab.org/toolboxes/Mathieu/>,Nikolay Strelkov.*
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.scilab.org/pipermail/users/attachments/20170323/a5adc61a/attachment.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: fir_test.zcos
Type: application/octet-stream
Size: 5408 bytes
Desc: not available
URL: <https://lists.scilab.org/pipermail/users/attachments/20170323/a5adc61a/attachment.obj>