# Help finding transfer function

I'm trying to find the transfer function of a circuit and i've created a 2x2 system with the unknown voltages of the nodes.

$$a= \begin{bmatrix} \frac{j*(L_1*L_2)}{w*L_1 + w*L_2} + w*j*C_1 & \frac{-j}{(w*L_2)} \\ \frac{j}{(w*L_2)} & w*j*C_2 - \frac{j}{(w*L_2)} + \frac{1}{R}\end{bmatrix}$$

$$s=\begin{bmatrix} \frac{V_s*j}{w*L_1} \\ \frac{V_s}{R}\end{bmatrix}$$

where $V_S$ is the input voltage and $\omega$ is the angular frequency.

The problem is that when I put this in matlab, because I define $V_S$ and $\omega$ as symbols I end up getting really big numbers. (45 digits and the like) And even trying to solve this by hand is too much work. I know that the transfer function should be 1 because i've checked it in multisim.

Anyone has any idea what to do?

-
Could you post your original circuit? –  suha May 26 '12 at 8:51
I don't get it. How can you get numbers as a result if you enter symbols. It should do it analytically, not numerically. Where does it get these numbers anyway? –  stevenvh May 26 '12 at 8:58
everything else except vs and w are given numerically –  user9970 May 26 '12 at 9:35
Before being able to answer your question, there's something wrong with your $\dfrac{j*(L_1*L_2)}{w*L_1 + w*L_2}$ term. It does not have dimensions of admittance (because the equivalent $\dfrac{L}{w}$ does not have, either). Check that part. –  Telaclavo May 26 '12 at 14:56
Yes I found that mistake too. I corrected it and when I don't use any symbols and do everything numerically the answer matches what I get in multisim. –  user9970 May 27 '12 at 7:33
show 1 more comment