Wrong output with MATLAB S Function

When I implement simple SOGI block to generate quadrature signal for given sinusoidla input, output magnitude comming exactly half. SOGI will give two outputs one is alpha and another is beta component. Beta component magnitude and phase is correct but alpha component magnitude is comming exactly half. Same discrete equations when I implemented in matlab function or tf block in simulink its giving correct. Same equations and everything same but when I implemented in s function its giving wrong.

Correct output with MATLAB Function: (Valpha and vin both having same phase and maginitude)

Wrong output with S Function (C code): (Valpha and vin both having same phase and maginitude)

I attached the simulation files of s function, This error is only because of s function related issues. Please check and let me know.

Thanks in advance.

Answer 1

[Divyajyoti Nayak]

Hi @AYYAPPA RUDRASIMHA,

The simulation results using the MATLAB function block and S-function block are different because the C code you have written for the S-function block is not equivalent to the code in MATLAB function block due to a small difference.

In the MATLAB function block, the ‘err’ variable is given the value of the input variable ‘ref’ after the assignment of variables ‘err_n2’and ‘err_n1’.

err_n2  = err_n1 ; %Initially Zero
err_n1  = err ; %Initially Zero because err = 0
err  = ref; %The input ref is assigned to err

While in the S-function block, the equivalent variables ‘SOGI_v.vin’, ‘SOGI_v.vin_n1’, and ‘SOGI_v.vin_n2’ are handled differently. SOGI_v.vin is assigned the input variable ‘vin’ before ‘SOGI_v.vin_n1’ and ‘SOGI_v.vin_n2’ have been assigned. This makes the initial state of the variables different and hence leading to the wrong result.

SOGI_v.vin = vin; %Assigned before SOGI_v.vin_n1 and SOGI_v.vin_n2
SOGI_v.d0 = (4 + 2 * K * Ts * w + (w * w * Ts * Ts));
d0 = (4 + (2 * K * Ts * w) + (w * w * Ts * Ts));
SOGI_v.b1 = (-8 + (2 * w * w * Ts * Ts)) / SOGI_v.d0;
SOGI_v.b2 = ((4 - (2 * K * Ts * w) + (w * w * Ts * Ts))) / SOGI_v.d0;
SOGI_v.a_alpha = (2 * K * Ts * w) / SOGI_v.d0;
SOGI_v.a_beta = (K * w * w * Ts * Ts) / SOGI_v.d0;
b1 = (-8 + (2 * w * w * Ts * Ts)) / (4 + (2 * K * Ts * w) + (w * w * Ts * Ts));
b2 = ((4 - (2 * K * Ts * w) + (w * w * Ts * Ts))) / (4 + (2 * K * Ts * w) + (w * w * Ts * Ts));
a1 = (2 * K * Ts * w) / (4 + (2 * K * Ts * w) + (w * w * Ts * Ts));


SOGI_v.vin_n2 = SOGI_v.vin_n1; %Initially zero
SOGI_v.vin_n1 = SOGI_v.vin; %Not initially zero as vin is already assigned

To fix the issue, just write the assignment of ‘SOGI_v.vin’ after assigning ‘SOGI_v.vin_n2’ and ‘SOGI_v.vin_n1’ and it should give the correct result.