Output setted from c-mex is sometime right, sometime null

Hi all! I stop my simulation with Stop block, setting from C-mex file the respective output to 1 when a condition is verified. If this condition is valid, before to stop, I set a few values in a second output vector that I want to read externally in a M-file. I don't know why, but some times it give me in output the right values, other times it give me a null output! This problem has to do with the ode solver I use (ode45)?

Thanks in advance!

Answer 1

[Kaustubha Govind]

More likely that there is something wrong with the way your S-function is written. Could you post your mdlOutputs implementation?

[Marcus]

It's very long and intricate to copy here. By the way I'm sure that it's right because: 1) sometimes goes well and sometimes not; 2) before stop the simulation I print the output values from c file (correct) but when I read them from M file they are null.

[Kaustubha Govind]

Then, perhaps you could show us how you are assigning the output to the S-function output signal?

[Marcus]

    real_T            *y1    = ssGetOutputPortRealSignal(S,1);
    real_T            *ystop    = ssGetOutputPortRealSignal(S,2);

    ...

    if (<condition>) {      //if condition is true
        y1[1] = <value1>;   //set some values in the y1 output
        y1[2] = <value2>;
        ...
        ssPrintf("%f\n", y1[1]);  //print on the console value1
        ystop[0] = 1;             //stop the simulation 
    }

It's right? Value1 is printed correctly on the console, but the respective output y1 (read in Matlab workspace) is all null.

[Kaustubha Govind]

Just to clarify - the ss* API uses zero-based indexing, so "y1" is writing to the second output port, and "ystop" is writing to the third output port - is this intended?

If it is, could you verify that ssGetOutputPortDataType(S,1) and ssGetOutputPortDataType(S,2) both return 0? This will ensure that both outputs are indeed expecting double values.

[Marcus]

I have another output port, so the number of that is correct. I tried also ssGetOutputDataType and it returns 0, so I think this is not the problem.

[Kaustubha Govind]

Also, not sure if it was a typo, but you should have:

        y1[0] = <value1>;   //set some values in the y1 output
        y1[1] = <value2>;
        ...

(Zero-based indexing)

[Marcus]

Indexing is not the problem (yes is a typo). I don't know what to think :/

[Kaustubha Govind]

Strange. I'm out of ideas - perhaps you should send your S-function to MathWorks Tech Support so they can investigate it more closely?