Error for multi output estimation of output-error model (oe function)

Hello,

I am using the system id toolbox to estimate Output Error models for an experiment with multi output data.

I keep getting an error when the number of outputs is larger than the number of inputs * 3. I.e. an error when numOutputs > numInputs*3.

For example, the following code gives an error at the fourth iteration.

% try to fit multi output data to oe model

% generate multi-ouput data with one input. 
% use noise to show the error
% single channel nonise as input
inputData=ones(1000,1);
sampRate=0.001;

% create output with various number of channels

for numOutputs=1:4

      outputData=ones(size(inputData,1),numOutputs);   
      testData=iddata(outputData,inputData,sampRate);

      %try these orders

       nb=4*ones(numOutputs,1);
       nf=5*ones(numOutputs,1);    
       nk=5*ones(numOutputs,1);

       chk=oe(testData,[nb nf nk] );

end

I have traced this to the check at line 428 of validatePEMInputs:

   % Polynomial model.
        if nc<nr
           error(message('Ident:general:InvalidSyntax',caller,caller))
        elseif ~DoubleData && Nydat~=nr
           error(message('Ident:estimation:polyestOrderRows',caller))
        end

Here nc=numOuputs and nr=numInputs*3.

I figured i'm, specifying the syntax (the orders) incorrectly, but I can't see what I'm doing wrong. Any ideas why this may be so?

Thanks,

Dror