Creating a loop for mle() function for 'tLocationScale' fit

Hello, my plroblem is the following. I have a vector of returns that has length of 4475. I want to create a moving window of 250 observations, so that for each pack of 250 observations I could use a function mle(data,'distribution','tLocationScale); to estimate distribution. But what is crucial for me is to extract vector of degrees of freedom. In a simple version i have:

parameters_data=mle(data,'distribution','tlocationscale'); 

degrees_of_freedom = parameters_data(3); 

if degrees_of_freedom > 3
 dof=degrees_of_freedom;
else
 dof=3
end.

For a version that I wanted to do I used

N = 4475;
l= 250; 
parameters_data=zeros(N-l+1,3);
for t = 1:(N-l+1)
    parameters_data(t,:)=mle(data(t:(t+l-1)),'distribution','tlocationscale'); 
end

But I have the following message "Warning: Maximum likelihood estimation did not converge. Iteration limit exceeded. > In tLocationScaleDistribution>tlsfit at 323 In tLocationScaleDistribution>tLocationScaleDistribution.fit at 177 In fitdist>localfit at 231 In fitdist at 178 In mle at 374 In Untitled at 113 "

Can someone advice me how should I change it so it would work correctly? Thank you in advance!