Why the out varies largely in fixed versus variable time step?

I have a model as shown,

for fixed time step the output from the scope 1 is

for variable step, the output is

why there is so much of variation?

Answer 1

[Davide Masiello]

That is because the fixed time step is not able to resolve the very steep change in the variable value that is clearly shown in the solution of the variable step setup.

You could actually obtain the same result with a fixed step by making the step size very small, but this would cause a much larger computation time.

Variable step is a far superior approach for this kind of problems.