Here you are:

As you can see the evaluation fails for the first and the last value of t.
You are using numerical differentiation for a function derived with "odesolve" for arguments from 0s to t.end
The algorithm used by Prime for this seems to use the function values of arguments before and after the point in question. But in the result of a 'function' determined by "odesolve" there are no function values for arguments before the first initial value of t (0 s) or after the last value (t.end). So the numerical derivative can't be calculated for these end values.
The evaluation should be possible without this problem using the function i.C1(t), which, as I have shown, is determined together with the others in the solution block.