Does $ \sum^{\infty}_{3} \frac{n^2}{(ln(ln(n)))^{ln(n)}} $ converge?
My initial feeling is no, due to the decreasing gradient of $ln(x)$ so I'd expect the individual terms to 'not tend to 0 fast enough'.
I have tried a few common convergence tests but I haven't spotted the conclusion:
The ratio test shows the individual terms to tend to something which could be one (which I haven't been able to evaluate properly) so it seems inconclusive.
The integral test doesn't seem to generate an easy solution as I don't see how to integrate such a function or compare it to another one.
My only ideas is some elementary solution using comparison test or perhaps the Cauchy condensation test but I don't see how it helps exactly.
Any help is appreciated.