By $F_{\sigma\delta}$, I mean that the set can be expressed as a countable intersection of $F_\sigma$ sets.
Let this sequence of functions be $f_n$, and the set of points where $f_n$ converges be $C$. Since $f_n$ must be Cauchy, I can define $C$ as
$$C = \{x \in \mathbb{R}\:|\: \forall\epsilon\in\mathbb{N} \; \exists N\in\mathbb{N} \::\:|f_m(x) - f_n(x)| < \frac{1}{\epsilon} \: \forall m,n > N\}$$.
Looking at this set, I'm trying to prove $F_{\sigma\delta}$ by firstly trying to find some countable intersection. I try to rewrite $C$ as
$$C = \bigcap\limits_{\epsilon \in \mathbb{N}}\{x \in \mathbb{R}\:|\: \exists N\in\mathbb{N} \::\:|f_m(x) - f_n(x)| < \frac{1}{\epsilon} \: \forall m,n > N\}$$.
Now I need to express each set being intersected as a countable union of closed sets to prove that it is $F_\sigma$. Then the countable intersection of these must be $F_{\sigma\delta}$. I tried to pick $N$ to describe this union, and rewrote $C$ as
$$C = \bigcap\limits_{\epsilon \in \mathbb{N}}\bigg[\;\bigcup\limits_{N \in \mathbb{N}}\{x \in \mathbb{R}\:|\:|f_m(x) - f_n(x)| < \frac{1}{\epsilon} \: \forall m,n > N\}\;\bigg]$$.
But this option confuses me because $N$ depends on $\epsilon$, so I don't know whether it would make sense to let both vary freely like that. Am I going in the wrong direction?