I'm trying to understand the phase-number uncertainty relation for superconductors, \begin{align} \Delta N \Delta \varphi \gtrsim 1. \end{align}
In particular, I'm trying to understand if it holds for the states \begin{align} \vert \psi_\varphi \rangle &= \prod_k (u_k + e^{i\varphi} \nu_k \hat{c}^{\dagger}_{k \uparrow}\hat{c}^{\dagger}_{-k \downarrow}) \vert \psi_0 \rangle \\ \vert \psi_N \rangle &= \int_0^{2\pi} e^{iN\varphi/2} \vert \psi_\varphi \rangle \end{align} of well defined phase and well defined particle number.
It seems that, for the first case $\Delta N$ is finite (estimated by Tinkham to be $\approx 10^9$ - for a macroscopic classic superconductor, I believe), while $\Delta \varphi$ is zero (since $\varphi$ is well defined). For the second case, on the other hand, I would say $\Delta N = 0$ and $\Delta_\varphi = 2 \pi$. In both cases, the product of uncertainties is zero. But if that is the case, the uncertainty relation above would be simply wrong.
Is there a flaw in my train of thought? If not, what is the usefulness of the uncertainty relation above if it does not apply to the two most basic states in this context?