Jim: I am currently thinking about picking an origin, and define all my fields by the ith charge with position vector $\vec{r}_i$. Then the field is zero only when they sum to the zero vector, which is always possible since the zero vector is always linearly dependent. Then it is easy to see that the charges will be distributed in a way that there exists a basis formed by the unit vectors such that it spans either $\mathbb{R}^3$ or its subspaces.
Since the picking of an origin is arbitrary and does not affect the physics, the above arugment should still hold and thus there is always a zero…