Issue to follow up on my comment on #2335 :

A user just had some simulations erroring because they had obviously set units wrong, which resulted in a grid spacing on the scale of 1e-9. However, this was still an auto grid (they had also set source frequency wrong), so it was not caught by this validator. A large number of our users probably use auto grid instead of uniform grid - it would be good to have a similar validator there. However, in that case it should be a Simulation post_init validator, as the grid needs to be constructed to examine the smallest dl.