See how far you can throw a rock into a stream of smaller particles falling off a conveyor belt and then try the same rock at the same speed straight down into the pile of dumped particles.
The rock will knock aside and penetrate the dispersed particles but makes much less progress into the packed bed.
But you've known this since you were a 3 year old playing in a sand box so don't bother. Instead apply it to the BNE:
Since the static acceleration of gravity + the cyclical acceleration tends to pack a bed of fine particles at the bottom of the container for a longer time than at the top, the larger particles will tend to progress upward over one or more strokes.
This explains why there's an optimum cyclical acceleration amplitude/number of cycles for separation. At high cyclical accelerations a packed bed forms on the ceiling of the container reducing the BNE. At low cyclical accelerations, of course, less or nothing happens at all.
This is also consistent with the fact that the BNE works in a vacuum.
The BNE has been studied because fertilizers and other granular materials tend to separate during shipment or anytime they are moved in bulk. Multi-story towers must be built at conveyor belt terminals in order to get a representative sample to sell the product.
The proper explanation of the BNE should have at least one beneficial effect. It could stop any further wishful thinking on easy ways to stop it.
You either reduce cyclical accelerations to below g or you re-mix the product.
Bret Cahill