The iterative process continues. Here's the latest attempt. 

One nice addition: accurately sized screws, to ensure clearance is ok. Additionally, the model is scaled up to prevent jams. Here's the concept of operations: 

Unfortunately, the baffle wall separating the chute from the wheel chamber extends too far down, and there is too much horizontal distance between the wheel and the chute. As a result, while the destructive jam problem remains fully fixed, the food get stuck in the chute and does not fall into the wheel chamber. 

The wheel design and chamber work well as long as food reaches the wheel, so I'm going to leave the wheel and only rework the chute for the next iteration.