Learning from my mistakes of last time, I decided to actually test the food deployment mechanism before designing the final product. Pictured above was my first attempt. The concept of operations is simple, as illustrated below. The "water wheel" is turned by a worm-gear DC motor connected via a flange coupler. Food drops from the hopper into the water wheel, which turns, dropping the food into the chute where it would be flowed to a food bowl for consumption.
My biggest problem to solve is to preventing kibbles from getting jammed in whatever mechanism I use and jamming up deployment. This model is specifically designed to prevent these jams with the curved lip at the bottom of the chute intended to block falling kibbles from getting stuck.
After designing and printing the wheel and hopper/chute parts, I designed the test fixture.
This uses the following parts:
Other materials included M4-7 bolts/nuts/washers, drywall screws, scrap wood, and L braces.
Finally, I connected it to a power supply and set it going!
Sadly, the results were disappointing. Although the fixture delivered food consistently and reliably, jammed kibbles was a big problems.
Roughly every ten seconds during active feeding, a kibble would become trapped between the wheel and the hopper chute. This would cause current to spike to 500mA (nominal 100mA), and more importantly, caused the entire fixture to jolt violently after the kibble got crushed or the jam cleared.
After only a minute of operation, there was significant damage to the lip of the hopper, which was specifically designed to prevent jams.
Sadly, I'm going to put this design in the bucket of unsatisfactory results.