After more testing with a slightly bigger hole and a heavier eccentric load on the motor, the real culprit appears to be too much load on the vibration plate by the column of nuts. When it's full the nuts put too much pressure on the dispenser and it has trouble vibrating enough. When the hopper is only half full it works great. I had added baffles a while ago to mitigate this, but need to rearrange them to reduce the load further on the vibration plate.

I can go back to a 22mm opening and I think I'll keep the heavier eccentric load Rearranging the baffles a bit should solve this.

Once this dispenser issue is solved I think it's "ready" as an MVP and I can build a second one.

 Everything else appears to be working fine: 

• activating when a bird has landed/stepped on the platform
• detecting deposited items
• structural and mechanical integrity

I hope to finish this up in the next few days. 

UPDATE

New baffles seem to be working, even when completely full of peanuts. It's a little louder now, but dispenses faster. We'll see how it does tomorrow.

UPDATE 2

With the 22mm hole I had used a hole punch. Usually one side of the hole has a burr on it (exit side for the punch blades). I had purposely put this on the upward-facing surface of the dispenser thinking it would prevent accidental dispensing if the machine was bumped. Turns out this small burr on the edge of the hole contributed to the jams. I smoothed it out with a dremel and things are working better. Will see how the rest of the day goes. I've already had to fill the bottlecap magazine 4 times today. 

LAST UPDATE ON THIS LOG POST

The following all helped:

• bigger eccentric load on the motor
• repositioned baffles
• filing off the burr around the lip of the dispenser hole

There is still a lot of room for improvement, but it's been a great learning experience so far.