I got a really great comment about my design by Exuvo. I wanted to reply but could not put any pictures in the comments. So I decided to make a project log about it.
Exuvo pointed out that in his attempts to make continuous track he had problems with material getting into the chain and not being about to be pushed out. Full scale tanks don't have this problem because the drive mechanism is open on both ends and the sprocket just pushed material out of the way.
When I actually thought about this when I was coming up with the design. As far as I could see there were two options.
Leave the chain open and as you drive over material it could get stuck in the chain. But this allows the sprocket to push it back out. But I was worried that this would increase the risk of the chain jumping off the sprocket.
Cover the chain on the drive side. This prevents material getting into the chain from the ground. But it could still get in from the sides. I felt that this was more likely to reduce the amount of material getting into the chain.
It was not so much an oversight as a design choice. I decided that covering the chain was a better starting point. The other part of this is that the track is rubber and not connected on the outside of the chain. This means that the sprocket can actually push material through the chain and the rubber tread will deform to let it out.
This is actually some of my older prototypes that have threaded rods and nylock nuts holding it together. But you can see that the track can flex out of the way. I am hoping this is enough to keep material out of the chain when you drive over it, but if something does get in the chain that the sprocket can push it out.
My plan is to build this design of the track, and get a working robot to drive around to test it out. I will do some experiments and see what works and what does not. I can always cut the rubber off my tracks and see if that works. I also have an idea about leaving the outside plate off my design and making it closer to the design on actual tanks.