Hi Ian, I just hooked up some power earlier and tested the prototype.

After testing it for only 5 mins I can see signs of the aluminum wearing.
It is also much louder than I thought it would be. I think I'll try Nylon next as per your suggestion.

If I can find some smaller Cam followers I'm going to scale it all down and use some 9mm or 12mm linear rails.
Nothing in this current prototype is really sized correctly. The whole design is not very forgiving, if anything is slightly out of alignment it won't run properly.
I reckon these systems call for extreme precision, I wonder If spring loading the rack and roller pinion together would work?

I too thought the reverse design you mentioned, was the ideal solution, (better to replace a worn plastic gear than a whole rack) but I spent a few hours on CAD investigating this option a while back and came to the conclusion that its impossible. I thought that it would work in the same manner as a chain and a sprocket, but when I tried to design it in CAD I couldn't design something that would work.
I think the issue is that a chain wraps around a sprocket, when a sprocket moves the chain is constantly changing form, having the rollers positioned in a straight line means that you can't get any rolling motion.
If you have an example of a design that works in this manner, I'd love to see it, I agree this would be a much better solution but I'm not sure it is possible.

I should probably start a separate post rather than hijacking this one.
Here is a video of it in action.

Nikki