[MasonMM]

We really like the 17T double sprocket approach to Texas Tube.

When we put two pieces of chain side by side on 16T double sprockets we found the width of the chain interfered with the bearings pressed into the side of the tube.

When we put a piece of chain on an 18T sprocket it crashed with the top and bottom walls of the extrusion.

The magic of the 17T is that the chain rides over the OD of the bearing that protrudes through the extrusion wall and the chain doesn't hit the top or bottom of the inner extrusion walls. Its such a perfect fit and the chain doesn't have anywhere to go even if it wanted to jump off of a sprocket tooth. As Aren mentioned, with the thinner walled VEX extrusion, you can squeeze an 18T sprocket in.

Prior to 2013 we talked about and CADed Chain-In-Tube (now Texas Tube), and then the smaller robot size in 2013 convinced us to try it. We machined 4 bearing holes in a 2x1 extrusion, waterjet a 17T tooth profile out of 3/8" aluminum, some post sprocket lathe work, and we had a working drive rail prototype in one night. We've run the 2013, 2014, and 2015 seasons with only one drivetrain failure, diagnosed as a piece of improperly built chain... but otherwise its extremely reliable.

Its a simple, robust, easy, light, and small way to build a drivetrain. This year we bought sprockets from 221, and machined our chassis rails on a manual mill. Ain't nobody got time to be reinventing a drivetrain every year.