[M. Mellott]

As Travis mentioned above, our team has used Brecoflex treads for 3 of the last 4 years. We opted for 50mm wide belts with the TK10-K13 self-tracking belt profile in order to use pulley wheels that did not have metal flanges on the outer edges. Depending on the backing you choose, belts that are spliced together for a continuous loop for a half-track design (900-1000mm) run $120-170, while belts used for a systems where they run the full length of the robot (1800-2000mm) can run up to $250. We used 32-tooth wheels for the primary drive wheels and 20-tooth wheels for the support (i.e bogey) wheels, where the wheels will run $38-50 each. We developed a bolt-on modular tread system that allowed us to remove and replace one whole tread side (wheels and all) in case of damage or belt breakage. We bought 6 belts for the '07 season (4+ events) because of manufacturing lead times (2 to 4+ weeks), tread wear (we opted for the cheaper SuperGrip tread), and potential breakage, as well as enough wheels for 4 modules. As mentioned by others, this system is not cheap compared to a wheeled drivetrain. BTW, from our experience, I will also put in my vote for the red Linatex backing--more expensive, but more than worth it with higher CoF and tread material duability.

Weight is also a major issue with this system. The wheels come as solid aluminum billets, with the 32-tooth wheels weighing 4 lb. each, so a lot of custom machining is necessary to save weight. Also, the treads need a lot of support with plates and spacers keeping the wheels from torquing due to belt tension and lateral impact forces from other robots.

Much has been said above about Brecoflex belts derailing or even snapping due to being pushed from the side. The main concern with tread breaking comes from the length of the tread and how it's supported. With tread belts that run the full length of a robot, it's easier for the belt to twist along the carpet due to a lateral force, which can cause derailing or breakage if the driver isn't careful. In order to keep it from twisting, many go to a much higher tension level on the belts, putting a lot of strain on the steel chords running the length of the belts, which can cause those chords to fatigue from the high tension and break that way. A better solution is to add smaller bogey wheels or Delrin sliders to support the lateral stresses on the belts (you typically need one in the center anyway that's lower than the outer wheels to ease turning). Of course, this extra support means extra weight. Another solution we are developing is a half-track system that uses much shorter belts, which we've found not only resists twisting due to the shorter length, but also requires less tension on the belt. The only issue we found at our live testing at IRI and Kettering was treadwear (we used the SuperGrip material at IRI, which wore away quickly, and used Linatex at Kettering which worked much better). I suppose another option might be looking at using tread belts with Kevlar or Stainless steel cords, but I've been too afraid to price them or be willing to purchase some for testing.

In the end, it's a matter of deciding which direction you want the development of your robot to go. With Brecoflex, you can develop a strong drivetrain (assuming you also have beefy transmissions to power those treads), but it's a heavy and often expensive system that takes away weight that could be utilized for other functionality.

I'm hoping to write up a white paper sometime soon on this, but I hope this helps until then. Feel free to contact me directly if you have any questions.