[Billfred]

Originally Posted by Al Skierkiewicz

Andrew,
I need to weigh in here from an electrical standpoint and add some other observations.
Drive systems that cannot steer, eat batteries when turning on carpet. The side loading is incredible and many teams have reported to me that they can barely last one match on a fully charged battery. (I have observed low battery conditions on several robots at the end of matches.) The data we collected with current monitoring suggests that most multi-wheel and tread drive systems regularly run near stall currents on all drive motors during turns. Slow, tight turns are obviously worse than higher speeds. The degree of battery drain depends on driver strategy and operation. A driver who only drives forward and back with little steering should not experience any greater battery drain than other designs. Due to the additional friction in the drive components, (pulleys, drive trucks, etc.) a treaded design generally runs higher electrical demand. Tighten up the belt to prevent it slipping off and those frictions increase. Only low friction bearings can help on these problems. When asked, I recommend that a robot should be able to run two to three matches on a single battery as a rule of thumb. This margin insures enough electrical power be available for a single match.
Although the discussion on treads and carpet friction is very interesting, I think there are too many variables (assumptions) for this to be a general discussion. Designs that have no way of transferring loads to the tread except at the wheels have to be analyzed differently than those that spread the load across the full length of tread in contact with the carpet. Even those designs that use center trucks that are lower than the either the front or back pulleys require a different analysis. Turning friction varies with design as well. Often in turns, belts tend to go "up on edge" so that the full width is no longer in contact with the carpet.
I am not saying that tank designs are bad but there obvious problems and tradeoffs. Teams that have successful tank designs do exist but further investigation into their design criteria is warranted.

Al,

I remember back in my BattleBots phase that one builder suggested running your robot against a curb for three minutes (in their case) to check and make sure you absolutely have enough power to last a shoving match. Would such a method work for testing FIRST robots as well?