[EricH]

One physics explanation coming right up!

Normal force, coefficient of friction, and torque exerted by the wheels all come into play here. Normal force is applied at each point of contact the robot has with the ground. The normal force * the coefficient of friction dictates the maximum force that can be applied horizontally at that point before something slips. (You've probably heard something like this in your physics classes--if you haven't, you will.)

Under the current setup, all 4 wheels are exerting a torque, which can be translated into a force tangent to the wheel (we'll call that tractive force for now) at the point of contact with the ground. This is why the robot moves in the first place.

But, when you remove two power sources, you remove two powered points of contact. The contact remains, which means that some friction remains (to be dealt with either by overwhelming force or by bearings on the robot). However, now all the available power runs through only two points of contact. The maximum possible force remains the same, however, due to the normal force not changing, so the maximum tractive force per wheel remains the same.

But wait! You just removed two places to apply power! Now you have less available tractive force; the two wheels left have to drive the whole robot; and friction hasn't changed.


The physics behind the "scrubbing" phenomenon are kind of similar: force at a distance. Let's figure that you turn around a point on your back axle line, for now; I'm sure that you can figure out what would happen if you moved that point forwards towards the middle.

You have two drive forces, one in each direction, at ~ 13" from the pivot point (I'm assuming it's in the center, and counting both wheels on the same side together). You have a moment (F*distance, in this case times 2). But, on the other end of the robot, ~36" away, you have the side friction forces of the wheels. Those also exert a moment, in the opposite direction of motion (because that's the way friction always goes). What often happens with a 4WD is that that moment due to friction is very close to the moment caused by the drive... and life gets very interesting until the robot starts bouncing and sliding.

A 6WD drop-center puts the center of rotation close to the center of the robot and cuts that long distance in half, while at the same time reducing the normal force (and therefore the friction force) at one end. Omni wheels can really cut down on the sideways friction force; casters can as well, but add in turning resistance (though ball casters add less than swivel casters).

Now, why does the robot work fine on non-carpet surfaces? Carpet has a rather high coefficient of friction with rubber. Like I said before, get some of the event carpet on Saturday and use it to test on next year.