[Joe Johnson]

Not everybody tells you to lower the wheels!

There are trade offs to every decision in a design. You pays your money and you takes your choice.

While I like 6WD with the lowered middle wheel, I have made 4WD robots with awful aspect ratios and with the wheels in the corners. As Wayne puts it, enough torque can just "overrpower the lateral friction" .

Now to the specific point of lowering the middle wheels: The decision to not to lower the middle wheels is not as bad as one may initially fear. In the worst case, it is still much better at turning than a 4WD robot with the wheels in the corners and in many cases it is as good at turning as a 6WD robot with the wheels lowered.

How so? In order to understand, you have to think about center of mass and free body diagrams.

Case 1:
CG very nearly over the middle wheels. Assume a rigid frame, flat floor, same size wheels, round wheels with axles in true center AND alligned axles.*

Then the weight on each wheel is W/6 (W=weight of the robot). It is the front and rear axles that will scrub (i.e. move laterally on the floor).

Compare this to the 4WD case where the front and rear wheels scrub exactly the same amount BUT in this case, the weigh on each wheel is W/4.

Bottom line, scrubbing effects are reduced by 33% ((1/4)-(1/6))/(1/4))

Case 2:
CG is not near the middle wheels (this is the more common case), same assumptions*.

The weight on the wheels that the CG is between share the full weight of the robot. The remaining wheels have no weight on them. For example if the weight is toward the back of the robot, the middle and rear wheels carry all the weight of the robot while the front wheels are unloaded.

Bottom line, this is the same performance as the 6WD case WITH lowered middle wheels.

Joe J.

*It is because none of these things are true that I recomend lowering the middle wheels. By lowering the wheels you allow things to work out well even with non perfect assumptions.

I was in a rush to get dinner when I posted the above. Since eating, the blood has returned to my brain. My analysis in Case 2 is just wrong. The case is a classic example of a "statically indeterminant" system. In such cases, you need resort to the stiffnesses of the members to determine the loading on each wheel. But... ...even so, in this case, to the extent that the middle wheels carry any load at all, it is better than the 4WD case with wheels in the corners. And with reasonable assumptions about the chassis it is not too hard to get the same answer as the one in the above message, even though I arrived at it in error.

Joe J.