We've been debating the "shift" vs "no shift" and "2 motors" vs "4 motors" on my team since the end of 2002's competition. We have not yet reached a conclusion!
However, the key debate points are:
1. resources (motors, machining time, battery current, materials [$$$]) allocated to the drive train
2. reliability and criticality of failure modes (ie a shifter will automatically be less reliable [more parts to fail] and a failure -may- be critical)
3. necessity to the game (ie if you cannot spare the resources and you don't need it for your strategy, it may cost you in other parts of the robot)
Our current position is:
If you want to engage in a pushing contest, it is essential to have four motors in the drive train and be able to shift into a very torquey low gear.
If you are really torquey but cannot move into high gear, you can be out-maneuvered in most matches.
I don't think we have the resources to engage in a brute force pushing contest at this point, so I doubt that we will go down that road.
The main problem which I have seen with many of the published shifter designs and which occurs with the Bosch gearbox is that they can get "out of gear." You can end up with no gear ratio in a failure. This seems to be too much of a drawback to risk those designs.
We have two designs (and whether we build one remains to be seen) which do not have this drawback. They default to the high gear (fast mode) in a failure. If we do build this, I'm sure we'll discover why this is a bad idea.
Another reason for a shifter comes from speed. Low gear gives you great acceleration. If you can shift on the fly, you will actually be faster in terms of time to objective, than a robot with a really high gear. This might rely on a vehicle speed measurement, which is difficult with the existing parts.
Do those of you with shifters on the fly do it via the driver or do you shift at a particular vehicle speed?
Andrew, Team 356