
you're right, didn't think of that ...
my fault.

Okay then, how about this:

Think of a differential, the std mower ones
which is set up in a 3sprocket jackshaft configuration.
Each axle carries a different size sprocket for one of the motors.
The center-Sprocket (the one that's attached to the differential box)
is then direct linked to the axle sprocket.
One motor has a centrifugal clutch the other one direct drive;
both motors are running.
Now the clutch one is the "main motor",
the other one will be our "kicker motor".
- Case1: both motors idle
The diff-axle of our main motor is spinning backwards, taking the load our kicker puts to the other side, differential will remain still: kart is not moving
- Case2: main motor engaged
The diff-axle of the mainmotor is spinning forward, the other side is prevented from spinning backwards by the kicker motor.
The differential case will turn and the kart starts to roll up to a point where the difference between the rpm of our main and the rpm of our kicker will prevent further acceleration
- Case3: main motor highest rpm, kicker revving up
More torque is send through the differential-jackshaft to our axle, making the kart faster until both motors are at their highest possible rpm up to a point where the kicker diff-sprocket has a higher rpm then the main diff-sprocket (due to the different sizes)
And we have "overdrive" of some sort.
Oh and no, I wont even try to figure what gear ratios will be need to make it useful.
But in theory this might work.
And counting a diff as ONE we have a low part count
that's better chris?
'sid