gaRCfield, be careful here, there are a number of systems within the servo model which exhibit "deadband" by definition.

Yes, deadband could be refered to as the time between pulses from the receiver AND the slop in the gearbox BUT in the context used in this thread it refers to the amount of change in the input signal (or error in output position) before the controller will provide a control signal to the motor to get things straight again. The word hysterysis could almost be used but is slightly different again.

With a 5 microsecond "deadband" the signal from the RX would need to deviate 5 microseconds before the controller would try to reposition the servo, 5 microseconds usually being about 1 click of trim depending on the radio. This allows the signal and servo to fluctuate a bit without the servo motor being given a drive signal.

Narrow deadband in this context is not how often the servo is being told to get with the program but refers to how far off track it can get before it gets a kick in the rear