The clear signal that is used to reset the tick counter can be extended/delayed by logic that operates with the fractional calibration value to provide fractional tick adjustment. Every time the fraction counter asserts an extend clear signal to the tick counter, the clear function to the tick counter remains asserted. Any inaccuracy in the oscillator is compensated for by adjusting the calibration value and making the remaining inaccuracy a fraction of a tick in every second. The impact of the remaining inaccuracy can be compensated for by using a fraction counter.
Every 16 seconds the accumulated inaccuracy can be approximated by the total number of ticks between zero and 16. This value is programmed in the fractional calibration segment of the calibration register. After 16 seconds, the fraction counter starts incrementing from zero to this value. During the time the fraction counter is incrementing, the clear signal to the tick counter stays asserted. As a result, the tick counter increments are delayed by the value of ticks every 16 seconds.
When the fraction comparator determines that the fraction counter value is equal to the maximum fractional calibration value, the fraction comparator releases the clear signal of the tick counter. This clear signal allows the fractional counter to start incrementing again. The fractional calibration register also includes an enable bit. When this bit is a 1, the fraction comparator performs the operations associated with fractional calibration, including the tick counter extend clear signal.