1. Field of the Invention
The present invention relates to a frequency locking structure, and more particularly to a frequency locking structure applied to phase-locked loops.
2. Background of the Related Art
Phase-locked loops (PLL) are generally used in the recovery of data and clock, modulation and demodulation of frequency and phase, and generation of stable and multiple clock. For the design of signal-mixing integrated circuits and system-on-chip (SOC), the phase-locked loops are necessary for the variable application requirement. Generally, there are specifications to be tested for the PLLs: lock-in range, lock-in time, and jitter.
The PLLs are capable of generating clean and stable clocks. However, the stability of the clocks may be deteriorated by noises and estimated by measuring the jitter outputted from the PLLs.
f_out=f_in*N/(M*P). (1)
wherein (N/(M*P)) is simplified to be a simple integer. Generally, the jitter from the PLL is derived from the design of the PLL 10, the reference input signal (ref_in) from the divider 12 and the jitter of the feedback signal (Vco_out) from the divider 16.
Accordingly, f_out is acquired by the formula (2) in condition that (fin) is 1 KHz and (f_out) is 12 MKHz as follows,
f_out=12 MKHz=1K*N/(M*P)
For example, M is 1 and P is 1 in the formula (2) to acquire N=12000. In the case, the reference input signal (ref_in) and the feedback signal (Vco_in) are identical, and f_in/M is 1 K. Thus, the reference input signal (ref_in) and the feedback signal (Vco_in) are 1 K, respectively. That is, (f_out) is 12 MKHz in conditions of ref_in =Vco_in =1K, M=P=1, and N=12000. The clock input of each (f_in) should generate the 1200 frequency output of (f_out). (Vco_in) and (ref_in) are identical in condition that N should be 12000. However, when the signal jitter of (f_in) is very high and the difference between (ref_in) and (Vco_out) is large, the clock jitter of (Vco_out) correspondingly increases.
The present invention is directed to a frequency locking structure applied to phase locked loops to reduce the difference between a reference input signal and an oscillator output signal for improving the jitter of the reference input signal and reducing the clock jitter of the oscillator output signal.
The present invention is directed to a frequency locking structure applied to phase locked loops using a common factor as a reduction ratio for the reduction of difference between the reference input signal (ref_in) and the oscillator output signal (Vco_out). The jitter of the reference input signal is improved and the clock jitter of the oscillator output signal is reduced to prevent erroneous input reference.
The present invention is directed to a frequency locking structure applied to phase locked loops to compare a clock signal and an input signal for the dynamic calculation of a common factor available in various condition. Furthermore, the present invention is directed to a frequency locking structure applied to phase locked loops to provide a circuit with the frequency locking structure having enhanced signal-to-noise ratio (SNR).
Accordingly, the present invention provides a frequency locking structure applied to phase-locked loops, which includes: a first divider for receiving an input signal and dividing the input signal by a first constant factor; a multiplexer for coupling the first divider receiving a signal from the first divider and multiplying the signal from the first divider together with a common factor to output a reference input signal; a phase-locked loop for coupling the multiplexer, and receiving the reference input signal and a feedback signal to output an oscillator output signal; a second divider coupled the phase-locked loop, receiving the oscillator output signal and dividing the oscillator output signal by a second constant factor to acquire the feedback signal, and transmitting the feedback signal into the phase-locked loop; and a third divider, coupling the phase-locked loop, receiving the oscillator output signal and dividing the oscillator output signal by a third constant factor to output an output signal.
Accordingly, the present invention provides a frequency locking circuit applied to phase-locked loops, which includes: a first divider for receiving an input signal and dividing the input signal by a first constant factor to output a signal; a counter coupled the first divider for receiving the signal from the first divider and a clock signal to output a count; a maximum common factor calculating device, coupled the counter to receive the count and a number generator to receive a number, and calculating a maximum common factor of the count and the number to output a clock adjustment and a feedback adjustment; a second divider, for receiving the clock signal, coupled the maximum common factor calculating device to receive the clock adjustment, and dividing the clock signal by the clock adjustment to acquire a reference input signal; a phase-locked loop coupled the second divider for receiving the reference input signal, receiving a feedback signal to output an oscillator output signal; a fourth divider coupled the phase-locked loop for receiving the oscillator output signal, dividing the oscillator output signal by the feedback adjustment to acquire a feedback signal, and transmitting the feedback signal into the phase-locked loop; and a fifth divider, coupled to the phase-locked loop for receiving the oscillator output signal and dividing the oscillator output signal by a second constant factor to output an output signal.
The invention is directed to a frequency locking structure applied to phase-locked loops. The difference between a reference input signal (ref_in) and an oscillator output signal (Vco_out) is reduced by utilizing a common factor to prevent the oscillator output signal (Vco_out) from clock jitter.
In an aspect of the invention, the difference between the oscillator output signal (Vco_out) and the reference input signal (ref_in) and the amount of jitter of the reference input signal (ref_in) may be reduced without changing the oscillator output signal (Vco_out). The reduction of clock jitter may prevent erroneous output frequency. First, the common factor (ω) is used as a reduction ratio and a formula (2) is derived from the formula (1):
The common factor (ω) is led in the formula (2) in the case that f_out, M, P are not changed. Thus, a new input signal f_in′=f_in*ω and constant factor N′=N/ω are derived. The reduction of difference between the input signal f_in′=f_in*ω and the oscillator output signal (Vco_out) can reduce the clock jitter of the oscillator output signal (Vco_out).
In view of the formula (2) associated with the aforementioned circuit, N/(M*P) is a simple number and (f_in*ω) is acquired from the clock signal (fclk). The clock signal (fclk) is used to count the input signal (f_in) to acquire the count R, as a result of the clock signal (fclk) of a frequency faster than the input signal (fin) and a period with jitter smaller than the input signal (f_in). Their relationship may be expressed by the following formula (3):
That is, f_in is equal to (fclk/R).
Furthermore, for the maximum common factor calculating device 37, the count R and the number N are acquired to generate the maximum common factor GCD (R,W)=ω for the requirement of an (N/ω) of integer. Thus, ω meets the integer requirement of the ratio of (N/ω) and (R/ω), respectively. Next, (f_in*ω) is acquired by the formula (4):
f_in*ω=(fclk/R)*w=fclk/(R/ω) (4)
The formula (5) may be derived from the formula (2):
f_out=(fclk/(R/ω))*((N/ω)/(M*P)) (5)
In a practical application, processing of audio signal which is sensitive to the clock jitter is taken as an example. The system is implemented by setting f_in=1 Khz, f_out=12 MHz, M=1, P=1 and N=12000. The results are indicated as follows.
SNR=56.12 dB when ω is 1.
SNR=60.97 dB when ω is 2.
SNR=64.80 dB when ω is 3.
SNR=65.56 dB when ω is 4.
SNR=67.66 dB when ω is 8.
Accordingly, the increase of ω enhances SNR.
The frequency locking structure of the invention applied to a circuit in practice, the optimized value of ω is dynamically acquired for various systemic conditions with the combination of f_in, f_out or fclk. Thus, the acquisition of ω is available for the various conditions, for example but not limited to, f_in =2K or 0.5K respectively in combination of f_out=24M and 40 MHz.
Although the present invention has been explained in relation to its preferred embodiment, it is to be understood that other modifications and variation can be made without departing the spirit and scope of the invention as hereafter claimed.