阅读说明：本技术 一种无参考时钟的时钟数据恢复电路 (Clock data recovery circuit without reference clock ) 是由 王腾佳 徐豪杰 李浩明 李国儒 于 2020-07-01 设计创作，主要内容包括：本发明涉及一种无参考时钟的时钟数据恢复(CDR)电路,所述电路的一个实施方式包括：第一鉴相器、第二鉴相器、鉴频器、译码器、电荷泵、二阶环路滤波器和正交压控振荡器；其中,第一鉴相器、第二鉴相器、鉴频器组成鉴频鉴相器,鉴频鉴相器、译码器、电荷泵、二阶环路滤波器和正交压控振荡器依次连接,正交压控振荡器向鉴频鉴相器进行反馈。该实施方式将电路的鉴频过程与鉴相过程完全分开从而使CDR电路具有更高的鲁棒性。(The invention relates to a Clock Data Recovery (CDR) circuit without a reference clock, one embodiment of the circuit comprises: the device comprises a first phase discriminator, a second phase discriminator, a frequency discriminator, a decoder, a charge pump, a second-order loop filter and an orthogonal voltage-controlled oscillator; the first phase discriminator, the second phase discriminator and the frequency discriminator form a frequency discrimination phase discriminator, the decoder, the charge pump, the second-order loop filter and the orthogonal voltage-controlled oscillator are sequentially connected, and the orthogonal voltage-controlled oscillator feeds back to the frequency discrimination phase discriminator. This embodiment completely separates the phase frequency discrimination process from the phase discrimination process of the circuit to make the CDR circuit more robust.)

1. A clock-data recovery circuit without a reference clock, comprising: the device comprises a first phase discriminator, a second phase discriminator, a frequency discriminator, a decoder, a charge pump, a second-order loop filter and an orthogonal voltage-controlled oscillator; wherein the content of the first and second substances,

an input signal to be recovered respectively enters a clock input end of a first phase discriminator and a clock input end of a second phase discriminator, a first clock signal fed back by a quadrature voltage-controlled oscillator enters a data input end of the first phase discriminator, a second clock signal fed back by the quadrature voltage-controlled oscillator and orthogonal to the first clock signal enters a data input end of the second phase discriminator, the first phase discriminator outputs a phase discrimination signal by comparing the phases of the input signal and the first clock signal, and the second phase discriminator outputs a phase comparison signal by comparing the phases of the input signal and the second clock signal;

the phase detection signal enters a clock input end of a frequency discriminator, the phase comparison signal enters a data input end of the frequency discriminator, and the frequency discriminator uses the edge of the phase detection signal to sample the phase comparison signal to generate the frequency discrimination signal;

the decoder generates a decoding differential signal according to the input frequency discrimination signal, the input phase discrimination signal and preset logic and sends the decoding differential signal to the charge pump, the charge pump generates a current signal according to the input decoding differential signal, the current signal forms a control signal after being filtered by the second-order loop filter, and the orthogonal voltage-controlled oscillator outputs a first clock signal and a second clock signal under the action of the control signal.

2. The circuit of claim 1, wherein the phase frequency signal and the phase detection signal are both differential signals; the phase frequency signals comprise first port phase frequency signals and second port phase frequency signals, and the phase frequency signals comprise first port phase detection signals and second port phase detection signals.

3. The circuit of claim 2, wherein the decoder comprises: the first amplifier, the second amplifier, the first NAND gate and the second NAND gate; wherein the content of the first and second substances,

the negative input end of the first amplifier is connected with the positive input end of the second amplifier, and the output end of the first amplifier is connected with the first input end of the first NAND gate;

the output end of the first NAND gate is connected with the first input end of the second NAND gate, and the output end of the second amplifier is connected with the second input end of the second NAND gate;

the positive input end of the first amplifier is used for accessing a first port phase-frequency discrimination signal, the negative input end of the second amplifier is used for accessing a second port phase-frequency discrimination signal, and the second input end of the first NAND gate is used for accessing a first port phase-discrimination signal and a second port phase-discrimination signal.

4. The circuit of claim 3, wherein the first amplifier and the second amplifier each comprise: the circuit comprises a first triode, a second triode, a first resistor, a second resistor and a first current source.

5. The circuit of claim 3, wherein the first NAND gate and the second NAND gate each comprise: the third triode, the fourth triode, the fifth triode, the sixth triode, the seventh triode, the third resistor, the fourth resistor and the second current source.

6. The circuit of claim 3, wherein the predetermined logic of the decoder is to: decoding the positive side signal in the differential signal into an OR operation result of the first result and the second result; wherein the content of the first and second substances,

the first result is the and operation result of the first port frequency discrimination signal and the first port phase discrimination signal, and the second result is the non-operation result of the second port frequency discrimination signal.

7. The circuit of claim 1, wherein the charge pump comprises: the current source circuit comprises an eighth triode, a ninth triode, a thirteenth triode, an eleventh triode, a third current source, a fourth current source and a fifth current source.

8. The circuit of claim 1, wherein the second order loop filter comprises a first capacitor, a second capacitor, and a fifth resistor connected in series.

9. The circuit of claim 8, wherein the first capacitance has a larger capacitance value than the second capacitance.

10. The circuit of any of claims 1-9, wherein the input signal is a non-return-to-zero NRZ signal.

Technical Field

The invention relates to the technical field of electronics, in particular to a clock data recovery circuit without a reference clock.

Background

The Clock Data Recovery (CDR) without a reference Clock is applied to high-speed interface communication, and the Data is re-sampled and then driven to be output, so that the transmission error rate caused by high-frequency Clock jitter, power supply noise and channel reflection or attenuation is reduced and improved. CDR architectureThe method is characterized by comprising a BBPD (nonlinear phase detector), an L initial PD (linear phase detector) and the like, wherein the most classical and common CDR architecture is proposed by Pottbaker in 1992, as shown in figure 1, the basic principle is that a high-speed trigger (PD in figure 1) is used for phase detection, two clock signals with orthogonal phases are used for frequency detection, and then the frequency detection signal and the phase detection signal are added to obtain a control voltage (V in figure 1)_c) The oscillator (VCO in fig. 1) is controlled to change frequency and phase.

In fig. 1, data in NRZ represents an input signal to be restored, NRZ represents Non return to zero Code (Non return zero Code), PDs (C L, D, Q1 are a clock input terminal, a data input terminal and an output terminal, respectively), QPDs (C L, D, Q2 are a clock input terminal, a data input terminal and an output terminal, respectively), FDs are Phase detectors (C L, D, Q3 are a clock input terminal, a data input terminal and an output terminal, respectively), PDs, QPDs and FDs may be collectively called Phase/Frequency detectors, a signal output from Q1 is a Phase detection signal (also called a Q1 signal), a signal output from Q3 is a Phase detection signal (also called a Q3 signal), a Filter adds a Q1 signal and a Q3 signal and performs low-pass filtering, and T/4 represents a delay line of 1/4 clock period.

The academic world and the industrial world have made continuous improvements on the above-mentioned architecture, such as an optical communication resampling chip published in the japanese fuji laboratory in 2015.

In current CDR framework, the conversion circuit of frequency discrimination signal, phase discrimination signal to control voltage is comparatively simple, directly adds frequency discrimination signal and phase discrimination signal, and then generates the control voltage signal, but current framework requires that the current mirror of frequency discrimination signal and phase discrimination signal control matches well, otherwise can arouse extra shake, worsens the performance of CDR, causes locking error even. The reason is explained below.

Taking fig. 1 as an example, the relative phase relationship (i.e., edge relationship) between data (i.e., input signal) and two quadrature clock signals (i.e., clock signal CK directly output from VCO and clock signal CKQ obtained through delay line) can be divided into four kinds, i.e., four intervals of ①, ②, ③, and ④ shown in fig. 2.

Fig. 2 is a schematic diagram of waveforms of an input signal and a quadrature clock signal according to an embodiment of the present invention, wherein ①, ②, ②, ② are a first interval, a second interval, a third interval, and a fourth interval, respectively, and a length of each interval is 1/4 of a clock signal period, specifically, assuming that a frequency of the clock signal is greater than that of the input signal DATA (i.e., DATA in NRZ in fig. 1), the DATA edge moves to the right with respect to the CK edge, due to the effect of the phase detection loop, Q1 is at a low level (hereinafter, represented by-1) in ②, ② intervals, and at ②, 365 intervals, Q6329 is at a high level (hereinafter, represented by 1) in the ①, ② intervals, and at ④ intervals, Q1 is at a middle level (i.e., at a middle level, which has no effect on the CDR loop), it can be seen that, when frequency locking is performed, a sum of Q1 and Q5 is at zero in the interval ②, when frequency locking is performed, at ③, when phase locking is performed, a phase locking is performed at a point 3, and Q is only indicated by an arrow that the last, the clock signal is controlled by a lower arrow of the clock signal, thus, and the last, the clock signal is indicated by a lower arrow indicated by zero arrow indicated by — indicated.

The key point of this architecture is that the frequency-to-phase conversion process must ensure that Q3 goes from 1 or-1 to zero in the architecture shown in fig. 1, when the relative frequency difference between the DATA and the clock is zero, the phase of DATA is theoretically only possible in the ③ or ④ interval (when the initial frequency of the clock signal is greater than the input signal frequency), because the Q1 and Q3 signals add to zero in the ①, ② interval, when the frequency discrimination loop does not act, and no frequency adjustment is performed, however, in practice, because Q1 and Q3 may generate mismatch due to the presence of excessive phase, the sum of Q1 and Q3 cannot return to zero in the ①, ② interval, leading to earlier frequency locking in the ①, ② interval, then the frequency adjustment continues in the ③, ④ interval due to the sum of Q1 and Q3 not being zero, so that the frequency locking cannot end, leading to Q3 being between 1 and CDR-1, it can be seen that the performance of the frequency discrimination signal is degraded due to the requirement that the frequency discrimination signal does not match well with the phase difference, otherwise it may be difficult to meet the actual high frequency detection requirement.

In view of the above disadvantages, it is desirable to provide a CDR circuit with more robustness.

Disclosure of Invention

The technical problem to be solved by the invention is as follows: a CDR circuit having higher robustness without adding a phase frequency detection signal to a phase detection signal is provided.

In order to solve the above technical problem, the present invention provides a clock data recovery circuit without a reference clock.

The clock data recovery circuit without the reference clock of the embodiment of the invention can comprise: the device comprises a first phase discriminator, a second phase discriminator, a frequency discriminator, a decoder, a charge pump, a second-order loop filter and an orthogonal voltage-controlled oscillator; the input signal to be recovered respectively enters a clock input end of a first phase discriminator and a clock input end of a second phase discriminator, a first clock signal fed back by an orthogonal voltage-controlled oscillator enters a data input end of the first phase discriminator, a second clock signal fed back by the orthogonal voltage-controlled oscillator and orthogonal to the first clock signal enters a data input end of the second phase discriminator, the first phase discriminator outputs a phase discrimination signal by comparing the phases of the input signal and the first clock signal, and the second phase discriminator outputs a phase comparison signal by comparing the phases of the input signal and the second clock signal; the phase detection signal enters a clock input end of a frequency discriminator, the phase comparison signal enters a data input end of the frequency discriminator, and the frequency discriminator uses the edge of the phase detection signal to sample the phase comparison signal to generate the frequency discrimination signal; the decoder generates a decoding differential signal according to the input frequency discrimination signal, the input phase discrimination signal and preset logic and sends the decoding differential signal to the charge pump, the charge pump generates a current signal according to the input decoding differential signal, the current signal forms a control signal after being filtered by the second-order loop filter, and the orthogonal voltage-controlled oscillator outputs a first clock signal and a second clock signal under the action of the control signal.

Preferably, the phase frequency detection signal and the phase detection signal are both differential signals; the phase frequency signals comprise first port phase frequency signals and second port phase frequency signals, and the phase frequency signals comprise first port phase detection signals and second port phase detection signals.

Preferably, the decoder comprises: the first amplifier, the second amplifier, the first NAND gate and the second NAND gate; the negative input end of the first amplifier is connected with the positive input end of the second amplifier, and the output end of the first amplifier is connected with the first input end of the first NAND gate; the output end of the first NAND gate is connected with the first input end of the second NAND gate, and the output end of the second amplifier is connected with the second input end of the second NAND gate; the positive input end of the first amplifier is used for accessing a first port phase-frequency discrimination signal, the negative input end of the second amplifier is used for accessing a second port phase-frequency discrimination signal, and the second input end of the first NAND gate is used for accessing a first port phase-discrimination signal and a second port phase-discrimination signal.

Preferably, the first amplifier and the second amplifier each include: the circuit comprises a first triode, a second triode, a first resistor, a second resistor and a first current source.

Preferably, the first nand gate and the second nand gate each include: the third triode, the fourth triode, the fifth triode, the sixth triode, the seventh triode, the third resistor, the fourth resistor and the second current source.

Preferably, the preset logic of the decoder is: decoding the positive side signal in the differential signal into an OR operation result of the first result and the second result; the first result is the and operation result of the first port frequency discrimination signal and the first port phase discrimination signal, and the second result is the non-operation result of the second port frequency discrimination signal.

Preferably, the charge pump comprises: the current source circuit comprises an eighth triode, a ninth triode, a thirteenth triode, an eleventh triode, a third current source, a fourth current source and a fifth current source.

Preferably, the second-order loop filter includes a first capacitor, a second capacitor and a fifth resistor connected in series in sequence.

Preferably, the capacitance value of the first capacitor is larger than that of the second capacitor.

Preferably, the input signal is a non-return-to-zero NRZ signal.

The technical scheme of the invention has the following advantages: the method is characterized in that the traditional CDR framework is innovated, a decoder is additionally arranged between a phase frequency detector (comprising a first phase detector, a second phase detector and a phase frequency detector) and a charge pump and used for decoding a phase frequency signal and a phase detection signal output by the phase frequency detector, the internal logic of the decoder is used for completely separating the phase frequency process and the phase detection process of the CDR, namely the phase frequency process is irrelevant to the phase detection signal and the phase detection process is irrelevant to the phase frequency signal, and therefore the defects that the phase frequency signal and the phase detection signal need to be added and well matched (but the addition is difficult to realize in practice) or the frequency cannot be locked in the prior art are overcome. In the embodiment of the invention, the addition of the phase frequency detection signal and the phase detection signal is avoided, and the phase frequency detection signal and the phase detection signal do not need to be well matched, so that the frequency locking cannot be finished due to the possible excessive phase.

Drawings

FIG. 1 is a schematic diagram of a prior art Pottbacker-type CDR architecture;

FIG. 2 is a schematic diagram of waveforms of an input signal and a quadrature clock signal in an embodiment of the present invention;

FIG. 3 is a schematic diagram of a clock data recovery circuit without a reference clock according to an embodiment of the present invention;

FIG. 4 is a schematic diagram of a clock data recovery circuit without a reference clock in an embodiment of the invention;

FIG. 5 is a schematic diagram of a decoder in an embodiment of the invention;

FIG. 6 is a schematic diagram of the first amplifier or the second amplifier in an embodiment of the invention;

FIG. 7 is a schematic diagram of a first NAND gate or a second NAND gate according to an embodiment of the present invention;

FIG. 8 is a schematic diagram of a charge pump in an embodiment of the invention;

FIG. 9 is a schematic diagram of behavior level simulation effects of an embodiment of the present invention;

FIG. 10 is a schematic diagram illustrating simulation effects of control voltage signals according to an embodiment of the present invention;

FIG. 11 is an eye diagram illustration of sampled output data in accordance with an embodiment of the present invention.

Detailed Description

In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all, embodiments of the present invention. All other embodiments, which can be obtained by a person skilled in the art without any inventive step based on the embodiments of the present invention, are within the scope of the present invention.

Fig. 3 is a schematic diagram of an architecture of a clock data recovery circuit without a reference clock in an embodiment of the present invention, fig. 4 is a schematic diagram of a principle of the clock data recovery circuit without the reference clock in the embodiment of the present invention, and as shown in fig. 3 and 4, the clock data recovery circuit without the reference clock in the embodiment of the present invention may include a first phase detector PD1, a second phase detector PD2, a frequency detector FD, a decoder, a charge pump CP, a second-order loop Filter L oop Filter, and a quadrature voltage controlled oscillator qvco, where the PD in fig. 3 includes a PD1 and a PD 2.

In the working process, an input signal DATA to be recovered respectively enters a clock input end of the first phase detector PD1 and a clock input end of the second phase detector PD2, a first clock signal CK fed back by the quadrature voltage controlled oscillator QVCO enters a DATA input end of the first phase detector, and a second clock signal CKQ fed back by the quadrature voltage controlled oscillator QVCO and orthogonal to the first clock signal CK enters a DATA input end of the second phase detector PD 2. The first phase detector PD1 outputs a phase detection signal PH by comparing the phase of the input signal DATA with the phase of the first clock signal CK, and the second phase detector PD2 outputs a phase comparison signal by comparing the phase of the input signal DATA with the phase of the second clock signal CKQ. It is to be understood that in the above process, the first phase detector PD1 samples the first clock signal CK using the input signal DATA to obtain the phase detection signal PH, and the second phase detector PD2 samples the second clock signal CKQ using the input signal DATA to obtain the phase comparison signal. Preferably, the input signal may be an NRZ signal.

In the practical application of the phase detection signal, the phase detection signal PH enters a clock input end of a phase frequency detector FD, the phase detection signal PH enters a data input end of the phase frequency detector PD, the phase frequency detector uses an edge of the phase detection signal to sample the phase detection signal to generate a phase detection signal FR., and the phase frequency detector PD1, the second phase detector PD2 and the phase frequency detector FD can all be realized based on a high-speed trigger (such as a D trigger), so that a phase detection loop and the phase detection loop can be formed.

The frequency discrimination signal FR and the phase discrimination signal PH enter a decoder, and the decoder generates a decoding differential signal (in which a positive terminal signal is UPP and a negative terminal signal is UPN) according to the input frequency discrimination signal FR, the input phase discrimination signal PH and a preset logic and sends the decoding differential signal to the charge pump CP. As a preferred solution, the logic may be: decoding a positive end signal UPP in the differential signal into an OR operation result of the first result and the second result; the first result is an and operation result of the first port frequency discrimination signal FRP and the first port frequency discrimination signal PHP, and the second result is a non-operation result of the second port frequency discrimination signal FRN, that is, UPP = FRP + PHP + FRN ', it can be understood that + represents an and operation, + represents an or operation,' represents a non-operation. The truth table for this logic is shown below (where # denotes arbitrary values):

FR=1	PH=#	UPP=1
			FR=0	PH=1	UPP=1
FR=0	PH=0	UPP=0
			FR=-1	PH=#	UPP=0

it can be seen that when FR is 1 or-1, the frequency discrimination signal FR controls the loop, the circuit is in the frequency locking state, and the value of the phase discrimination signal PH does not work at all at this time; when FR is zero, the phase discrimination signal PH controls the loop, the circuit is in a phase-locked state, and the phase discrimination signal FR does not work at this time. Therefore, the invention realizes the complete separation of the frequency locking process and the phase locking process through the decoder, avoids the defect that the frequency locking signal FR and the phase locking signal PH need to be added to require the frequency locking signal FR and the phase locking signal PH to be well matched in the prior art, ensures that the circuit cannot lock the frequency due to the existence of excessive phases, is beneficial to reducing extra jitter and improving the CDR performance. In addition, through the arrangement, six states (namely six states formed by combining three states of 1, -1 and zero FR and two states of 1-1 and PH) in the frequency and phase locking process in the conventional Pottbacker architecture can be simplified into four states in the truth table, and the frequency and phase locking logic of the CDR circuit is simplified.

Fig. 5 is a schematic diagram of a decoder in an embodiment of the invention, with which the above logic can be implemented. As shown in fig. 5, the decoder includes: a first amplifier U1, a second amplifier U2, a first nand gate G1, and a second nand gate G1. The negative input end of the first amplifier U1 is connected with the positive input end of the second amplifier U2, and the output end of the first amplifier U1 is connected with the first input end of the first NAND gate G1; the output of the first nand gate G1 is connected to a first input of the second nand gate G2, and the output of the second amplifier U2 is connected to a second input of the second nand gate G2. It can be seen that the positive input terminal of the first amplifier U1 is used for receiving the first port frequency-discrimination signal FRP, the negative input terminal of the second amplifier U2 is used for receiving the second port frequency-discrimination signal FRN, the second input terminal of the first nand gate G1 is used for receiving the first port frequency-discrimination signal PHP and the second port frequency-discrimination signal PHN, and the output terminal of the second nand gate G2 can output the decoding differential signals UPP and UPN. In the embodiment of the present invention, Vcm denotes a reference level.

Fig. 6 is a schematic diagram of the first amplifier or the second amplifier in the embodiment of the present invention. As shown in fig. 6, the first amplifier U1 or the second amplifier U2 may include: the circuit comprises a first triode T1, a second triode T2, a first resistor R1, a second resistor R2 and a first current source I1. One end of the first resistor R1 is connected to one end of the second resistor R2, the other end of the first resistor R1 is connected to a collector of the first transistor T1, the other end of the second resistor R2 is connected to a collector of the second transistor T2, the first transistor T1 is connected to an emitter of the second transistor T2 and is connected to one end of the first current source I1, and the other end of the first current source I1 is grounded. The bases of the first transistor T1 and the second transistor T2 are the positive input terminal (indicated by vin +) and the negative input terminal (indicated by vin-) of the amplifier, respectively, and vout + and vout-indicate the output terminals of the amplifier in fig. 6.

Fig. 7 is a schematic diagram of a first nand gate or a second nand gate in the embodiment of the present invention, and as shown in fig. 7, the first nand gate G1 or the second nand gate G2 may include: the driving circuit comprises a third triode T3, a fourth triode T4, a fifth triode T5, a sixth triode T6, a seventh triode T7, a third resistor R3, a fourth resistor R4 and a second current source I2. One end of the third resistor R3 is connected to one end of the fourth resistor R4, and is also connected to the collector of the fifth transistor T5. The other end of the third resistor R3 is connected to the collector of the third transistor T3, and the other end of the fourth resistor R4 is connected to the collector of the fourth transistor T4. The third transistor T3 is connected to the emitter of the fourth transistor T4 and to the collector of the sixth transistor T6. An emitter of the fifth triode T5 is connected to a collector of the seventh triode T7, the sixth triode T6 is connected to an emitter of the seventh triode T7, and is also connected to one end of the second current source I2, and the other end of the second current source I2 is grounded. In fig. 7, A, A ' is the first input of the nand gate, B, B ' is the second input of the nand gate, and Y, Y ' is the output of the nand gate. It can be seen that the first nand gate G1 or the second nand gate G2 both employ NPN transistors.

The decoder realized by the circuit is a high-speed decoder and can work in a high-frequency environment of 25 GHz. It will be appreciated that the above decoder logic, decoder circuit, amplifier circuit, and nand gate circuit are preferred implementations and are not intended to limit any other possible implementations.

The decoded differential signals UPP and UPN output from the decoder enter the charge pump CP, which can generate a current signal according to the input decoded differential signals UPP and UPN. In practical application, the charge pump CP can be constructed by using an NPN transistor as a switching tube, thereby supporting high-frequency operation. Fig. 8 is a schematic diagram of a charge pump in an embodiment of the present invention, and as shown in fig. 8, the charge pump CP may include: an eighth transistor T8, a ninth transistor T9, a thirteenth diode T10, an eleventh transistor T11, a third current source I3, a fourth current source I4, and a fifth current source I5. The eighth transistor T8 is connected to the collector of the ninth transistor T9, and is connected to one end of the third current source I3 and one end of the fourth current source I4, and the other end of the fourth current source I4 is grounded; the thirteenth diode T10 is connected to the collector of the eleventh transistor T11; the eighth transistor T8 is connected to the emitter of the eleventh transistor T11, and is also connected to one terminal of the fifth current source I5, and the other terminal of the fifth current source I5 is grounded. The bases of the eighth transistor T8 and the thirteenth transistor T10 are used for inputting the UPP signal, the bases of the ninth transistor T9 and the eleventh transistor T11 are used for inputting the UPN signal, and vout indicates the output terminal of the charge pump. It is understood that the above is only one implementation circuit selectable for the charge pump, and does not form any limitation on other implementation circuits.

The current signal output by the charge pump CP forms a control signal (i.e., a control voltage signal) after passing through the second-order loop Filter L oop Filter, and the quadrature voltage controlled oscillator QVCO outputs the first clock signal CK and the second clock signal CKQ under the action of the control signal and respectively feeds back to the first phase detector PD1 and the second phase detector PD 2. the structure of the second-order loop Filter L oop Filter can be seen in fig. 3, and it can be seen that the second-order loop Filter L oop Filter can include a first capacitor C1, a second capacitor C2, and a fifth resistor (i.e., a resistor in the L oop Filter in fig. 3) which are connected in series in sequence, and when the Filter is designed, the capacitance value of the first capacitor C1 can be much larger than that of the second capacitor C2 (which is much larger than that can be quantized according to actual requirements), for example, the capacitance value of the first capacitor C1 is 16 times that of the second capacitor C2.

In general, the second order loop Filter L oop Filter is a low pass Filter, the long term slew rate of which is determined by the first capacitor C1, and the short term slew rate of which is determined by the second capacitor C2, the long term slew rate can determine the change rate and the change direction of the signal phase in a long term, and the short term slew rate can determine the change rate and the change direction of the signal phase in a short term.

In the frequency locking process of the clock data recovery circuit without the reference clock according to the embodiment of the present invention, taking the frequency of the clock signal (CK and CKQ) as an example that is greater than the frequency of the input signal, in the interval ① and ② shown in fig. 2, the frequency discrimination signal FR is 1, the UPP signal output by the decoder is 1, and due to the action of the frequency discrimination loop, the frequency of the clock signal output by the quadrature voltage controlled oscillator QVCO will decrease, in the interval ③ and ④, the frequency discrimination signal FR is zero, the phase discrimination signal PH is 1, the UPP signal output by the decoder is 1, and the frequency of the clock signal output by the quadrature voltage controlled oscillator QVCO will also decrease.

In the conversion process from frequency locking to phase locking, taking the clock signal frequency greater than the input signal frequency as an example, if the frequency locking process ends in the interval ③ or ④, the conversion process is the same as the existing architecture, if the frequency locking process ends in ① or ②, the possible excessive phase difference can be filtered out by the second-order loop Filter L oop Filter, therefore, even if the frequency locking process ends in ① or ②, the input signal phase still converges to make the frequency discrimination signal FR zero.

FIG. 9 is a schematic diagram of behavior-level simulation effects of an embodiment of the present invention, where in FIG. 9, the abscissa is time and the unit is microseconds; the ordinate is the voltage in volts. The clock signal frequency (initial frequency) is 25GHz, the input signal frequency offset is 100MHz, and the simulation results of the frequency discrimination signal FR and the phase discrimination signal PH in the simulation locking process and the control voltage signal Vout _ CP output by the second-order loop filter are shown in fig. 9. It can be seen that the initial state of the loop is a frequency locking stage, and the charge pump charges the second-order loop filter to enable the constant slew rate of Vout _ CP to rise; when the frequency of the clock signal is close to that of the input signal, the loop switches from the frequency-locked state to the phase-locked state, an excessive phase occurs in this embodiment, and Vout _ CP exhibits a waveform that gradually converges to oscillation by the excessive phase locking. The loop finally enters a phase-locking stage, the Vout _ CP and the PH signals present periodic characteristics, and the period length and the variation amplitude of the Vout _ CP are related to loop parameters; the FR signal goes to zero, which can be used as a lock indication signal for the loop. It can be seen that the CDR circuit of the embodiment of the present invention can also successfully lock the frequency even if there is an excessive phase.

Fig. 10 is a schematic diagram of simulation effects of a control voltage signal according to an embodiment of the present invention, and fig. 11 is a schematic diagram of an eye diagram of sampled output data according to an embodiment of the present invention, where both simulations are circuit-level simulations. In fig. 10, the abscissa is time in microseconds; the ordinate is the voltage in volts. In fig. 11, the sampled output data refers to data sampled and output from an input signal by using a clock signal recovered from the input signal, and the abscissa is time in picoseconds; the ordinate is the voltage in millivolts. In the circuit level simulation, the initial frequency of the clock signal is designed to be 25GHz, the input data has 1MHz sinusoidal jitter, and the CDR circuit of the embodiment of the invention works well as can be seen from FIGS. 10 and 11.

In summary, in the technical solution of the embodiment of the present invention, an existing CDR architecture is innovated, a decoder is added between a phase frequency detector and a charge pump to decode a phase frequency signal and a phase detection signal output by the phase frequency detector, and the internal logic of the decoder completely separates the phase frequency process and the phase detection process of the CDR, that is, the phase frequency process is unrelated to the phase detection signal and the phase detection process is unrelated to the phase frequency signal, so as to solve the problem that the phase frequency signal and the phase detection signal need to be added and well matched, otherwise the frequency cannot be locked in the prior art. In the embodiment of the invention, the addition of the phase frequency detection signal and the phase detection signal is avoided, and the phase frequency detection signal and the phase detection signal do not need to be well matched, so that the frequency locking cannot be finished due to the possible excessive phase.

Finally, it should be noted that: the above examples are only intended to illustrate the technical solution of the present invention, but not to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.