DETAILED DESCRIPTION OF EMBODIMENTS
[0029] Various embodiments of the present invention will be described with reference to the accompanying drawings. It is to be noted that the same or similar reference numerals are applied to the same or similar parts and elements throughout the drawings, and the description of the same or similar parts and elements will be omitted or simplified.
[0030] FIG. 1 is a block diagram showing a receiver according to an embodiment of the present invention. The receiver includes a data read circuit 6 and a phase interpolator 7 . The phase interpolator 7 receives four-phase (0, π/2, π, 3π/2) clock signals (Clock_In, CML) 8 and provides four-phase clock signals (Reclock_In) 9 whose phases are synchronous with those of serial data 2 ′. The serial data 2 ′ is sent from a transmitter (not shown) to the data read circuit 6 . The data read circuit 6 reads the serial data 2 ′ based on the clock signals 9 , provides the phase interpolator 7 with a phase information signal (UP/DN) 10 containing information about the phases of the clock signals and serial data 2 ′, and outputs data 11 , which is synchronous with clock signals in the receiver, and clock signals 9 ′, which have phases adjusted to those of the data 11 , to subsequent circuits for further processing.
[0031] FIG. 2 is a schematic view showing the phase interpolator 7 according to the embodiment. The phase interpolator 7 has an IDAC (switched current digital to analog converter) controller 51 to convert an input signal into a current output 56 , a mixer 52 to receive the output 56 of the IDAC controller 51 and the clock signals 8 , shift the phases of the clock signals 8 according to the output 56 , and provide differential clock signals, and an output circuit 53 to receive the outputs of the mixer 52 .
[0032] The phase information signal 10 from the data read circuit 6 (not shown in FIG. 2 ) is received by the IDAC controller 51 , which converts the signal 10 into the current signal 56 indicative of an up/down current value. The four-phase clock signals 8 are received by the mixer 52 , which mixes the clock signals 8 with the current signal 56 , shifts the phases of the clock signals 8 , and provides the phase-shifted clock signals to the output circuit 53 .
[0033] FIG. 3 shows the output circuit 53 according to the embodiment. The output circuit 53 has an integrator 62 to receive and modify the differential clock signals 61 whose phases have been adjusted to those of data, an output buffer 63 to amplify the outputs of the integrator 62 , a duty correction circuit 64 to receive the amplified signals from the output buffer 63 , adjust the phases of the received signals, and feed phase adjusted signals back to the output buffer 63 , and a controller 71 to control the modification carried out by the integrator 62 on the differential clock signals 61 . The controller 71 checks to see if the voltage of the output signals 61 from the mixer 52 is amplifiable by the output buffer 63 , and according to a result of the checking, changes the capacitance of the integrator 62 . The integrator 62 makes a signal incline so as to reduce the influence of phase shift by phase insertion. Namely, the integrator 62 makes a signal waveform more slewed. By changing the capacitance of the integrator 62 , it is possible to control the rising speed of each clock signal edge and restore the amplitude of each clock signal to a voltage level that is amplifiable by the output buffer 63 . The output circuit 53 handles one differential clock pair among the four-phase clock signals 61 . The phase interpolator 7 incorporates two systems, one for a differential clock pair of 0 and 180 degrees and the other for a differential clock pair of 90 and 270 degrees. Namely, in the phase interpolator 7 , a combination of the output circuit 53 , mixer 52 , and IDAC controller 51 is prepared for the differential clock pair of 0 and 180 degrees, and another combination of the same is prepared for the differential clock pair of 90 and 270 degrees. The following explanation is conducted only for one of them, and it should be understood that the same explanation is applicable to the other.
[0034] FIG. 4 is a circuit diagram showing examples of the controller 71 and integrator 62 according to a first embodiment of the present invention. The controller 71 has an operational amplifier 21 , a capacitor 12 , and an inverter 13 . The integrator 62 has an NMOS transistor 14 a connected to one ( 61 a ) of the differential clock signals, a capacitor 15 a connected to the NMOS transistor 14 a , an NMOS transistor 14 b connected to the other ( 61 b ) of the differential clock signals, and a capacitor 15 b connected to the NMOS transistor 14 b . The operational amplifier 21 finds a difference between the positive signal 61 a or Integp and the negative signal 61 b (an inverted signal due to the CML) or Integn. The operational amplifier 21 then compares the found differential voltage with a reference signal 17 or Vref. If the differential voltage is smaller than the reference signal Vref, the operational amplifier 21 provides a detected signal. Here, the capacitor 12 is able to prevent a switching of the inverter 13 upon first receiving a detected signal. Namely, only when the differential voltage is smaller than the reference signal Vref several times, does the output of the operational amplifier 21 make the voltage of a node 18 exceed a threshold of the inverter 13 , so that the inverter 13 may provide an OFF signal to a line 19 . The OFF signal turns off the NMOS transistors 14 a and 14 b so as to disconnect the capacitors 15 a and 15 b of the integrator 62 . As a result, the capacitance of the integrator 62 decreases, quickening the rise of each clock edge and restoring the amplitude of the clock signals. Here, capacitance means a function of accumulating charge and includes wiring load capacitance and input terminal capacitance.
[0035] According to the embodiment, the capacitor 12 accumulates a predetermined amount of charge. According to another embodiment, the capacitor 12 may be omitted. In this case, a single pulse of a detected signal makes the voltage of the node 18 exceed the threshold of the inverter 13 , thus providing an OFF signal to the line 19 . In this way, if it is necessary to invert a signal after one or a small number occurrences of the differential voltage becoming smaller than the reference signal Vref, the capacitor 12 can be omitted or replaced with a smaller one. If it is necessary to invert a signal after a larger number of times of such occurrence, the capacitor 12 can be replaced with a larger one.
[0036] FIG. 5 shows an example of the operational amplifier 21 of FIG. 4 . According to this example, the operational amplifier 21 has operational amplifiers 40 a and 40 b . Each of the operational amplifiers 40 a and 40 b finds a difference between a positive signal 41 and a negative signal 42 and compares the difference with a reference signal 43 . If any one of the differences “positive signal 41 —negative signal 42 ” and “negative signal 42 —positive signal 41 ” is smaller than the reference signal 43 , an exclusive OR gate 45 causes an inversion to provide a high output signal 46 . This results in accumulating charge in the capacitor 12 . If this consecutively occurs several times, the accumulated charge exceeds the threshold of the inverter 13 , which then provides an inverted signal.
[0037] FIG. 6 shows an example of one of the operational amplifiers 40 a and 40 b of FIG. 5 . An operational amplifier 81 compares the positive signal 41 with the negative signal 42 and provides a difference between them as a signal 83 . An operational amplifier 82 compares the signal 83 with the reference signal 43 . If the signal 83 is greater than the reference signal 43 , it is considered that the signals are amplifiable by the output buffer 63 , and the amplifier 82 provides a high output signal 84 , which is inverted by an inverter 85 into a low signal 44 . If the operational amplifier 82 determines that the signal 83 is lower than the reference signal 43 , it is determined that the signals are not amplifiable by the output buffer 63 . Then the operational amplifier 82 provides a low output signal 84 , which is inverted by the inverter 85 into a high output signal 44 .
[0038] FIG. 7 explains amplitude expansion according to the embodiment. In the controller 71 , the voltage of the node 18 may exceed the threshold of the inverter 13 , to provide an OFF signal to the line 19 . This OFF signal turns off the NMOS transistors 14 to disconnect the capacitors 15 of the integrator 62 . This reduces the capacitance of the integrator 62 , quickening the rise of each clock edge and restoring the amplitude of each clock signal. This controls changes exerted by the integrator 62 on the differential clock signals, expands amplitude (dot-and-dash lines) more than the related art (solid and broken lines), and enables all signals to be amplified by the output buffer 63 . Consequently, the embodiment allows the data read circuit 6 to correctly recognize all clock signals and correctly read data, prevents the destruction of duty ratios of clock signals due to phase shift in the clock signals, and ensures data read operation through phase adjustment between data and clock signals.
[0039] FIG. 8 is a circuit diagram showing examples ( 71 a , 62 a ) of the controller 71 and integrator 62 according to a second embodiment of the present invention. Compared with FIG. 4 , the controller 71 a has two operational amplifiers 21 a and 21 b that receive different reference signals Vref 1 and Vref 2 , respectively. The operational amplifier 21 a finds a difference between the signals Integp and Integn and compares the differential voltage with the reference signal Vref 1 . If the differential voltage is smaller than the reference signal Vref 1 , a predetermined amount of charge is accumulated in a capacitor 12 a . On the other hand, the operational amplifier 21 b finds a difference between the signals Integp and Integn and compares the differential voltage with the reference signal Vref 2 . If the differential voltage is smaller than the reference signal Vref 2 , a predetermined amount of charge is accumulated in a capacitor 12 b . Based on these different standards, the operational amplifiers 21 a and 21 b turn off NMOS transistors 14 and disconnect capacitors 15 of the integrator 62 a . For example, the reference signals Vref 1 and Vref 2 are set to Vref 1 >Vref 2 . The operational amplifier 21 b that receives the reference signal Vref 2 may have no capacitor. That results that an inverter 13 b inverts in response to a single input pulse. The capacitors 15 a and 15 b and the capacitors 15 c and 15 d may have different capacitance values. That results in differ load of the capacitance when the NMOS transistors 14 are turned off. According to the capacitance of the disconnected capacitors, the rising speed of each clock edge changes so as to control the degree of restoration of clock amplitude. This technique realizes precise control of clock amplitude.
[0040] FIG. 9 is a circuit diagram showing examples ( 71 b , 62 b ) of the controller 71 and integrator 62 according to a third embodiment of the present invention. The controller 71 b has an operational amplifier 22 a that receives the signals Integp, Integn, and Vref 1 , an operational amplifier 22 b for receiving the signals Integp, Integn, and Vref 2 , a delay circuit 25 for delaying the output of the operational amplifier 22 b , a flip-flop (FF) for receiving the output of the delay circuit 25 at a clock input terminal and the output of the operational amplifier 22 a at an input terminal, an inverter 26 for inverting the output of the delay circuit 25 , a flip-flop 24 for receiving the output of the inverter 26 at a clock input terminal and the output of the flop-flop 23 at an input terminal, and a NAND gate 27 for receiving the outputs of the flip-flops 23 and 24 and providing a NAND of the received signals. The controller 71 b is an example of a counter circuit. The operational amplifier 22 a receives the signals Integp and Integn ( 61 a and 61 b ), finds a difference between them, and compares the difference with the reference signal Vref 1 . If the difference is smaller than the reference signal Vref 1 , the operational amplifier 22 a provides an output signal 33 . Similarly, the operational amplifier 22 b finds a difference between the signals Integp and Integn and compares the difference with the reference signal Vref 2 . Here, the reference signal Vref 2 is set to have a voltage level that always generates clock signals. Accordingly, an output signal 34 from the operational amplifier 22 b generates clock signals. The operational amplifiers 22 a and 22 b may have the same structures as those of FIGS. 5 and 6 . The output signal 34 is slightly delayed by the delay circuit 25 so that the output signal 33 is properly received by the FF 23 . The signal 33 hit by a clock signal 35 from the delay circuit 25 is sent to the next FF 24 . The clock signal 35 is inverted by the inverter 26 , and a half-clock-shifted clock signal 36 hits the FF 24 , which provides a signal 38 . Signals 37 and 38 are passed to the NAND gate 27 . During a high period of the signal 33 , i.e., only when voltage variations are outside of an allowable range, does the NAND gate 27 provide a low signal 39 to turn off NMOS transistors 28 a and 28 b and disconnect capacitors 29 a and 29 b of the integrator 62 b . This reduces the capacitance of the integrator 62 b , quickening the rise of each clock edge and restoring clock amplitude. This embodiment employs two flip-flop stages so that a switching operation may take place only when the reference signal 31 has been exceeded at least two times. It is possible to employ three, four, or more flip-flop stages, so as to trigger a switching operation upon three or more occurrences of exceeding the reference signal. In this way, the configuration of FIG. 9 is capable of defining the number of occurrences necessary to trigger a switching operation. The structure of FIG. 9 can turn off the NMOS transistors 28 a and 28 b upon two occurrences of the difference between the signals Integp and Integn ( 61 a and 61 b ) being below the reference signal Vref 1 . It is thus possible to optionally set the number of occurrences of the difference between the signals Integp and Integn being lower than the reference signal Vref 1 to turn off the NMOS transistors 28 a and 28 b.
[0041] FIG. 10 shows another example ( 62 c ) of the integrator 62 . This example halves the capacitance of the integrator 62 c in an OFF operation, instead of completely disconnecting the capacitance. More precisely, one capacitor 15 e of the integrator 62 c is configured so as to be switched off with an NMOS transistor 14 e , and the other capacitor 14 f thereof is fixed. Even if the NMOS transistor 14 e is turned off so as to disconnect the capacitor 15 e , the capacitor 15 f remains active, thus restricting the rise of each clock edge.
[0042] Although this example employs only one switching element, more reference signals and more switching stages may be employed in order to precisely control the capacitance of the integrator. Voltage applied to the switching elements may analogously be controlled in response to a voltage difference between the signals 61 , in order to analogously control the capacitance of the integrator.
[0043] As explained above, the phase interpolator and receiver according to any one of the embodiments of the present invention determine, with a controller incorporated in the phase interpolator, whether or not the voltage of signals from a mixer is amplifiable by an output buffer, and according to a result of the determination, change the capacitance of an integrator incorporated in the phase interpolator. By changing the capacitance of the integrator, the phase interpolator controls the rising speed of each clock edge and restores the amplitude of clock signals so that the clock signals may have voltage levels amplifiable by the output buffer. This results in providing correct clock signals that correctly hit data in a data read circuit and ensure the normal operation of the receiver.
[0044] The present invention may be embodied in other specific forms without departing from the spirit or essential characteristics thereof. The embodiments are therefore to be considered in all respects as illustrative and not restrictive, the scope of the present invention being indicated by the appended claims rather than by the foregoing description, and all changes which come within the meaning and range of equivalency of the claims are therefore intended to be embraced therein.