Edge-triggered latch with symmetric complementary pass-transistor logic data path

The present application is related to the following co-pending U.S. patent applications: U.S. patent application Ser. No. 09/810,026 filed on Mar. 15, 2001, titled “Edge-Triggered Latch With Balanced Pass-Transistor Logic Trigger.” The above mentioned patent application is assigned to the assignee of the present invention. The content of the cross referenced co-pending application is hereby incorporated herein by reference thereto.

1. Technical Field

The present invention relates in general to storage devices within digital systems, and in particular to an improved edge-triggered latch that combines the data and clock generation path utilizing pass-transistor logic.

2. Description of the Related Art

Digital circuits, such as microprocessors and memory devices, typically use flip-flops as temporary storage devices. The most basic type of flip-flops operate with signal levels and are often referred to as latches. A latch with clock pulses as its control input is essentially a flip-flop that is triggered every time the clock pulse goes to a one or zero logic level. For ease of reference, it will be assumed that, as utilized herein, “latch” incorporates flip-devices and all clock-controlled latches.

Several categories of latches are known in the art including level-sensitive, master-slave, and edge-triggered. The present invention is directed to edge-triggered lateh design. With reference to FIG. 1, there is depicted a conventional D-latch 100 . In accordance with well-known D-latch design standards, D-latch 100 has two inputs, D (data) at a data input node 106 , and C (control) at a clock input node 108 . D-latch 100 generates a differential output at an output node 114 and a complementary output node 112 . The outputs at nodes 112 and 114 cannot change state while the clock input at node 108 is at a logic 0 regardless of the value of D at data input node 106 . A graphical representation of the input and output signals to and from D-latch 100 is provided in FIG. 2 .

D-latch 100 is a positive edge-triggered latch, meaning that it triggers a data input at input node 106 only during a positive transition of a clock signal C. Two such positive clock signal transitions, 212 and 214 , are depicted in FIG. 2 . Pulse generating circuitry including Complementary Metal Oxide Semiconductor (CMOS) P-type Field Effect Transistors (PFETs) P 2 and P 3 , and N-type Field Effect Transistors (NFETs) N 3 , N 5 , N 4 , and N 6 , is utilized to enable edge-triggered data propagation within D-latch 100 as follows.

Three CMOS inverters, I 1 , I 2 , and I 3 produce a delayed complementary version of control signal C, illustrated in FIG. 2 as {overscore (C)} dd . Control signal C is logically ANDed with its delayed complementary counterpart {overscore (C)} dd by series NFET pairs N 3 and N 5 , and N 4 and N 6 , to produce a data evaluation window having a width 210 . The resulting data evaluation window is felt at a pair of pull-down nodes 118 and 120 at the sources of NFETs N 1 and N 2 , respectively.

During a data evaluation window at pull-down nodes 118 and 120 data propagates through D-latch 100 . At the onset of a data evaluation window (i.e., at a rising edge of C), the biasing at pull-down nodes 118 and 120 enables a pair of CMOS inverters comprising P 1 and N 1 , and P 4 and N 2 , respectively, to pass the data through an internal data path node 104 and into a storage node 102 . From storage node 102 , the data propagates through inverter I 6 to complementary output node 112 and through inverters I 4 and I 7 to output node 114 . PFETs P 2 and P 3 suppress noise by maintaining a logic high at internal data path node 104 during non-data transfer cycles.

An ideal D-latch design provides a temporally symmetric complementary output as well as an optimized balance between performance and electrical efficiency. However, as illustrated in FIG. 2, the data path for Q is longer (four logic stages) than that for {overscore (Q)} (three logic stages), resulting in an unbalanced differential output. In addition to having an unbalanced output, D-latch 100 suffers a reduced power efficiency due to periodic “glitches” experienced by internal data path node 104 wherein a high-to-low transition occurs for each clock cycle even in the absence of a change in the data input. Finally, the performance of D-latch 100 in terms of transfer speed is reduced by its CMOS edge-triggering design wherein additional clock pulse generating elements N 3 , N 5 , N 4 , and N 6 are required to enable edge-triggered data propagation through the CMOS inverters in the data path.

From the foregoing, it can be appreciated that a need exists for an improved clock-triggered latch wherein the foregoing limitations of conventional latches are minimized.

An edge-triggered latch having improved clock-to-output performance and greater efficiency is disclosed herein. The edge-triggered latch of the present invention includes a data input and a clock input. Multiple source-to-drain connected pass-transistor logic (PTL) transistors are incorporated in the data path of the edge-triggered latch for converting a clock signal from the clock input into an edge-triggered data evaluation window. The PTL transistors propagate data from the data input into a storage node during the edge-triggered data evaluation window.

All objects, features, and advantages of the present invention will become apparent in the following detailed written description.

The novel features believed characteristic of the invention are set forth in the appended claims. The invention itself however, as well as a preferred mode of use, further objects and advantages thereof, will best be understood by reference to the following detailed description of an illustrative embodiment when read in conjunction with the accompanying drawings, wherein:

FIG. 1 illustrates a conventional edge-triggered D-latch;

FIG. 2 is a timing diagram depicting the operation of the conventional D-latch shown in FIG. 1;

FIG. 3 illustrates an edge-triggered D-latch designed in accordance with one embodiment of the present invention;

FIG. 4 is a timing diagram depicting the improved clock triggering performance realized by the edge-triggered D-latch shown in FIG. 3;

FIG. 5A illustrates a positive edge-triggered D-latch having a pass-transistor logic triggering mechanism in accordance with an alternate embodiment of the present invention;

FIG. 5B depicts a negative edge-triggered D-latch having a pass-transistor logic triggering mechanism in accordance with an alternate embodiment of the present invention;

FIG. 6 illustrates an edge-triggered D-latch wherein transmission gate logic is combined with a pass-transistor logic triggering mechanism in accordance with an alternate embodiment of the present invention;

FIG. 7 is a timing diagram depicting the improved clock triggering performance realized by the edge-triggered D-latch shown in FIG. 5A; and

FIG. 8 is a timing diagram depicting the improved clock triggering performance realized by the edge-triggered D-latch shown in FIG. 6 .

This invention is described in a preferred embodiment in the following description with reference to the figures. While this invention is described in terms of the best mode for achieving this invention's objectives, it will be appreciated by those skilled in the art that variations may be accomplished in view of these teachings without deviating from the spirit or scope of the present invention.

The present invention is directed to improving both the clock-to-output (C-to-Q) and data-to-output (D-to-Q) response for edge-triggered latches. Complementary Metal Oxide Semiconductor (CMOS) static logic design is employed in many latch designs. CMOS circuits consume less power and typically have higher performance in terms of switching speed that other N-channel Metal Oxide Semiconductor (NMOS) or P-channel Metal Oxide Semiconductor (PMOS) circuits. D-latch 100 in FIG. 1, is a static CMOS circuit wherein the latch data path from input node 106 to storage node 102 includes two inverters comprising NFET/PFET pairs N 1 /P 1 and N 2 /P 4 . Complementary NFET/PFET design is also incorporated within the pull-up/pull-down clock pulse generating mechanism within D-latch 100 comprising P 2 , P 3 , N 3 , N 5 , N 4 , and N 6 .

An inherent problem with the CMOS design depicted in FIG. 1, however, is that a leakage current flows within each CMOS NFET/PFET complementary device pair when the FETs are simultaneously switched on by an input signal. Thus, for high-speed, low-power applications, CMOS design techniques often impose severe speed versus power tradeoffs, thereby limiting their design flexibility.

The present invention replaces conventional CMOS clock pulse triggering and data path propagation logic with a pass-transistor logic (PTL) configuration that provides improved power efficiency and data transfer speed as well as an optimally balanced differential output.

With reference to the figures, and in particular with reference to FIG. 3, there is illustrated an edge-triggered D-latch 300 designed in accordance with one embodiment of the present invention. As illustrated in FIG. 3, D-latch 300 receives differential inputs D and {overscore (D)} from a pair of input nodes 310 and 312 , respectively. A pair of CMOS inverters I 7 and I 8 are utilized to generate differential inputs D and {overscore (D)}. A clock signal C is input into D-latch 300 at a control input node 316 . Three series-connected inverters I 9 , I 10 , and I 11 , invert and delay clock signal C, with the resultant signal {overscore (C)} dd being applied to a control input node 314 . FIG. 4 illustrates the timing relationship between signals C and {overscore (C)} dd .

In accordance with the depicted embodiment, pass-transistor logic (PTL) NFETs N 7 , N 8 , N 9 , and N 10 are incorporated within the data paths for D-latch 300 . As depicted in FIG. 3, delayed inverted clock signal {overscore (C)} dd provides the gate control input for NFETs N 7 and N 9 at control input node 314 , while clock signal C provides the gate control input for NFETs N 8 and N 10 at control input node 316 . Under the control of C and {overscore (C)} dd , NFETs N 7 , N 8 , N 9 , and N 10 serve as pass-gates with respect to the differential data inputs.

NFET pairs N 7 and N 8 , and N 9 and N 10 , are source-to-drain connected such that the clock signal at control input node 316 is logically ANDed with the delayed inverted clock signal at node 314 . As illustrated in FIG. 4, a positive edge-triggered data evaluation window having a width 410 results during which data at differential input nodes 310 and 312 propagates through pass-gate NFET pairs N 7 and N 8 , and N 9 and N 10 into a pair of storage nodes 306 and 308 , respectively. A pair of source-to-gate cross-coupled PFETs PS and P 6 are employed as keepers within D-latch 300 to facilitate rising edge transitions for the pass-gate NFETS. The present invention thus incorporates pass-gate transistors N 7 , N 9 , N 8 , and N 10 as active components within the latch triggering mechanism for D-latch 300 since the data evaluation window 410 is a combined effect of the control signals applied to the gates of N 7 , N 9 , N 8 , and N 10 .

With continued reference to FIGS. 3 and 4, upon the occurrence of a rising edge 414 of clock signal C, pass-gate NFETS N 7 , N 9 , N 8 , and N 10 are simultaneously opened, thus enabling differential inputs D and {overscore (D)} to propagate to a storage nodes 306 and 308 , respectively. From storage nodes 306 and 308 , the data values are output to a pair of differential output nodes 318 and 320 as Q and {overscore (Q)}, respectively. It should be noted that the delay between the rising edge of clock signal C 414 and the edge transitions of Q and {overscore (Q)} at 406 and 408 is significantly reduced in comparison to the analogous delay between the rising edge of clock signal C 214 and the transitions of the differential Q-outputs of conventional D-latch 100 in FIG. 1 . It should be further noted that the differential outputs Q and {overscore (Q)} are temporally balanced as a result of the symmetric data path design of D-latch 300 .

As illustrated in FIG. 3, the data input D follows a first data path comprising NFET pair N 9 and N 7 , while the complementary data input {overscore (D)} follows a parallel data path comprising NFET pair N 7 and N 8 . In an important feature of the depicted embodiment, it should be noted that the delayed inverted clock signal {overscore (C)} dd is applied as the control input for the first pass-gate transistor within each of the first and second data paths (i.e., NFETs N 7 and N 9 ), while the clock signal C is applied as the control input for the second pass-gate transistor within each of the first and second data paths (i.e., NFETs N 8 and N 10 ). This ordering of the application of {overscore (C)} dd and C to the pass-gates is required to ensure that data is only evaluated at the rising edge of C.

With reference to FIG. 5A, there is illustrated a positive edge-triggered D-latch 500 having a PTL latch trigger circuit in accordance with an alternate embodiment of the present invention. As depicted in FIG. 5A, D-latch 500 includes a clock input C at node 516 and differential data inputs D and {overscore (D)} at nodes 510 and 512 , respectively. In embodiment depicted in FIG. 5A, the latch trigger circuit for D-latch 500 is a pulse generator that generates a pulse at low-to-high clock signal transitions.

As further illustrated in FIG. 5A, D-latch 500 receives differential inputs D and {overscore (D)} from a pair of input nodes 510 and 512 , respectively. A pair of CMOS inverters I 16 and I 17 are utilized to generate differential inputs D and {overscore (D)}. A clock signal C is input into D-latch 500 at a control input node 516 . PTL NFETs N 14 and N 12 are incorporated as pass-gates within the data paths for D-latch 500 . Data path NFETs N 14 and N 12 are controlled by a latch trigger pulse applied to a single control node 514 . As illustrated in FIG. 7, a latch control pulse 710 is generated at control node 514 to permit the differential data input at nodes 510 and 512 to propagate through pass-gate NFETs N 14 and N 12 into a pair of storage nodes 506 and 508 , respectively.

The pulse generator for D-latch 500 includes clock input node 516 , which serves as an input to a pair of series-connected inverters 140 and 141 . Referring back to FIG. 7, latch control pulse 710 at control node 514 has a pulse width that corresponds to the delay imparted on clock signal C by inverters I 40 and I 41 . The delayed version of clock signal C at node 515 is depicted in FIG. 7 as delayed clock signal C dd . It can be further seen from FIG. 7 that the resultant triggering pulses are asserted at the rising edge of clock signal C and deasserted at the rising edge of delayed clock signal C dd .

The pulse generator for D-latch 500 further includes a PFET P 8 , which serves as a pass-gate for transmitting portions of clock signal C from clock input node 516 to control node 514 in accordance with the delayed clock C dd applied to the gate terminal of PFET P 8 at node 515 . A pass-gate NFET N 16 is also incorporated within the pulse generator to generate latch control pulse 710 at control node 514 .

FIG. 7 depicts an exemplary pulse generation operation implemented by PFET P 8 and NFET N 16 as follows. While delayed clock signal C dd at node 515 is at a logic level high, PFET P 8 is switched off, thus preventing clock signal C from reaching control node 514 . In addition, a logic high at node 515 switches NFET N 16 on, thus enabling NFET N 16 to pass ground (logic low) to control node while PFET P 8 is being switched off. When delayed clock signal C dd transitions to a logic level low, NFET N 16 switches off and PFET P 8 switches on to permit clock signal C to propagate to control node 514 . The overlapping portion of a low C dd and a high C is captured by latch trigger transistors PFET P 8 and NFET N 16 to produce a latch control pulse at control node 514 having a rising edge that corresponds to the rising edge of C.

The resultant positive-level pulses at control node 514 are applied to the gate terminals of data path pass-gate NFETs N 14 and N 12 , resulting in the differential data inputs at nodes 510 and 512 being sampled only in response to rising edge clock transitions.

With continuing reference to FIG. 7 in conjunction with FIG. 5A, upon the occurrence of latch control pulse 710 , pass-gate NFETS N 14 and N 12 are simultaneously opened, thus enabling differential inputs D and {overscore (D)} to propagate to a storage nodes 506 and 508 , respectively. From storage nodes 506 and 508 , the data values are output to a pair of differential output nodes 518 and 520 as Q and {overscore (Q)}, respectively.

It should be noted that the delay between the rising edge of clock signal C 714 and the edge transitions of Q and {overscore (Q)} at 506 and 508 are significantly reduced in comparison to the analogous delay between the rising edge of clock signal C 214 and the transitions of the differential Q-outputs of conventional D-latch 100 in FIG. 1 . It should be further noted that due to the pulse triggering functionality of the latch trigger circuit for D-latch 500 , only one pass-gate transistor is required for each data path, thus substantially reducing the data-to-output (D-to-Q) propagation time.

An alternative embodiment of D-latch 500 wherein negative (high-to-low) pulses are employed as the latch triggering mechanism is illustrated in FIG. 5B. A D-latch 550 is illustrated, which incorporates pass-gate PFETs P 10 and P 12 rather than NFETs into the differential data paths. The device polarity for the pulse generator is also reversed for D-latch 550 .

With reference to FIG. 6, there is depicted an edge-triggered D-latch 600 wherein transmission gate logic is combined with a pass-transistor logic triggering mechanism in accordance with an alternate embodiment of the present invention. A differential data input comprising data D and inverted data {overscore (D)} at nodes 610 and 612 , respectively, is generated by inverters I 28 and I 29 . As with D-latches 500 and 550 , D-latch 600 includes a single pass-data stage that is controlled by a latch trigger circuit.

The pass-data stage for D-latch 600 comprises a pair of transmission gates 630 and 632 . In accordance with conventional transmission gate design, transmission gates 630 and 632 comprise NFET/PFET pairs N 241 P 18 and N 26 /P 20 , respectively. The latch trigger circuit for D-latch 600 generates a control input for the transmission gates at control nodes 614 and 615 . A simultaneous logic low at node 615 and logic high at node 614 switches transmission gates 630 and 632 on and permits the differential data to propagate to a pair of storage nodes 606 and 608 .

The latch trigger circuit for D-latch 600 includes a set of series-connected inverters I 30 , I 31 , I 32 , and I 33 that receives a clock signal input at a clock source node 616 . The latch trigger circuit further comprises two NFET/PFET pass-gate pairs, N 28 /P 22 and N 22 /P 16 , which generate latch trigger signals at control nodes 614 and 615 . As depicted in FIG. 6, PFET P 22 serves as a pass-gate for transmitting portions of clock signal C from clock input node 616 to control node 614 in accordance with a delayed clock applied at node 622 to the gate terminal of PFET P 22 . A pass-gate NFET N 28 shares delayed output node 622 and control node 614 with the input and control pass-gate terminals of PFET P 22 .

Referring to FIG. 8 in conjunction with FIG. 6, the operation of the trigger circuit for D-latch 600 is exemplified. NFET/PFET pair N 22 /P 16 generates a latch control signal at control node 615 that switches data path PFETs P 18 and P 20 on or off. The clock signal C at node 616 propagates through inverter I 30 to produce a delayed inverted signal at node 626 , which is applied to the drain terminal of passgate NFET N 22 . The control signal applied to the gate terminals of N 22 /P 16 for passing the signal at node 626 through NFET N 22 into latch control node 615 is obtained after the clock signal is further delayed through a pair of inverters I 31 and I 32 .

The complementary transmission gate control signal for data path NFETs N 24 and N 26 is generated by NFET/PFET pair N 28 /P 22 at control node 614 . As depicted in FIG. 8, the latch control signal at node 614 coincides with the overlap of a high clock signal C at node 616 and a low delayed clock signal at node 622 . The latch trigger circuit for D-latch 600 produces a positive (low-to-high) pulse at control node 614 and a negative (high-to-low) pulse at control node 615 to enable to pass through transmission gates 630 and 632 . As explained below the generation of the latch trigger pulses occurs in response only to a rising clock edge at clock input node 616 .

The generation of a latch trigger window commences with a positive transition 814 of clock signal C at node 616 . A pulse having a rising edge 818 is generated at control node 614 until the delayed clock signal at node 622 transitions to a logic high. The pulse width of the pulse at control node 614 is delimited by the rising edge of the delayed clock signal at node 622 . Also initiated by the rising edge 814 of clock signal C, a negative pulse is generated within the rising and falling boundaries of the generated positive pulse. A logic high at node 624 enables pass-gate NFET N 22 to pass the falling edge of delayed clock signal 626 to control node 615 as falling edge 816 .

It should be noted that the delay between the rising edge of clock signal C at node 616 and the edge transition of Q at 812 is significantly reduced in comparison to the analogous delay between the rising edge of clock signal C 214 and the transitions of the differential Q-outputs of conventional D-latch 100 in FIG. 1 . It should be further noted that the symmetric control and data paths of D-latch 600 result in balanced C-to-Q and D-to-Q responses. Furthermore, the pulse generation functionality provided by the latch trigger circuit within D-latch 600 allows for a simplified data path wherein only a single transmission gate in each of the differential data paths is required for each data path, thus substantially reducing the data-to-output (D-to-Q) propagation time.

While the invention has been particularly shown and described with reference to a preferred embodiment, it will be understood by those skilled in the art that various changes in form and detail may be made therein without departing from the spirit and scope of the invention.