Current source evaluation sense-amplifier

Abstract

A sense-amplifier based on current-source-evaluation. Compared to Conventional sense-amplifiers, a design based on static-current sources scales better to small transistor geometries. The design has lower power consumption, reduced noise, and improved clock scaling.

Description

BACKGROUND

[0003] 1. Field

[0004] This invention relates to sense amplifiers and, in one embodiment, sense amplifiers for deep sub-micron designs.

[0005] 2. Description of Related Art

[0006] A. Overview

[0007] Sense amplifiers (“SAs”) were traditionally used in dynamic and static (cache, buffers, etc.) random access memory (RAM) to improve memory-access latency. Due to increasing clock speeds of digital logic circuits and increasing on-chip interconnect parasitics, SAs have also gained popularity in modern digital logic circuit designs. In static and dynamic memory, SAs are used to improve the read-access time. In digital logic, SAs are used to improve the performance of flip-flops (“FFs”), the most commonly used digital building block.

[0008] Flip-flops are an integral part of digital design because of their use in retime, deskew and receive circuitry. Previous work on flip-flops has included efforts at utilizing SA front-ends to improve sensitivity and speed of conventional flip-flops. The basic SA flip-flop (“SAFF”) circuit contains a SA front-end followed by a latch stage.

[0009] Previous work on high-performance SAFFs has concentrated on improving the latch stage and has mostly ignored the SA front-end. The SAFF reported in J. Montanaro, R. T. Witek, et al., “A 160 MHz, 32-b 0.5 W CMOS RISC microprocessor,” IEEE J. Solid-State Circuits, vol. 31, pp. 1703-1714(November 1996), uses a NAND set reset (SR) latch; H. Partovi, R. Burd, et al., “Flow-through latch and edge-triggered flip-flop hybrid elements,” ISSCC Dig. Tech. Papers, pp 138-139 (1996), uses a hybrid-latch flip-flop; a semi-dynamic flip-flop in F. Klass, “Semi-dynamic and dynamic flip-flops with embedded logic,” Symp. VLSI Circuits Dig. Tech Papers, pp. 108-109 (1998), uses dynamic-style flip-flops; B. Nikolic, V. Oklobdzija, V. Stohanovic, W. Jia, “Improved sense-amplifier-based flip-flop: Design and measurements,” IEEE J. of Solid-State Circuits, pp. 876-884, vol. 35 (June 2000) uses a cross-coupled inverter latch; and J. C. Kim, Y. C. Jang and H. J. Park, “CMOS sense-amplifier-based flip-flop with N-C2MOS output latches,” Electronic Letters, vol. 36, pp.498-500 (March 2000), uses a N-C2MOS latch.

[0010] Common to all these designs, which are incorporated herein by reference, is a standard sense-amplifier front-end. Each of these types of sense-amplifiers are referred to in this application as a conventional voltage mode sense-amplifier (“CVSA”).

[0011] Research on SA designs have analyzed the sensitivity of the CVSA design to process variations such as transistor mismatch and have also improved the original design such as in the case of pMOS cross-coupled SA (“pCCSA”). These approaches to SA design assumed and were optimized for relatively large transistor geometries and higher voltage supply-rails than commonly in use today.

[0012] For a higher-voltage supply-rail, it is important to eliminate DC current paths (zero static power consumption) to achieve low power. In both CVSA and pCCSA, this is achieved by including clocked transistors in the evaluation path. Over the years, despite some optimization that has been made to earlier designs, the core architecture of the clocked voltage SA has remained similar to the initial designs with clocked transistors in the evaluation path.

[0013] Today, deep sub-micron CMOS technology has relatively low-voltage supply rails. As a consequence, it is important to reduce the evaluation chain depth to improve circuit performance and process scaling. To employ SAs in high-speed designs, they should also be able to achieve a multi-GHz operating frequency without compromising power, sensitivity and clock-load. Unfortunately, conventional voltage sense-amplifier designs carry a high clock-load burden that limits their frequency scaling. The experimental results for earlier SA designs have therefore been limited to sub-GHz operating frequencies. See B. Nikolic, V. Oklobdzija, V. Stohanovic, W. Jia, “Improved sense-amplifier-based flip-flop: Design and measurements,” IEEE J. of Solid-State Circuits, pp. 876-884, vol. 35 (June 2000), the content of which is incorporated herein by reference.

[0014] To improve scaling of SAs into deep sub-micron designs, it is therefore important to reduce the evaluation chain depth, reduce the noise generated by the SA, and reduce the SA clock-load.

[0015] B. Detailed Description of Prior Art CVSA and pCCSA

[0016] The operation of a clocked sense-amplifier consists of a pre-charge/discharge phase, also known as an equalization phase, and an evaluation phase. To eliminate DC power consumption, commonly used sense-amplifier architectures have a clocked transistor in the evaluation chain. Two of the most popular SA architectures are the conventional voltage sense amplifier (“CVSA”) and pMOS cross-coupled sense-amplifier (“pCCSA”), which is a modified version of the CVSA.

[0017] (1) Conventional Voltage Sense Amplifier (CVSA)

[0018] FIG. 1 us a schematic of a prior art conventional voltage-mode sense-amplifier (CVSA). The sense-amplifier has a clocked nMOS transistor MNC in the evaluation path 104 and an evaluation path depth of three. One path is MN3, MN1 and MNC; the other is MN4, MN2 and MNC.

[0019] As is well understood by those skilled in the art, VDD is a reference to the supply voltage; φ is a reference to the clock input, D is a reference to the data input; {overscore (D)} is a reference to the complement of the data input; OUT is a reference to the data output; {overscore (OUT)} is a reference to the complement of the data output; the first letter “M” in the labels signifies a transistor; the second letter “P” or “N” specifies that the transistor is a “P” or “N” channel transistor; the “C” subscript specifies that the transistor is managing the clock signal; the “D” subscript specifies that the transistor is managing the data signal; and the subscript number represents an arbitrary number to distinguish each transistor from others of the same type and management function. This same nomenclature is used throughout all of the figures in this application.

[0020] During the equalization phase, the clock φ is low; transistor MNC is switched off; transistors MPC1 and MPC2 are switched on; and the nodes OUT and {overscore (OUT)} are precharged to VDD. The high output keeps transistors MN3 and MN4 switched on. Since there is no current path to ground, the intermediate nodes 102 and 103 are precharged to VDD-VthN, where VthN is the threshold voltage of the N transistors.

[0021] In addition, node 100 is precharged to VIN—HIGH-VthN, where VIN—HIGH is the higher of the two voltages at inputs D and {overscore (D)}. The lower limit on the size of transistors MPC1 and MPC2 is determined by their ability to fully precharge the nodes in half a clock cycle, i.e., in a single clock phase. The precharge speed of the CVSA in FIG. 1 is proportional to the precharge transistor conductivity of MPC1 and MPC2 and is inversely proportional to the SA load and parasitic capacitance.

[0022] At the clock φ goes high, the SA enters the evaluation phase. During the evaluation phase, transistors MPC1 and MPC2 are switched off and transistor MNC is switched on. For small input signals (D−{overscore (D)}=±ΔV), transistors MN1 and MN2 remain switched on and, as a result, there is a current path to ground from both OUT and {overscore (OUT)}. However, due to the voltage differential between the inputs, one of the paths will sink more current compared with the other. The evaluation path 104 connected to the higher input voltage will have higher current-sinking capability, and the output of that evaluation path will get pulled down to ground.

[0023] From the previous description, it will be understood by those skilled in the art that the evaluation speed of the SA in FIG. 1 is proportional to the input voltage differential, the conductivity of the evaluation path, and is inversely proportional to the SA load and parasitic capacitance. The conductivity of the evaluation chain is proportional to the nMOS transistor size, and is inversely proportional to the number of series transistors. As seen in FIG. 1, the CVSA has a 3-transistor deep evaluation path 104.

[0024] The pMOS and nMOS transistors can be interchanged in the CVSA design to generate a predischarge sense-amplifier. When the evaluation path is changed to pMOS, the transistor sizes must be increased to compensate for the reduced speed of pMOS (nMOS is typically 3-4 times faster than PMOS). Since the parasitic capacitance of the SA increases as an unfortunate side effect of increased pMOS transistor sizes, the actual increase in the pMOS transistor sizes will be greater than the mobility ratio between nMOS and pMOS.

[0025] If the parasitic contribution from the evaluation chain is assumed to remain constant, moving from precharge to predischarge results in 3 times the reduction in the predischarge transistor size. As a result, the capacitive contribution from predischarge transistors to the output node capacitance and clock-load will be reduced by a factor of 3. Unfortunately, since the pMOS evaluation transistors need to be resized to achieve the required evaluation speed, moving from precharge to predischarge would be likely to increase the parasitic capacitance of the evaluation chain by more than 3 times. Therefore, an increase in total clock-load capacitance and SA node capacitance is expected. Thus, the predischarge type CVSA will in general under perform its precharge counterpart.

[0026] For slower clock speeds, where the parasitic capacitance of the CVSA is smaller than the load capacitance, transistors can be resized to achieve better performance without significantly increasing the total output node capacitance. As a result, at slower clock speeds there is a linear relationship between the CVSA performance and the clock-load. A CVSA designed in 0.18 μm CMOS technology optimized to operate at 1.0 GHz clock frequency has a clock-load that is 1.8 times greater than a CVSA optimized for a 500 MHz clock frequency.

[0027] Hspice simulation results for a conventional voltage SAs (CVSA) optimized for different operating frequencies simulated in 0.18 μm CMOS is shown in Table 1. The output load is 10.0 fF and the target input sensitivity is 100 mVPP.

TABLE 1
Optimized operating
frequency	Delay (ps)	Clock-load (fF)	RMS Power (μW)
1.5 GHz	96	10	240
2.0 GHz	72	18	410
2.5 GHz	63	27	655
3.0 GHz	57	39	980

[0028] On the other hand, for frequencies above 1.0 GHz, transistor resizing is of decreasing benefit. This is because the CVSA output parasitic capacitance is comparable to the load capacitance and resizing to improve CVSA performance also increases the total CVSA capacitance, limiting the overall performance improvement. At a clock frequency of 1.5 GHz, CVSA output parasitic to load capacitance ratio is 0.75. When the transistors are doubled in size, even though the output node capacitance increases by a factor of 1.4, there is still a 1.4 times improvement in the speed. On the other hand, if the transistors are resized by a factor of 4, due to the 2.3 times increase in the output capacitance, there is only a 1.75 times speed improvement. This analysis agrees with the Hspice simulation results given in Table 1. Compared to the 1.5 GHz CVSA, the 3.0 GHz CVSA achieved a 1.7 times improvement in the evaluation delay at the cost of a 4 times increase in the clock-load. Clearly the CVSA has serious performance limitations as clock frequency is increased.

[0029] (2) pMOS and nMOS Cross-Coupled Sense-Amplifiers (PCCSA and nCCSA)

[0030] FIG. 2(a) is a schematic of a prior art PMOS cross-coupled SA design. This is a modification to the conventional architecture and does not include a cross-couple transistor pair in the evaluation path 201a. Since both the CVSA design of FIG. 1 and the cross-coupled SA design in FIG. 2(a) share the same basic design, the operation of the two SAs are similar and also have similar drawbacks.

[0031] Eliminating the cross-couple transistors from the evaluation chain reduces the evaluation path depth and the number of parasitic nodes in the SA. In a pCCSA, when the nMOS cross-coupled transistors MN3 and MN4 in FIG. 1 are eliminated from the evaluation path, the pMOS cross-couple pair MP1 and MP2 need to be resized to compensate. The resulting increase in the pMOS cross-couple pair MP1 and MP2 increases the SA parasitic capacitance and reduces the SA frequency response.

[0032] The frequency response of the cross-couple SA can be improved by swapping the pMOS and nMOS transistors, resulting in the prior art nMOS cross-couple sense amplifier (nCCSA) shown in FIG. 2b, with an evaluation path 201b.

[0033] Compared with a pMOS cross-couple load, an nMOS cross-couple load is able to provide stronger positive feedback. In addition, since the nCCSA uses nMOS predischarge, compared to the pMOS counterpart, the equalization phase has faster response time and lower capacitance. In addition, the analysis of the clock-load characteristics shows that predischarge architecture also reduces the equalization phase clock-load significantly.

[0034] Though there is a significant reduction in equalization clock-load, it does not translate into a significant reduction in the total clock-load. In nCCSA, the clocked pMOS evaluation transistor MPC dominates the clock-load, and the capacitance contribution of MPC is 4 times greater than the capacitive contribution of MNC1 and MNC2 for 3.0 GHz SA in 0.18 μm technology.

SUMMARY

[0035] A clocked sense amplifier may have an evaluation path that is not clocked and that includes two substantially matched current sources.

[0036] The clocked sense amplifier may include a current mode core that has a transistor stack consisting of no more than one transistor.

[0037] The evaluation path may include a pMOS transistor.

[0038] The clocked sense may be configured to operate with an equalization phase and an evaluation phase. The equalization phase may include a predischarge phase. An nMOS transistor may be used in the predischarge phase.

[0039] The clocked sense amplifier may include an input differential transconductance stage for sensing the input to the amplifier.

[0040] The clocked sense amplifier may include an input pass transistor stage.

[0041] The clocked sense amplifier may be connected to a current-steering latch.

[0042] The clocked sense amplifier may be connected to an nMOS dynamic latch.

[0043] The clocked sense amplifier may be connected to a current-steering dynamic latch.

[0044] Two of the clocked sense amplifiers may be configured to form a demultiplexer.

[0045] The clocked sense amplifier further may include an input and a terminating resistor in communication with the input configured to create a high speed receiver.

[0046] Two of the clocked sense amplifiers may be configured to form a multiplexer.

[0047] The clocked sense amplifier may include an output and at terminating resistor in communication with the output configured to reduce power supply noise.

[0048] The clocked sense amplifier may include an input, an optical receiver in communication with the input, and a photodiode in communication with the optical receiver, configured to create a low power, low noise quantizer.

[0049] The clocked sense amplifier may be constructed using sub-micron technology. 0.18 μm or 0.13 technology may be used, as well as other technology, including smaller technologies as they are developed.

[0050] The clocked sense amplifier may have an evaluation path that includes a pMOS transistor that is not clocked, a current mode core that has a transistor stack consisting of no more than one transistor, and two substantially matched current sources.

[0051] A method of sense amplification is provided that may include an equalization step during a first portion of a clock cycle and an evaluation step during a second portion of a clock cycle that utilizes an evaluation path that is not clocked.

[0052] An amplified latch is provided that may include a clocked sense amplifier having an evaluation path that is not clocked and two substantially matched current sources, the amplifier having an output and a latch having an input connected to the output of the clocked sense amplifier.

[0053] These as well as still further features, objects and benefits of the invention will now become clear from consideration of the following detailed description of illustrative embodiments, taken in conjunction with the attached drawings.

BRIEF DESCRIPTION OF DRAWINGS

[0054] FIG. 1 is a schematic of a prior art conventional voltage-mode sense-amplifier (CVSA).

[0055] FIGS. 2(a) and (b) are schematics of prior art cross-couple SA designs, FIG. 2(a) using pMOS cross-coupled transistors and FIG. 2(b) using nMOS cross-coupled transistors.

[0056] FIG. 3 is a schematic of one embodiment of the invention in which certain sections are represented by functional blocks, rather than specific components.

[0057] FIG. 4 is a schematic of one implementation of the embodiment of the invention shown in FIG. 3.

[0058] FIGS. 5(a)-(c) are block diagram of sense amplifier and latch circuits made in accordance with the invention being used in connection with latch circuits.

[0059] FIG. 6 is a block diagram of sense amplifier and latch circuits made in accordance with the invention being cross-coupled to create a de-multiplexer.

[0060] FIG. 7 is a block diagram of a sense amplifier and latch circuit made in accordance with the invention and configured to implement a high-speed receiver.

[0061] FIG. 8 is a block diagram of two sense amplifier and latch circuits made in accordance with the invention configured to implement a multiplexer.

[0062] FIG. 9 is a block diagram of a sense amplifier and latch circuit made in accordance with the invention and configured to reduce power supply noise.

[0063] FIG. 10 is a block diagram of a sense amplifier and latch combination made in accordance with the invention and configured to provide a low power, low noise, quantizer optical receiver.

[0064] FIG. 11 illustrates another embodiment of the invention using an input pass-transistor stage.

DETAILED DESCRIPTION OF ILLUSTRATIVE EMBODIMENTS

[0065] FIG. 1 is a schematic of a prior art conventional voltage-mode sense-amplifier (CVSA) and has been discussed above in the Description of Related Art.

[0066] FIGS. 2(a) and (b) are schematics of prior art cross-couple SA designs, FIG. 2(a) using pMOS cross-coupled transistors and FIG. 2(b) using nMOS cross-coupled transistors. These figures have also been discussed above in the Description of Related Art.

[0067] FIG. 3 is a schematic of one embodiment of the invention in which certain sections are represented by functional blocks, rather than specific components.

[0068] As shown in FIG. 3, the clocked evaluation path 104 in FIG. 1 and 201a and 201b in FIG. 2 is replaced with a current mode core 301 that includes an evaluation path 307 that is not clocked and that includes two substantially-matched current sources 303 and 305.

[0069] The current mode core 301 also includes nMOS cross-coupled transistors MN1 and MN2 for positive feedback, and an input differential transconductance stage 308 for converting the input voltage to current. Although a particular type of differential transconductance stage is shown in the figure, it is, of course, to be understood that any other type could instead be used in other embodiments.

[0070] As is well understood by those skilled in the art, the nomenclature VBIAS is a reference to a bias voltage source and the subscript “S” as part of a transistor designation indicates that the transistor is processing the bias voltage. This same nomenclature applies to all of the figures in which this nomenclature appears.

[0071] Removal of the clock transistor from the evaluation chain 307 reduces the clock-load, improves the evaluation chain turn-on delay, and improves sense amplifier noise and threshold variation tolerance.

[0072] FIG. 4 is a schematic of one implementation of the embodiment of the invention shown in FIG. 3. To further reduce the clock-load and SA output parasitic capacitance due to clocked transistors, the pMOS precharge of the prior art CVSA shown in FIG. 1 is replaced with nMOS predischarge transistors MNC1, MNC2 and MNC3, and the evaluation chain 407 is changed from nMOS to pMOS transistors MP1 and MP2. The improvement in the turn-on delay from static current sources, the reduction in the evaluation chain depth, and the reduced clock transistor parasitic allows the pMOS current source transistors MP1 and MP2 to be sized comparable to those of the CVSA cross-couple nMOS transistors, without impacting the evaluation delay.

[0073] When clock signal φ is high in FIG. 4, both OUT and {overscore (OUT)} are predischarged to ground. At the falling edge of φ, if D is low (V(D-{overscore (D)})=ΔVVIN<0), ID+½IS+ΔI) current flows through transistor MNC1 and ID+(½IS−ΔI) current flows through MNC2. As a result of the current disparity, OUT changes from 0 to 1, while {overscore (OUT)} remains at ground. When optimizing for higher frequencies, current ID is selected so that at the target frequency, transistors MP1 and MP2 are always in saturation (VOUT≦VBIAS+|VthP|, where VthP is the threshold voltage of the pMOS).

[0074] To improve the turn-on time of the evaluation transistors MN1 and MN2 and to reduce the clock-load, the predischarge transistor MNC1 and MNC2 width is selected so that during the predischarge phase the outputs remain just below VthN. The transistor MNC3 is used for equalization of the output nodes during predischarge. In 0.13 μm CMOS technology, the CSESA with 50 mVPP input sensitivity has a power consumption of 500 μW, a clock-load of 4.5 fF and an operating frequency of 5.0 GHz for a 3 times inverter load. On the other hand, a conventional SA with the same input sensitivity has an operating frequency of only 4.0 GHz, and a clock-load of 29.3 fF for a 3 times inverter load.

[0075] A brief performance comparison of the two sense amplifier architectures is given in Table 2. Table 2 compares the Hspice simulated performance of the invention with the best achievable speed that applicant has been able to obtain for the prior art CVSA, nCCSA and CSESA. The SAs are optimized for a fan-out of 3 (3 times minimum size inverter).

TABLE 2
TSMC
0.13 μm CMOS	CVSA	nCCSA	Invention	Units
Power supply	1.2	1.2	1.2	Volts
Speed	4.0	4.4	5.0	GHz
Clock-load	29.3	18	4.5	fF
Power	0.44	0.36	0.5	mW
consumption
(RMS)
Power supply	1.4	0.8	0.12	mA
noise peak-peak
(Vdd)

[0076] FIGS. 5(a)-(c) are block diagrams illustrating sense amplifiers 501, 505 and 509 made in accordance with the invention being used in connection with latch circuit 503, 507, 511 and 513, respectively.

[0077] As illustrated in FIG. 5(a), the complementary outputs of the sense amplifier 501 is connected to the set and reset inputs of the latch 503 which, in this instance, is a set-reset latch. The combination collectively constitutes a current-steering latch.

[0078] As illustrated in FIG. 5(b), the complementary outputs of the sense amplifier 505 is connected to the set and reset inputs of the latch 507 which, in this example, is a clocked set-reset latch. This collectively creates a current-steering dynamic latch.

[0079] FIG. 5(c) illustrates an alternate embodiment of the current-steering dynamic latch in which the complementary outputs of the sense amplifier 509 are each

[0080] connected to the D input of latches 511 and 513 which, in this example, are clocked D memories.

[0081] FIG. 6 is a block diagram of sense amplifier and latch circuits made in accordance with the invention being cross-coupled to create a de-multiplexer. Two clocked sense amplifiers/latch combinations 601 and 603 made in accordance with the invention, such as the combinations shown in FIG. 5(b) or (c), are connected in parallel to the D and D input, creating a de-multiplexer of a multiplex D signal. The de-multiplexing occurs during the different phases of the clock Q signal.

[0082] FIG. 7 is a block diagram of a sense amplifier and latch circuit made in accordance with the invention and configured to implement a high-speed receiver, such as an LVD or CMC. As shown in FIG. 7, a clocked sense amplifier/latch combination 701 made in accordance with the invention, such as the combination shown in FIG. 5(b) or (c), has resistors 703 and 705 connected to the D and D inputs, respectively, thus acting as termination resistors.

[0083] FIG. 8 is a block diagram of two sense amplifier and latch circuits made in accordance with the invention configured to implement a multiplexer. As shown in FIG. 8, two clocked sense amplifier/latch combinations 801 and 803 made in accordance with the invention, such as the combination shown in FIG. 5(b) or (c), are connected to a combiner block 805 that may include NOR gates.

[0084] FIG. 9 is a block diagram of a sense amplifier and latch circuit made in accordance with the invention and configured to reduce power supply noise. As shown in FIG. 9, a sense amplifier/latch combination 901 made in accordance with the invention, such as the combination shown in FIG. 5(b) or (c), has resistors 903 and 905 connected to the OUT and {overscore (OUT)} outputs, respectively, thus acting as termination resistors.

[0085] FIG. 10 is a block diagram of a sense amplifier and latch combination made in accordance with the invention and configured to provide a low power, low noise, quantizer optical receiver. As shown in FIG. 10, a sense amplifier and latch combination 1001 made in accordance with the invention, such as the combination shown in FIG. 5(b) or (c), has its inputs connected to an optical receiver 1003 driven by a photodiode 1005.

[0086] As should now be apparent to those skilled in the art, the embodiments that have now been discussed are merely illustrative. The invention has wide applicability to numerous applications and configurations in numerous forms.

[0087] For example, the invention is useful in deep sub-micron designs, such as 0.18 μm technology and 0.13 μm technology, and is useful in other size technologies as well.

[0088] FIG. 11 illustrates but another embodiment of the invention using an input pass-transistor stage. As shown in FIG. 11, an input pass-transistor stage 1101 receives the D and {overscore (D)} data inputs. The other portions of the circuit are the same and operate in substantially the same way as the embodiment of the invention shown in FIG. 4 and discussed above in connection with FIG. 4.

[0089] In short, the invention is limited solely by the claims that now follow.

Claims

1. A clocked sense amplifier having an evaluation path that is not clocked and that includes two substantially matched current sources.

2. The clocked sense amplifier of claim 1 further including a current mode core that has a transistor stack consisting of no more than one transistor.

3. The clocked sense amplifier of claim 1 wherein the evaluation path includes a pMOS transistor.

4. The clocked sense amplifier of claim 1 configured to operate with an equalization phase and an evaluation phase.

5. The clocked sense amplifier of claim 4 wherein the equalization phase includes a predischarge phase.

6. The clocked sense amplifier of claim 5 further including a nMOS transistor used in the predischarge phase.

7. The clocked sense amplifier of claim 1 further including an input differential transconductance stage for sensing the input to the amplifier.

8. The clocked sense amplifier of claim 1 further including an input pass transistor stage.

9. The clocked sense amplifier of claim 1 connected to a current-steering latch.

11. The clocked sense amplifier of claim 1 connected to a nMOS dynamic latch.

12. The clocked sense amplifier of claim 1 connected to a current-steering dynamic latch.

13. Two clocked sense amplifiers of the type recited in claim 1 configured to form a de-multiplexer.

14. The clocked sense amplifier of claim 1 further including an input and a terminating resistor in communication with the input configured to create a high speed receiver.

15. Two clocked sense amplifiers of the type recited in claim 1 configured to form a multiplexer.

16. The clocked sense amplifier of claim 1 further including an output and a terminating resistor in communication with the output configured to reduce power supply noise.

17. The clocked sense amplifier of claim 1 further including an input, an optical receiver in communication with the input, and a photodiode in communication with the optical receiver configured to create a low power, low noise quantizer.

18. The clocked sense amplifier of claim 1 constructed using sub-micron technology.

19. The clocked sense amplifier of claim 18 constructed using 0.18 um technology.

20. The clocked sense amplifier of claim 18 constructed using 0.13 um technology.

21. A clocked sense amplifier having an evaluation path that includes a pMOS transistor and that is not clocked, a current mode core that has a transistor stack consisting of no more than one transistor, and two substantially matched current sources.

22. A method of sense amplification comprising: a. an equalization step during a first portion of a clock cycle; and b. an evaluation step during a second portion of the clock cycle that utilizes an evaluation path that is not clocked.

23. An amplified latch comprising: a. a clocked sense amplifier having an evaluation path that is not clocked and including two substantially matched current sources, said amplifier having an output; and b. a latch having an input connected to the output of said clocked sense amplifier.