Imported: 13 Feb '17 | Published: 18 Jan '11
USPTO - Utility Patents
An input buffer having a comparator that receives an input signal, a reference signal and a positive feedback. The comparator compares the input signal relative to the reference signal and generates an output signal transitioning between a first logic state and a second logic state responsive to the magnitude of the input signal transitioning through the magnitude of the reference signal. The comparator intensifies the output signal in response to the positive feedback from the output of the comparator while the output signal transitions from the first logic state to the second logic state.
This application is a continuation of U.S. patent application Ser. No. 11/639,452, filed Dec. 15, 2006 now, U.S. Pat. No. 7,558,125. This application is incorporated by reference herein.
This invention relates generally to integrated circuits, and more specifically to an apparatus and method for a comparator circuit that uses AC positive feedback to reduce false switching due to slope reversals of a received signal.
Input buffers are commonly used in a wide variety of integrated circuits. Buffers generally perform a number of advantageous functions when used in digital circuits. For example, buffers generally provide a high input impedance to avoid excessively loading circuits to which they are connected, and they have a low output impedance to simultaneously drive electrical circuits without excessive loading. Buffers can condition the signals applied to internal circuits so that the internal signals have well-defined logic levels and transition characteristics. Buffers are used, for example, to couple command, address and write data signals from command, address and data buses, respectively, of memory devices, including dynamic random access memory (“DRAM”) devices.
There are also several types of input buffers. For example, there are single ended input buffers in which a single input signal is applied to the buffer to cause the buffer to transition when the input signal transitions through predetermined voltage levels. Single-ended input buffers may also be used to compare the input signal to a reference voltage so that when the input signal transitions through the reference voltage the output of the buffer also transitions. Differential input buffer circuits are useful in digital circuits for determining whether an unknown input voltage is either above or below a fixed reference voltage. A conventional differential input buffer 100 is shown in FIG. 1 that includes a pair of differential amplifiers 101, 103, and an output coupled to an inverter 140. The amplifiers 101, 103 are connected in parallel between a PMOS transistor 102 that is coupled to a supply voltage VCC and an NMOS transistor 104 that is coupled to ground. When enabled by an active low signal EN_, the supply voltage VCC supplies a current through the PMOS transistor 102 to a node 105. As a result, a constant current is provided to the amplifier 101 and a PMOS transistor 122 coupled to the amplifier 103. Similarly, the supply voltage VCC is directly coupled to the gate of the transistor 104 such that a constant current is coupled from a node 115 through the transistor 104, thereby drawing current through an NMOS transistor 110 coupled to the amplifier 101 and to the amplifier 103. Therefore, the transistor 122 functions as a current source providing constant current to amplifier 103, and transistor 110 functions as a current sink to discharge a constant current from amplifier 101.
The two differential amplifiers 101, 103 have essentially the same components, but are complementary configured with respect to each other. The differential amplifier 101 includes a pair of PMOS transistors 116, 118 whose gates are coupled to each other and to node 105, and further coupled to the drain of the PMOS transistor 118. The transistors 116, 118 are coupled to each other in a current mirror configuration so they both have the same gate-to-source voltage. As a result, the transistors 116, 118 have the same source-to-drain resistance. The drains of the transistors 116, 118 are coupled to the drains of two NMOS transistors 112, 114 respectively, and receive an input signal VIN and a reference signal VREF at their respective gates. When the magnitude of the VIN signal is at ground potential, the transistor 112 is turned OFF. As a result, an output node 108 at which an OUT_signal is generated is driven high through the PMOS transistor 116. An inverter 140 having an input coupled to the node 108 thus generates a low DIFF_OUT signal. When the magnitude of the VIN signal is at VCC, the transistor 112 is turned ON with a significantly higher gate-to-source voltage than the gate-to-source voltage of the PMOS transistor 116. As a result, the resistance of the transistor 112 is significantly lower than the resistance of the transistor 116. The voltage at the output node 108 is therefore low enough so that the inverter 140 outputs a high DIFF_OUT signal. As the magnitude of the VIN signal passes through the magnitude of the VREF signal, which is typically VCC/2, the NMOS transistors 112, 114 have the same gate-to-source voltage and hence the same resistance. Furthermore, the NMOS transistors 112, 114 will have the same gate-to-source voltage as the PMOS transistors 116, 118. If the NMOS transistors 112, 114 have the same electrical characteristics as the PMOS transistors 116, 118, the PMOS transistors 116, 118 will then have the same resistance as the NMOS transistors 112, 114. In such case, the OUT_voltage will be equal to VCC/2. Therefore, decreasing the magnitude of the VIN signal increases the resistance across the transistor 112, reducing the current through the transistors 112, 116 to cause the magnitude of the OUT_signal to increase. Conversely, increasing the magnitude of the VIN signal decreases the resistance across the transistor 112, increasing the current through the transistors 112, 116 to cause the magnitude of the OUT_signal to decrease.
The amplifier 103 includes components that are the same as the amplifier 101, and thus for the sake of brevity, the components to the amplifier 103 will not be described in detail. The amplifier 103 has a topology that is complementary to the topology of the amplifier 101 so that the gates of NMOS transistors 128, 130 are coupled together in a current mirror configuration so that both transistors 128, 130 have the same resistance. As the magnitude of the VIN signal increases, the resistance of the PMOS transistor 126 increases to cause the magnitude of the OUT_signal to decrease. Conversely, as the magnitude of the VIN signal decreases, the resistance of the PMOS transistor 126 decreases to cause the magnitude of the OUT_signal to increase.
When the magnitude of the VIN signal decreases below VT, where VT is the threshold voltage of the NMOS transistor 112, the transistor 112 is turned OFF and thus no longer responds to changes in the magnitude of VIN. Similarly, when the magnitude of the VIN signal increases above VCC−VT, where VT is the threshold voltage of the PMOS transistor 126, the transistor 126 is turned OFF and thus no longer responds to changes in the magnitude of VIN. Thus, the buffer 100 can operate at all values of VIN from 0 to VCC, but only one amplifier 101 or 103 is operable with the magnitude of VIN below VT or above VCC−VT.
When the difference between VIN and VREF is small, such as when VIN transitions through VREF, the integrity of the input signal can be easily compromised by a number of interferences, such as improper bus termination, reflections, signal noise, and VREF noise. These factors can result in false switching of the buffer 100 as shown in FIG. 2. For example, due to the presence of noise, the VREF signal may fluctuate about its predetermined value, such as VCC/2. As the VIN signal approaches the VREF signal, the noise interference on the VREF signal may overlap the VIN signal such that a slope reversal 205 occurs, where the buffer 100 detects VREF to be greater than VIN, when in fact VIN is intended to be greater than VREF, but may not be due to noise and or signal reflections. Consequently, the buffer 100 may falsely switch its output, thereby generating an incorrect response to the input signal and causing delays or resulting in errors to the overall operation of the system or component that relies on the buffer 100.
Therefore, there is a need for a low current input buffer that reduces false switching in the presence of noise due to input signal slope reversals, and restores signal integrity.
Embodiments of the present invention are directed to an input buffer with AC positive feedback. Certain details are set forth below to provide a sufficient understanding of the invention. However, it will be clear to one skilled in the art that the invention may be practiced without these particular details. In other instances, well-known circuits, control signals, and timing protocols have not been shown in detail in order to avoid unnecessarily obscuring the invention.
FIG. 3 shows a block diagram of a buffer 300 according to an embodiment of the invention. The buffer 300 includes a differential comparator 302 that receives an input signal VIN and a reference signal VREF. Similarly to the conventional input buffer 100, the comparator 302 of the buffer 300 compares the two input signals and generates an inverted output signal OUT_depending on whether the VIN signal is above or below the reference VREF. An inverter 304 drives and inverts the OUT_signal to generate a buffer output signal DIFF. However, as previously described, when the VIN signal approaches the trip point determined as VREF, false switching due to input signal slope reversals can occur. In such cases, the buffer 300 reduces false switching by coupling the DIFF signal at a node 308 to the comparator 302, thereby providing positive feedback as the output signal transitions from one logic level to another. The feedback loop includes a capacitor 306 that creates AC positive feedback for a small period of time as it charges and discharges in response to the DIFF signal swings. The positive feedback provided by the output signal can overcome noise interferences of the VIN or VREF signals when the signal difference is small to maintain signal integrity at the switching point. If VIN is a periodic signal, such as a clock signal, the capacitance of the capacitor 306 is preferably chosen so that it charges from and discharges into the amplifier 302 over a duration that is less than one-half period of a periodic signal.
A differential input buffer 400 according to one embodiment of the invention is shown in FIG. 4. Similar to the amplifiers 101, 103 of the buffer 100 (FIG. 1), the buffer 400 includes two differential amplifiers 401, 403 whose components are essentially the same, but are complementary configured with respect to each other. The buffer 400 includes many of the same components as the buffer 100 operating in the same manner and, in the interest of brevity, these same components will not be described in detail.
The buffer 400 differs from the conventional buffer 100 shown in FIG. 1 in two respects. Most significantly, as explained in greater detail below, the buffer 400 receives AC positive feedback that makes it more immune to false switching resulting from noise. Second the buffer 400 includes three inverters 440, 442, 444 coupled to the node 108 to invert the OUT_signal and to provide a low impedance to the output at the node 108. The use of three inverters 440, 442, 444 provides greater amplification of the OUT_signal so that the DIFF signal transitions high or low to VCC or to zero, respectively, well prior to the OUT_signal completely transitioning low-to-high or high-to-low.
The AC positive feedback mentioned above is provided by applying the DIFF signal at the output of the inverter 444 to the amplifiers 401, 403 through respective third inputs of the amplifiers 401, 403 at respective nodes 407, 409. The DIFF signal is applied to the nodes 407, 409 through respective capacitors 406, 408 to provide AC positive feedback to increase the drive of the OUT_signal at the node 108 as it transitions high or low. The AC positive feedback does not change the VREF trip point, and provides positive feedback for only a small period of time that the DIFF signal is transitioning from one logic level to another. The positive feedback provided as the VIN signal approaches the VREF reference results in a more stable, uniform transition characteristic since it counteracts any input signal slope reversals due to noise. For example, assume the OUT_signal is transitioning low and the DIFF signal is transitioning high in response to the VIN signal transitioning high. The capacitors 406, 408 couple the low-to-high transition of the DIFF signal to the nodes 407, 409 of the amplifiers 401, 403, causing the voltage at the nodes 407, 409 to be driven high. The increased voltage at the node 407 of the amplifier 401 increases the resistance of the PMOS transistor 116, thereby further decreasing the magnitude of the OUT_signal. The increased voltage at the node 409 of the amplifier 403 decreases the resistance of the NMOS transistor 130, thereby also decreasing the magnitude of the OUT_signal. Thus, a rising VIN signal results in a falling OUT_signal and a rising DIFF signal. The rising DIFF signal further decreases the magnitude of the OUT_signal, thereby providing positive feedback during the time that the DIFF signal is rising. The amplifiers 401, 403 respond to a falling VIN signal in the same manner to provide a falling DIFF signal to the nodes 407, 409 that decrease the resistance of the PMOS transistor 116 in the amplifier 401 and increase the resistance of the NMOS transistor 130 in the amplifier 403, thereby further increasing the OUT_signal.
The amount of positive feedback that is provided depends primarily on the size of the capacitor and gain of the amplifier at the nodes 407, 409, and are determined as part of the design parameters for the particular buffer 400. In the ideal case, the AC positive feedback is provided for less than half the clock cycle of the VIN signal. For example, assuming the OUT_signal is pulled down and the DIFF signal is driven high in response to the OUT_signal. The capacitor 406 couples the low-to-high transition of the DIFF signal to the node 407. The capacitor 406 is then discharged as current is drawn from the capacitor 406 by the node 407. The time constant of the capacitor 406 and resistance at the node 407 should be set so that the capacitor 406 is substantially discharged by the time the DIFF signal transitions low.
The buffer 400 or another buffer according to an embodiment of the invention is shown in a synchronous dynamic random access memory (“SDRAM”) device 500. The SDRAM device 500 includes an address register 512 that receives either a row address or a column address on an address bus 514, preferably by coupling address signals corresponding to the addresses though one embodiment of input buffers 516 according to the present invention. The address bus 514 is generally coupled to a memory controller (not shown). Typically, a row address is initially received by the address register 512 and applied to a row address multiplexer 518. The row address multiplexer 518 couples the row address to a number of components associated with either of two memory banks 520, 522 depending upon the state of a bank address bit forming part of the row address. Associated with each of the memory banks 520, 522 is a respective row address latch 526, which stores the row address, and a row decoder 528, which applies various signals to its respective array 520 or 522 as a function of the stored row address. The row address multiplexer 518 also couples row addresses to the row address latches 526 for the purpose of refreshing the memory cells in the arrays 520, 522. The row addresses are generated for refresh purposes by a refresh counter 530, which is controlled by a refresh controller 532.
After the row address has been applied to the address register 512 and stored in one of the row address latches 526, a column address is applied to the address register 512 and coupled through the input buffers 516. The address register 512 couples the column address to a column address latch 540. Depending on the operating mode of the SDRAM 500, the column address is either coupled through a burst counter 542 to a column address buffer 544, or to the burst counter 542 which applies a sequence of column addresses to the column address buffer 544 starting at the column address output by the address register 512. In either case, the column address buffer 544 applies a column address to a column decoder 548 which applies various signals to respective sense amplifiers and associated column circuitry 550, 552 for the respective arrays 520, 522.
Data to be read from one of the arrays 520, 522 is coupled to the column circuitry 550, 552 for one of the arrays 520, 522, respectively. The data is then coupled through a read data path 554 to a data output register 556. Data from the data output register 556 is coupled to a data bus 558 through data output buffers 559. Data to be written to one of the arrays 520, 522 is coupled from the data bus 558 to a data input register 560 through data input buffers 561 according to an embodiment of the invention. The data input register 560 then couples the write data to the column circuitry 550, 552 where they are transferred to one of the arrays 520, 522, respectively. A mask register 564 may be used to selectively alter the flow of data into and out of the column circuitry 550, 552, such as by selectively masking data to be read from the arrays 520, 522.
The above-described operation of the SDRAM 500 is controlled by a command decoder 568 responsive to command signals received on a control bus 570 though command input buffers 572 according to an embodiment of the invention. These high level command signals, which are typically generated by a memory controller (not shown), are a clock enable signal CKE*, a clock signal CLK, a chip select signal CS*, a write enable signal WE*, a row address strobe signal RAS*, and a column address strobe signal CAS*, which the “*” designating the signal as active low. Various combinations of these signals are registered as respective commands, such as a read command or a write command. The command decoder 568 generates a sequence of control signals responsive to the command signals to carry out the function (e.g., a read or a write) designated by each of the command signals. These command signals, and the manner in which they accomplish their respective functions, are conventional. Therefore, in the interest of brevity, a further explanation of these control signals will be omitted.
Although, the memory device illustrated in FIG. 5 is a synchronous dynamic random access memory (“SDRAM”) 500 that includes the buffer 400 according to an embodiment of the invention, the buffer 400 can be used in other types of memory devices, as well as other types of digital devices.
FIG. 6 shows a computer system 600 containing the SDRAM 500 of FIG. 5. The computer system 600 includes a processor 602 for performing various computing functions, such as executing specific software to perform specific calculations or tasks. The processor 602 includes a processor bus 604 that normally includes an address bus, a control bus, and a data bus. In addition, the computer system 600 includes one or more input devices 614, such as a keyboard or a mouse, coupled to the processor 602 to allow an operator to interface with the computer system 600. Typically, the computer system 600 also includes one or more output devices 616 coupled to the processor 602, such output devices typically being a printer or a video terminal. One or more data storage devices 618 are also typically coupled to the processor 602 to allow the processor 602 to store data in or retrieve data from internal or external storage media (not shown). Examples of typical storage devices 618 include hard and floppy disks, tape cassettes, and compact disk read-only memories (CD-ROMs). The processor 602 is also typically coupled to cache memory 626, which is usually static random access memory (“SRAM”), and to the SDRAM 100 through a memory controller 630. The memory controller 630 is coupled to the SDRAM 500 through the normally control bus 570 and the address bus 514. The data bus 558 is coupled from the SDRAM 500 to the processor bus 604 either directly (as shown), through the memory controller 630, or by some other means.
From the foregoing it will be appreciated that, although specific embodiments of the invention have been described herein for purposes of illustration, various modifications may be made without deviating from the spirit and scope of the invention. For example, many of the components described above may be implemented using either digital or analog circuitry, or a combination of both. Accordingly, the invention is not limited except as by the appended claims.