US20200350002A1 - Static Power Reduction in SRAM - Google Patents
Static Power Reduction in SRAM Download PDFInfo
- Publication number
- US20200350002A1 US20200350002A1 US16/399,942 US201916399942A US2020350002A1 US 20200350002 A1 US20200350002 A1 US 20200350002A1 US 201916399942 A US201916399942 A US 201916399942A US 2020350002 A1 US2020350002 A1 US 2020350002A1
- Authority
- US
- United States
- Prior art keywords
- circuit
- bias
- bit cells
- row
- voltage
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 230000003068 static effect Effects 0.000 title claims abstract description 16
- 230000009467 reduction Effects 0.000 title claims description 9
- 238000000034 method Methods 0.000 claims abstract description 81
- 230000000694 effects Effects 0.000 claims description 8
- 239000000758 substrate Substances 0.000 claims description 3
- 239000003990 capacitor Substances 0.000 description 63
- 230000007704 transition Effects 0.000 description 46
- 238000013461 design Methods 0.000 description 18
- 238000004891 communication Methods 0.000 description 16
- 238000007796 conventional method Methods 0.000 description 10
- 230000008859 change Effects 0.000 description 8
- 230000007246 mechanism Effects 0.000 description 6
- 230000006399 behavior Effects 0.000 description 5
- 230000006872 improvement Effects 0.000 description 5
- 238000012986 modification Methods 0.000 description 4
- 230000004048 modification Effects 0.000 description 4
- 230000002411 adverse Effects 0.000 description 2
- 230000005540 biological transmission Effects 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 230000014759 maintenance of location Effects 0.000 description 2
- 239000000872 buffer Substances 0.000 description 1
- 230000003139 buffering effect Effects 0.000 description 1
- 238000007599 discharging Methods 0.000 description 1
- 230000008054 signal transmission Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G11—INFORMATION STORAGE
- G11C—STATIC STORES
- G11C11/00—Digital stores characterised by the use of particular electric or magnetic storage elements; Storage elements therefor
- G11C11/21—Digital stores characterised by the use of particular electric or magnetic storage elements; Storage elements therefor using electric elements
- G11C11/34—Digital stores characterised by the use of particular electric or magnetic storage elements; Storage elements therefor using electric elements using semiconductor devices
- G11C11/40—Digital stores characterised by the use of particular electric or magnetic storage elements; Storage elements therefor using electric elements using semiconductor devices using transistors
- G11C11/41—Digital stores characterised by the use of particular electric or magnetic storage elements; Storage elements therefor using electric elements using semiconductor devices using transistors forming static cells with positive feedback, i.e. cells not needing refreshing or charge regeneration, e.g. bistable multivibrator or Schmitt trigger
- G11C11/413—Auxiliary circuits, e.g. for addressing, decoding, driving, writing, sensing, timing or power reduction
- G11C11/417—Auxiliary circuits, e.g. for addressing, decoding, driving, writing, sensing, timing or power reduction for memory cells of the field-effect type
-
- G—PHYSICS
- G11—INFORMATION STORAGE
- G11C—STATIC STORES
- G11C5/00—Details of stores covered by group G11C11/00
- G11C5/06—Arrangements for interconnecting storage elements electrically, e.g. by wiring
- G11C5/063—Voltage and signal distribution in integrated semi-conductor memory access lines, e.g. word-line, bit-line, cross-over resistance, propagation delay
-
- G—PHYSICS
- G11—INFORMATION STORAGE
- G11C—STATIC STORES
- G11C11/00—Digital stores characterised by the use of particular electric or magnetic storage elements; Storage elements therefor
- G11C11/21—Digital stores characterised by the use of particular electric or magnetic storage elements; Storage elements therefor using electric elements
- G11C11/34—Digital stores characterised by the use of particular electric or magnetic storage elements; Storage elements therefor using electric elements using semiconductor devices
- G11C11/40—Digital stores characterised by the use of particular electric or magnetic storage elements; Storage elements therefor using electric elements using semiconductor devices using transistors
- G11C11/41—Digital stores characterised by the use of particular electric or magnetic storage elements; Storage elements therefor using electric elements using semiconductor devices using transistors forming static cells with positive feedback, i.e. cells not needing refreshing or charge regeneration, e.g. bistable multivibrator or Schmitt trigger
- G11C11/412—Digital stores characterised by the use of particular electric or magnetic storage elements; Storage elements therefor using electric elements using semiconductor devices using transistors forming static cells with positive feedback, i.e. cells not needing refreshing or charge regeneration, e.g. bistable multivibrator or Schmitt trigger using field-effect transistors only
-
- G—PHYSICS
- G11—INFORMATION STORAGE
- G11C—STATIC STORES
- G11C11/00—Digital stores characterised by the use of particular electric or magnetic storage elements; Storage elements therefor
- G11C11/21—Digital stores characterised by the use of particular electric or magnetic storage elements; Storage elements therefor using electric elements
- G11C11/34—Digital stores characterised by the use of particular electric or magnetic storage elements; Storage elements therefor using electric elements using semiconductor devices
- G11C11/40—Digital stores characterised by the use of particular electric or magnetic storage elements; Storage elements therefor using electric elements using semiconductor devices using transistors
- G11C11/41—Digital stores characterised by the use of particular electric or magnetic storage elements; Storage elements therefor using electric elements using semiconductor devices using transistors forming static cells with positive feedback, i.e. cells not needing refreshing or charge regeneration, e.g. bistable multivibrator or Schmitt trigger
- G11C11/413—Auxiliary circuits, e.g. for addressing, decoding, driving, writing, sensing, timing or power reduction
- G11C11/417—Auxiliary circuits, e.g. for addressing, decoding, driving, writing, sensing, timing or power reduction for memory cells of the field-effect type
- G11C11/419—Read-write [R-W] circuits
-
- G—PHYSICS
- G11—INFORMATION STORAGE
- G11C—STATIC STORES
- G11C5/00—Details of stores covered by group G11C11/00
- G11C5/02—Disposition of storage elements, e.g. in the form of a matrix array
- G11C5/025—Geometric lay-out considerations of storage- and peripheral-blocks in a semiconductor storage device
-
- G—PHYSICS
- G11—INFORMATION STORAGE
- G11C—STATIC STORES
- G11C5/00—Details of stores covered by group G11C11/00
- G11C5/14—Power supply arrangements, e.g. power down, chip selection or deselection, layout of wirings or power grids, or multiple supply levels
- G11C5/147—Voltage reference generators, voltage or current regulators; Internally lowered supply levels; Compensation for voltage drops
-
- G—PHYSICS
- G11—INFORMATION STORAGE
- G11C—STATIC STORES
- G11C11/00—Digital stores characterised by the use of particular electric or magnetic storage elements; Storage elements therefor
- G11C11/21—Digital stores characterised by the use of particular electric or magnetic storage elements; Storage elements therefor using electric elements
- G11C11/34—Digital stores characterised by the use of particular electric or magnetic storage elements; Storage elements therefor using electric elements using semiconductor devices
- G11C11/40—Digital stores characterised by the use of particular electric or magnetic storage elements; Storage elements therefor using electric elements using semiconductor devices using transistors
- G11C11/41—Digital stores characterised by the use of particular electric or magnetic storage elements; Storage elements therefor using electric elements using semiconductor devices using transistors forming static cells with positive feedback, i.e. cells not needing refreshing or charge regeneration, e.g. bistable multivibrator or Schmitt trigger
- G11C11/413—Auxiliary circuits, e.g. for addressing, decoding, driving, writing, sensing, timing or power reduction
- G11C11/417—Auxiliary circuits, e.g. for addressing, decoding, driving, writing, sensing, timing or power reduction for memory cells of the field-effect type
- G11C11/418—Address circuits
-
- G—PHYSICS
- G11—INFORMATION STORAGE
- G11C—STATIC STORES
- G11C2207/00—Indexing scheme relating to arrangements for writing information into, or reading information out from, a digital store
- G11C2207/22—Control and timing of internal memory operations
- G11C2207/2227—Standby or low power modes
-
- G—PHYSICS
- G11—INFORMATION STORAGE
- G11C—STATIC STORES
- G11C8/00—Arrangements for selecting an address in a digital store
- G11C8/12—Group selection circuits, e.g. for memory block selection, chip selection, array selection
Definitions
- the present invention relates to the field of integrated circuits.
- the present invention relates to a circuit for reducing static power in SRAM and methods for using the same.
- FIG. 1A illustrates a modeling of a signal line between a transmitter and a receiver.
- a transmitter 102 may be configured to transmit a signal through a signal line to a receiver 104 .
- the transmitter 102 drives a transmitted signal at node 103 and the receiver 104 receives a received signal at node 105 .
- the signal line may be modelled as a series of resistor-capacitor networks, labelled as 106 a through 106 j .
- the propagation delay through the signal line depends on electrical characteristics of the signal line, which is a function of the series of resistor-capacitor networks.
- FIG. 1B illustrates a conventional method of signal communication among a transmitter and multiple receivers in a system.
- a transmitter 112 may be configured to transmit a signal through a signal line to a plurality of receivers, namely 114 a through 114 j .
- the signal line may be modelled as a series of resistor-capacitor networks, labelled as 116 a through 116 m .
- a plurality of buffers such as 118 a through 118 j , may be used to reduce propagation delay and ensure sufficient signal strength received at a receiver.
- the communication between transmitter 112 and a receiver such as receiver J 114 j becomes one directional. In other words, another set of signal line would be used to transmit a signal from receiver J 114 j to transmitter 112 .
- the number of receiver increases in a system with a 32 bit or wider system bus, the number of signal paths and the components and areas used to support such signal paths increase exponentially, which can significantly increase the cost and complexity of the system.
- FIG. 1C illustrates a conventional design of a subsection of a SRAM.
- a subsection of a SRAM may include memory bank 0 ( 122 a ) and memory bank 1 ( 122 b ).
- Memory bank 0 may be accessed through column multiplexor 0 ( 124 a ) and similarly memory bank 1 may be accessed through column multiplexor 1 ( 124 b ).
- a set of sense amplifiers 126 are shared by the memory bank 0 and memory bank 1 through the column multiplexor 0 and column multiplexor 1, respectively.
- the set of sense amplifiers 126 drive a set of global bit lines corresponding to local bit lines of memory bank 0 and memory bank 1.
- This subsection of SRAM can be part of a SRAM memory block, where multiple SRAM memory blocks may form the memory core of the SRAM.
- FIG. 1D illustrates a conventional design of a SRAM memory bank of FIG. 1C .
- a memory bank may include multiple rows and columns of bit cells (BCs), represented by BC 128 .
- BCs bit cells
- a 32-bit memory bank width is shown.
- different memory bank widths may be designed, such as memory bank width of 8, 16, 64, 128, 256 bits may be implemented.
- each row of the memory bank is controlled by a word line, for example WL 0 controls the top row of the memory bank.
- Each column of the memory bank is controlled by a pair of differential signal lines, for example Bit0 and its inverse _Bit0 control the left column, and Bit31 and its inverse _Bit31 control the left column.
- a typical conventional design would pack many rows of bit cells into a memory bank. To reduce power consumption, each bit cell in the typical conventional design would have a weak drive strength. These design criteria impose adverse impact on the performance and power consumption of the SRAM.
- FIG. 1E illustrates a conventional design of a bit cell of a SRAM.
- a bit cell may include transistors M 1 , M 2 , M 3 , M 4 , M 4 , M 5 , and M 6 ; and the transistors M 1 , M 2 , M 3 , M 4 , M 4 , M 5 , and M 6 are connected in the manner shown in FIG. 1E .
- the transistors are controlled by a word line (for example WL J ) and by a pair of differential bit lines (for example Bit K and _BitK).
- the drain terminals of transistors M 5 , and M 6 are typically pulled up to Vdd, and the source terminals of transistors M 3 and M 4 are coupled to Vss.
- one leakage path is through M 1 and M 3 ; another leakage path is through M 5 and M 3 ; yet another leakage path is through M 6 and M 4 ; and yet another leakage path is through M 2 and M 4 .
- the leakage through the NMOS transistors are larger compare to the leakage through the PMOS transistors, because the size of the NMOS transistors is larger than the size of the PMOS transistors.
- FIG. 1F illustrates a conventional method of static power reduction in a SRAM. As shown in the conventional method of FIG. 1F , during the access mode, the drain terminals of transistors M 5 , and M 6 are typically pulled up to Vdd, and the source terminals of transistors M 3 and M 4 are coupled to Vss.
- the drain terminals of transistors M 5 , and M 6 are typically pulled up to a retention voltage V RETENTION , which is lower than Vdd, and the source terminals of transistors M 3 and M 4 remain at Vss. In this manner, power leakage can be reduced.
- V RETENTION retention voltage
- one drawback with the above conventional method is that there is a long delay to transition from the standby mode back to the access mode, because the adjustment of power supply is done to the entire SRAM or may be done to the entire circuit that includes the SRAM. The long delay makes frequent switching between the access mode and standby mode impractical when a device is being used.
- a circuit for reducing static power in SRAM includes a plurality of memory blocks, where a memory block in the plurality of memory blocks includes a plurality memory banks, where a memory bank in the plurality of memory banks includes a plurality bit cells.
- the circuit further includes a bias circuit configured to produce a bias voltage to a row of bit cells, where the bias circuit is coupled to a circuit ground terminal of the row of bit cells in the plurality of bit cells, and a controller configured to control the bias circuit to produce a first set of bias settings in an access mode and control the bias circuit to produce a second set of bias settings in a standby mode of the SRAM.
- a method for reducing static power in SRAM includes providing a plurality of memory blocks, where a memory block in the plurality of memory blocks includes a plurality memory banks, where a memory bank in the plurality of memory banks includes a plurality bit cells; producing, by a bias circuit, a bias voltage to a row of bit cells, where the bias circuit is coupled to a circuit ground terminal of the row of bit cells in the plurality of bit cells; and controlling, by a controller, the bias circuit to produce a first set of bias settings in an access mode and control the bias circuit to produce a second set of bias settings in a standby mode of the SRAM.
- FIG. 1A illustrates a modeling of a signal line between a transmitter and a receiver.
- FIG. 1B illustrates a conventional method of signal communication among a transmitter and multiple receivers in a system.
- FIG. 1C illustrates a conventional design of a SRAM data signals.
- FIG. 1D illustrates a conventional design of a SRAM memory bank of FIG. 1C .
- FIG. 1E illustrates a conventional design of a bit cell of a SRAM.
- FIG. 1F illustrates a conventional method of static power reduction in a SRAM.
- FIG. 2A illustrates an exemplary representation of signal communication between two transceivers of a system according to aspects of the present disclosure.
- FIG. 2B illustrates an exemplary representation of signal communication among multiple transceivers in a system according to aspects of the present disclosure.
- FIG. 3A illustrates an exemplary implementation of a transmitter of FIG. 2A according to aspects of the present disclosure.
- FIG. 3B illustrates an exemplary implementation of a transmitter of FIG. 2A according to aspects of the present disclosure.
- FIG. 3C illustrates another exemplary implementation of a transmitter of FIG. 2A according to aspects of the present disclosure.
- FIG. 3D illustrates another exemplary implementation of a transmitter of FIG. 2A according to aspects of the present disclosure.
- FIG. 3E illustrates an exemplary implementation of a receiver of FIG. 2A according to aspects of the present disclosure.
- FIG. 3F illustrates another exemplary implementation of a receiver of FIG. 2A according to aspects of the present disclosure.
- FIG. 3G illustrates yet another exemplary implementation of a receiver of FIG. 2A according to aspects of the present disclosure.
- FIG. 4A illustrates exemplary node voltages at an output node of a transmitter and an input node of a receiver based on a conventional implementation of low to high signal transition.
- FIG. 4B illustrates exemplary node voltages at an output node of a transmitter and an input node of a receiver based on the implementation of FIG. 2A for a low to high signal transition according to aspects of the present disclosure.
- FIG. 4C illustrates another exemplary node voltages at an output node of a transmitter and an input node of a receiver based on a conventional implementation of high to low signal transition.
- FIG. 4D illustrates another exemplary node voltages at an output node of a transmitter and an input node of a receiver based on the implementation of FIG. 2A for a high to low signal transition according to aspects of the present disclosure.
- FIG. 4E illustrates yet another exemplary node voltages at an output node of a transmitter and an input node of a receiver based on the implementation of FIG. 3C for a low to high signal transition.
- FIG. 4F illustrates yet another exemplary node voltages at an output node of a transmitter and an input node of a receiver based on the implementation of FIG. 3C for a high to low signal transition.
- FIG. 4G illustrates an exemplary implementation of a data bus and its control signals according to aspects of the present disclosure.
- FIG. 5A illustrates an exemplary method of signal communication between a transmitter and a receiver according to aspects of the present disclosure.
- FIG. 5B illustrates another exemplary method of signal communication between a transmitter and a receiver according to aspects of the present disclosure.
- FIG. 5C illustrates an exemplary method of detecting a transmitted signal according to aspects of the present disclosure.
- FIG. 5D illustrates yet another exemplary method of signal communication between a transmitter and a receiver according to aspects of the present disclosure.
- FIG. 6A illustrates an exemplary implementation of a SRAM according to aspects of the present disclosure.
- FIG. 6B illustrates an exemplary implementation of SRAM signal lines according to aspects of the present disclosure.
- FIG. 6C illustrates another exemplary implementation of SRAM signal lines according to aspects of the present disclosure.
- FIG. 6D illustrates an exemplary bridge circuit according to aspects of the present disclosure.
- FIG. 6E illustrates another exemplary bridge circuit according to aspects of the present disclosure.
- FIG. 6F illustrates an exemplary holding circuit according to aspects of the present disclosure.
- FIG. 7A illustrates an exemplary method of reducing dynamic power in a SRAM according to aspects of the present disclosure.
- FIG. 7B illustrates an exemplary implementation of bridging between a set of memory bank signal lines and a set of memory block signal lines according to aspects of the present disclosure.
- FIG. 7C illustrates an exemplary method of performing a read operation of a SRAM according to aspects of the present disclosure.
- FIG. 8A illustrates an exemplary implementation of controlling a bit cell in a SRAM according to aspects of the present disclosure.
- FIG. 8B illustrates an exemplary implementation of generating a bias voltage to circuit ground of a bit cell in a SRAM according to aspects of the present disclosure.
- FIG. 8C illustrates an exemplary implementation of controlling a word line of a bit cell in a SRAM according to aspects of the present disclosure.
- FIG. 8D illustrates an exemplary plot of drain to source voltage of a NMOS transistor in a bit cell of a SRAM according to aspects of the present disclosure.
- FIG. 9A illustrates an exemplary method of reducing static power in a SRAM according to aspects of the present disclosure.
- FIG. 9B illustrates an exemplary method of controlling a bias circuit according to aspects of the present disclosure.
- FIG. 2A illustrates an exemplary representation of signal communication between two transceivers of a system according to aspects of the present disclosure.
- a first transceiver 202 may be configured to transmit a signal through a signal line 206 to a second transceiver 204 .
- the first transceiver 202 may include a first transmitter 203 and a first receiver 205 .
- the second transceiver 204 may include a second transmitter 207 and a second receiver 209 .
- the signal line may be modelled as a series of resistor-capacitor networks, and the propagation delay through the signal line depends on electrical characteristics of the signal line, which is a function of the series of resistor-capacitor networks.
- transceiver 1 ( 202 ) and transceiver 2 ( 204 ) are bi-directional.
- transmitter 203 of transceiver 1 ( 202 ) may be configured to send a transmitted signal through the signal line 206 and be received as a received signal at receiver 209 of transceiver 2 ( 204 ).
- transmitter 207 of transceiver 2 ( 204 ) may be configured to send a transmitted signal through the signal line 206 and be received as a received signal at receiver 205 of transceiver 1 ( 202 ).
- FIG. 2B illustrates an exemplary representation of signal communication among multiple transceivers in a system according to aspects of the present disclosure.
- a first transceiver TR 1 212 may be configured to communication with a plurality of transceivers, labelled as TR 2 214 a , TR 3 214 b , through TR 1 214 j , etc. Similar to FIG. 2A , the communication between TR 1 212 and the plurality of transceivers (TR 2 214 a , TR 3 214 b , through TR J 214 j ) are bi-directional.
- FIG. 3A illustrates an exemplary implementation of a transmitter of FIG. 2A according to aspects of the present disclosure.
- a transmitter may include a driver capacitor C DRIVE , and a first transistors M 1 configured to discharge the driver capacitor C DRIVE .
- the transmitter may further include a second transistor M 2 configured to charge a node 303 at the transmitter end of a signal line such as signal line 206 .
- the transmitter may further include a switch S 1 configured to connect or separate the driver capacitor C DRIVE and the signal line during different phases of an operation.
- the driver capacitor C DRIVE , the first transistor M 1 , the second transistor M 2 , and the switch S 1 are connected in the manner shown in FIG. 3A .
- the driver capacitor C DRIVE is discharged to logic low through the control of the M 1 transistor while node 303 is charged to logic high through the control of M 2 transistor. Then in phase two, switch S 1 is closed, causing charge sharing among the driver capacitor C DRIVE and the capacitors of the signal line, resulting in a controlled high to low voltage transition at the receiver end of the receiving transceiver.
- the controlled high to low voltage transition may have a magnitude that is proportional to a ratio of capacitance between the driver capacitor C DRIVE and the capacitors of the signal line.
- controlled low to high voltage transition can be captured and sensed by a sense amplifier to determine the high to low signal transition by the receiver at the receiving transceiver.
- the transistors M 1 , M 2 , and the switch S 1 are controlled in the manner as described above for a high to low signal transition.
- FIG. 3B illustrates an exemplary implementation of a transmitter of FIG. 2A according to aspects of the present disclosure.
- a transmitter may include a driver capacitor C DRIVE , and a first transistors M 1 configured to charge the driver capacitor C DRIVE .
- the transmitter may further include a second transistor M 2 configured to discharge a node 313 at the transmitter end of a signal line such as signal line 206 .
- the transmitter may further include a switch S 1 configured to connect or separate the driver capacitor C DRIVE and the signal line during different phases of an operation.
- the driver capacitor C DRIVE , the first transistor M 1 , the second transistor M 2 , and the switch S 1 are connected in the manner shown in FIG. 3B .
- the driver capacitor C DRIVE is charged to logic high through the control of the M 1 transistor while node 313 is discharged to logic low through the control of M 2 transistor. Then in phase two, switch S 1 is closed, causing charge sharing among the driver capacitor C DRIVE and the capacitors of the signal line, resulting in a controlled low to high voltage transition at the receiver end of the receiving transceiver.
- the controlled low to high voltage transition may have a magnitude that is proportional to a ratio of capacitance between the driver capacitor C DRIVE and the capacitors of the signal line.
- controlled low to high voltage transition can be captured and sensed by a sense amplifier to determine the low to high signal transition by the receiver at the receiving transceiver.
- the transistors M 1 , M 2 , and the switch S 1 are controlled in the manner as described above for a low to high signal transition.
- FIG. 3C illustrates another exemplary implementation of a transmitter of FIG. 2A according to aspects of the present disclosure.
- a transmitter may include a driver capacitor C DRIVE , and a first transistor M 1 configured to discharge the driver capacitor C DRIVE .
- the transmitter may further include a second transistor M 2 configured to charge a node 323 at the transmitter end of a signal line such as signal line 216 .
- the transmitter may further include a switch S 1 configured to connect or separate the driver capacitor and the signal line during different phases of an operation.
- the transmitter further includes a pulse generator circuit 322 configured to generate control signals to control transistors M 1 , M 2 , and the switch S 1 .
- the driver capacitor C DRIVE , the first transistor M 1 , the second transistor M 2 , the switch S 1 , and the pulse generator circuit 322 are connected in the manner shown in FIG. 3C .
- the driver capacitor C DRIVE is discharged to logic low through the control of the M 1 transistor while node 323 is charged to logic high through the control of M 2 transistor.
- switch S 1 is closed, causing charge sharing among the driver capacitor C DRIVE and the capacitors of the signal line, resulting in a controlled high to low voltage transition at the receiver end of the receiving transceiver.
- the controlled high to low voltage transition may have a magnitude that is proportional to a ratio of capacitance between the driver capacitor C DRIVE and the capacitors of the signal line.
- the controlled low to high voltage transition can be captured and sensed by a sense amplifier to determine the high to low signal transition by the receiver at the receiving transceiver.
- the pulse generator circuit 322 generates control signals configured to control transistors M 1 , M 2 , and the switch S 1 , in the manner as described above for a high to low signal transition.
- the pulse generator circuit 322 is configured to generate a pulse onto the signal line based on the signal to be transmitted; and where a duration of the pulse is programmable.
- FIG. 3D illustrates another exemplary implementation of a transmitter of FIG. 2A according to aspects of the present disclosure.
- a transmitter may include a driver capacitor C DRIVE , and a first transistor M 1 configured to charge the driver capacitor C DRIVE .
- the transmitter may further include a second transistor M 2 configured to discharge a node 325 at the transmitter end of a signal line such as signal line 216 .
- the transmitter may further include a switch S 1 configured to connect or separate the driver capacitor C DRIVE and the signal line during different phases of an operation.
- the transmitter further includes a pulse generator circuit 324 configured to generate control signals to control transistors M 1 , M 2 , and the switch S 1 .
- the driver capacitor C DRIVE , the first transistor M 1 , the second transistor M 2 , the switch S 1 , and the pulse generator circuit 324 are connected in the manner shown in FIG. 3D .
- the driver capacitor C DRIVE is charged to logic high through the control of the M 1 transistor while node 325 is discharged to logic low through the control of M 2 transistor. Then in phase two, switch S 1 is closed, causing charge sharing among the driver capacitor C DRIVE and the capacitors of the signal line, resulting in a controlled low to high voltage transition at the receiver end of the receiving transceiver.
- the controlled low to high voltage transition may have a magnitude that is proportional to a ratio of capacitance between the driver capacitor C DRIVE and the capacitors of the signal line.
- the controlled low to high voltage transition can be captured and sensed by a sense amplifier to determine the low to high signal transition by the receiver at the receiving transceiver.
- the pulse generator circuit 324 generates control signals configured to control transistors M 1 , M 2 , and the switch S 1 , in the manner as described above for a low to high signal transition.
- the pulse generator circuit 324 is configured to generate a pulse onto the signal line based on the signal to be transmitted; and where a duration of the pulse is programmable.
- FIG. 3E illustrates an exemplary implementation of a receiver of FIG. 2A according to aspects of the present disclosure.
- a receiver may include transistors M 1 , M 2 , and M 7 , a pair of cross-coupled inverters implemented by M 3 , M 4 , M 5 and M 6 , a latch implemented by a pair of cross-coupled Nand gates N 1 and N 2 .
- the transistors M 1 , M 2 , M 3 , M 4 , M 5 , M 6 , M 7 , and the pair of cross-coupled Nand gates N 1 and N 2 are connected in the manner shown in FIG. 3E .
- the receiver receives an input signal (In) at transistor M 2 , and a control signal (En_In) configured to control transistors M 1 , M 2 , and M 7 .
- the receiver is configured to generate an output signal (Out).
- transistors M 1 , M 2 may be replaced with switches.
- FIG. 3F illustrates another exemplary implementation of a receiver of FIG. 2A according to aspects of the present disclosure.
- a receiver may include transistors M 1 , M 2 , and M 7 , a pair of cross-coupled inverters implemented by M 3 , M 4 , M 5 and M 6 , a latch implemented by a pair of cross-coupled Nand gates N 1 and N 2 .
- the transistors M 1 , M 2 , M 3 , M 4 , M 5 , M 6 , M 7 , and the pair of cross-coupled Nand gates N 1 and N 2 are connected in the manner shown in FIG. 3F .
- the receiver receives an input signal (In) at transistor M 1 , and a control signal (En_In) configured to control transistors M 1 , M 2 , and M 7 .
- the receiver is configured to generate an output signal (Out).
- transistors M 1 , M 2 may be replaced with switches.
- FIG. 3G illustrates yet another exemplary implementation of a receiver of FIG. 2A according to aspects of the present disclosure.
- the receiver includes transistors M 1 , M 2 , M 3 , and a receive capacitor C R .
- the transistors M 1 , M 2 , M 3 , receive capacitor C R , an input signal (In), and an output signal Out are connected in the manner shown in FIG. 3G .
- a transmitter circuit such as the transmitter circuit of FIG. 3A or the transmitter circuit of FIG. 3C , is configured to cause a deviation in current at a receiver node of the signal line.
- a receiver circuit such as the receiver circuit of FIG. 3E or the receiver circuit of FIG. 3F , is configured to reproduce the transmitted signal using the deviation in current at the receiver node of the signal line.
- the receiver circuit includes a current mirror, formed by M 2 and M 3 , configured to multiply the deviation in current detected at the receiver node (input In) of the signal line, the receive capacitor C R is configured to store charges produced by the current mirror, where the charges represent a signal strength of the transmitted signal.
- FIG. 4A illustrates exemplary node voltages at an output node of a transmitter and an input node of a receiver based on a conventional implementation of low to high signal transition.
- line 402 represents a node voltage at an output node of a transmitter.
- Line 404 represents a node voltage at an input node of a receiver.
- the voltage at the output node of the transmitter can change quickly.
- the voltage at the input node of the receiver can rise slowly. There can be a significant delay before the change in voltage at the input node of the receiver can be recognized. This delay is represented by the time T 1 .
- FIG. 4B illustrates exemplary node voltages at an output node of a transmitter and an input node of a receiver based on the implementation of FIG. 2A for a low to high signal transition according to aspects of the present disclosure.
- line 406 represents a node voltage at an output node of a transmitter.
- Line 408 represents a node voltage at an input node of a receiver. The voltage at the output node of the transmitter can change quickly.
- the transmitter circuit is configured to cause a deviation in voltage at a receiver node of the signal line and the receiver circuit is configured to detect the deviation in voltage from a reference voltage at the receiver node of the signal line, the magnitude of the voltage deviation, represented by Vs1 with respect to Vss, to be sensed at the receiver end is significantly reduced.
- a steady voltage of Vs1 at the input node of the receiver can be reached in significantly shorter time, represented by T 2 .
- T 2 time
- the time delay (T 2 vs T 1 ) before the receiver can start sensing a low to high signal transition on the signal line can be significantly reduced with the disclosed implementation, which can lead to significant improvement in performance of the system.
- the amount of voltage change at the input node of the receiver (line 408 vs line 404 ) can be significantly reduced with the disclosed implementation, and this reduction in voltage swing can result in power savings for the system.
- FIG. 4C illustrates another exemplary node voltages at an output node of a transmitter and an input node of a receiver based on a conventional implementation of high to low signal transition.
- line 412 represents a node voltage at an output node of a transmitter.
- Line 414 represents a node voltage at an input node of a receiver.
- the voltage at the output node of the transmitter can change quickly.
- the voltage at the input node of the receiver can decrease slowly. There can be a significant delay before the change in voltage at the input node of the receiver can be recognized. This delay is represented by the time T 3 .
- FIG. 4D illustrates another exemplary node voltages at an output node of a transmitter and an input node of a receiver based on the implementation of FIG. 2A for a high to low signal transition according to aspects of the present disclosure.
- line 416 represents a node voltage at an output node of a transmitter.
- Line 418 represents a node voltage at an input node of a receiver. The voltage at the output node of the transmitter can change quickly.
- the transmitter circuit is configured to cause a deviation in voltage at a receiver node of the signal line and the receiver circuit is configured to detect the deviation in voltage from a reference voltage at the receiver node of the signal line, the magnitude of the voltage deviation, represented by Vs2 with respect to Vdd, to be sensed at the receiver end is significantly reduced.
- Vs2 the magnitude of the voltage deviation
- Vdd the voltage deviation
- the time delay (T 4 vs T 3 ) before the receiver can start sensing a high to low signal transition on the signal line can be significantly reduced with the disclosed implementation, which can lead to significant improvement in performance of the system.
- the amount of voltage change at the input node of the receiver (line 418 vs line 414 ) can be significantly reduced with the disclosed implementation, and this reduction in voltage swing can result in power savings for the system.
- FIG. 4E illustrates yet another exemplary node voltages at an output node of a transmitter and an input node of a receiver based on the implementation of FIG. 3C for a low to high signal transition.
- line 422 represents a node voltage at an output node of a transmitter based on the exemplary implementation FIG. 3C for a low to high signal transition.
- Line 424 represents a node voltage at an input node of a receiver based on the exemplary implementation FIG. 3C for a low to high signal transition.
- the receiver may be configured to sense the voltage deviation Vs1 with respect to Vss at time T 5 .
- the transmitter circuit can stop driving the signal line after a duration controlled by the pulse generator circuit, for example after time T 6 .
- the input node at the receiver may be returned to a standby state, which is Vss in this example.
- the behavior of the transmission mechanism is similar to that of FIG. 4B , and the description is not repeated here.
- FIG. 4F illustrates yet another exemplary node voltages at an output node of a transmitter and an input node of a receiver based on the implementation of FIG. 3C for a high to low signal transition.
- line 426 represents a node voltage at an output node of a transmitter based on the exemplary implementation FIG. 3C for a low to high signal transition.
- Line 428 represents a node voltage at an input node of a receiver based on the exemplary implementation FIG. 3C for a low to high signal transition.
- the receiver may be configured to sense the voltage deviation Vs2 with respect to Vdd at time T 8 .
- the transmitter circuit can stop driving the signal line after a duration controlled by the pulse generator circuit, for example after time T 9 .
- the input node at the receiver may be returned to a standby state, which is Vdd in this example.
- the behavior of the transmission mechanism is similar to that of FIG. 4D , and the description is not repeated here.
- FIG. 4G illustrates an exemplary implementation of a data bus and its control signals according to aspects of the present disclosure.
- transceiver 1 ( 432 ) is configured to communicate with transceiver 2 ( 434 ) using a signal bus.
- Transceiver 1 receives a 32-bit input data (Data_in) and a control signal Enable_In (En_In), and produces a 32-bit output data (Data_Out), along with control signals (En_Out) and an optional pre-charge timing reference (PT_Ref).
- each line of the 32-bit output Data_Out is transmitted using the mechanism and/or circuits described in association with FIG. 2A-2B , FIG. 3A-3E , and FIG.
- control signals En_out and PT_Ref may be communicated to transceiver 2 using a conventional buffering mechanism.
- the PT_Ref signal is configured to assist the receiver about the timing and state of charging and/or discharging of the output node at the transmitter of transceiver 1, for example in implementations described in association with FIG. 4E and FIG. 4F .
- FIG. 5A illustrates an exemplary method of signal communication between a transmitter and a receiver according to aspects of the present disclosure.
- the method drives a transmitted signal onto a signal line.
- the method detects the transmitted signal based on a deviation of a received signal from a reference signal on the signal line.
- the method uses the received signal to communicate the transmitted signal.
- the signal line is configured to transmit bi-directional signals for different periods of signal transmission. An amount of deviation of the signal from the reference signal to be detected by the receiver is programmable.
- FIG. 5B illustrates another exemplary method of signal communication between a transmitter and a receiver according to aspects of the present disclosure.
- the method causes, by the transmitter circuit, a deviation in voltage at a receiver node of the signal line; and detects, by the receiver circuit, the deviation in voltage from a reference voltage at the receiver node of the signal line.
- the transmitter circuit includes a capacitor configured to share charges with the signal line; a pull down transistor configured to discharge the capacitor via control of a first switch and a second switch, or a pull up transistor configured to charge the capacitor via control of the first switch and the second switch; where the pull down transistor or the pull up transistor is controlled to drive the signal line.
- the methods performed in blocks 512 and 514 may additionally and/or optionally include the methods performed in blocks 516 , 518 , and 520 .
- the method pre-charges the capacitor and drives the signal line using the pre-charged capacitor.
- the method discharges the capacitor and drives the signal line using the discharged capacitor.
- the method generates, by a pulse generator, a pulse onto the signal line based on the transmitted signal; where the pulse generator is coupled to the pull down transistor and the pull up transistor, and where a duration of the pulse is programmable.
- FIG. 5C illustrates an exemplary method of detecting a transmitted signal according to aspects of the present disclosure.
- the method detects, by a sense amplifier, the deviation of the received signal from the reference signal on the receiver node of the signal line.
- the method performed in block 522 may additionally and/or optionally include the method performed in block 524 .
- the method maintains, by a holding circuit, an input voltage at the receiver node of the signal line.
- FIG. 5D illustrates yet another exemplary method of signal communication between a transmitter and a receiver according to aspects of the present disclosure.
- the method causes, by the transmitter circuit, a deviation in current at a receiver node of the signal line; and reproduces, by the receiver circuit, the transmitted signal using the deviation in current at the receiver node of the signal line.
- the method performed in block 532 may additionally and/or optionally include the method performed in block 534 .
- the method multiplies, by a current mirror, the deviation in current detected at the receiver node of the signal line; and stores, by a capacitor, charges produced by the current mirror, where the charges represent a signal strength of the transmitted signal.
- FIG. 6A illustrates an exemplary implementation of a SRAM according to aspects of the present disclosure.
- a SRAM may include a plurality of memory blocks.
- a memory block in the plurality of memory blocks may include a plurality memory banks, such as memory banks 0 through 3, memory bank L ⁇ 1, memory bank L, memory bank W ⁇ 1, and memory bank W, labelled 602 a through 602 h , respectively.
- a memory bank in the plurality of memory banks includes a plurality bit cells, such as the bit cell shown in FIG. 1E .
- two adjacent memory banks may share a bridge circuit.
- memory bank 0 and memory bank 1 share bridge circuit 1 (labeled as 604 a ).
- a block of a SRAM may include a set of memory bank signal lines, labelled as 603 a through 603 h , and a corresponding bridge circuit. A bit cell in a memory bank can be accessed through the set of memory bank signal lines.
- the block of a SRAM may further include a set of memory block signal lines 606 shared across the plurality of memory banks in the memory block through bridge circuits (such as 604 a , 604 b , 6041 and 604 w ) couple between the set of memory bank signal lines and the set of memory block signal lines 606 .
- the set of memory bank signal lines includes a plurality of differential bit lines.
- the block of a SRAM may further include a set of sense amplifiers 608 corresponding to the set of memory block signal lines and additionally/optionally include holding circuits, such as holding circuit 610 , where the set of sense amplifiers are shared among the plurality of memory banks in the memory block.
- a controller (not shown) may be used to control an access of one or more bit cells in the plurality bit cells.
- the controller may include an address decoder configured to generate control signals to the bridge circuit and the plurality bit cells.
- FIG. 6B illustrates an exemplary implementation of SRAM signal lines according to aspects of the present disclosure.
- a signal path of FIG. 6A is shown, for example from memory bank 0 ( 602 a ) via bank signal lines ( 603 a ), bridge circuit ( 604 a ), memory block signal lines 606 , to sense amplifiers 608 .
- the bridge circuit 604 a can be configured to disconnect the set of memory bank signal lines 603 a and the set of memory block signal lines 606 during a standby mode of a memory bank, such as memory bank 0 or memory bank 1.
- the bridge circuit 604 a can be further configured to connect the set of memory bank signal lines 603 a and the set of memory block signal lines 606 during an access mode of a memory bank.
- the bridge circuit 604 a can be configured to cause a deviation in voltage in the plurality pairs of differential bit lines in a memory bank.
- the set of sense amplifiers 608 can be configured to detect the deviation in voltage in the plurality pairs of differential bit lines, as described above in association with FIG. 3A-3E and FIG. 4A-4F .
- FIG. 6C illustrates another exemplary implementation of SRAM signal lines according to aspects of the present disclosure. Similar to the example to FIG. 6B , a signal path of FIG. 6A is shown, for example from memory bank 0 via bank signal lines 603 a , bridge circuit ( 604 a ), a first optional holding circuit 610 a , memory block signal lines 606 , a second optional holding circuits 610 b , to sense amplifiers 608 .
- a holding circuit, such as 610 a or 610 b can be configured to maintain a voltage at an output node of the bridge circuit 604 a or at an input node of a sense amplifier 608 .
- the set of sense amplifiers 608 can be configured to detect the deviation in voltage in the plurality pairs of differential bit lines, as described above in association with FIG. 3A-3E and FIG. 4A-4F .
- a sense amplifier may include a differential amplifier configured to detect a voltage differential from a pair of memory block signal lines.
- FIG. 6D illustrates an exemplary bridge circuit according to aspects of the present disclosure.
- a bridge circuit may include a pull up transistor M 1 , a switch S, and a pull down transistor M 2 .
- the pull up transistor M 1 , switch S, and pull down transistor M 2 are connected in the manner as shown in FIG. 6D .
- the pull-up transistor M 1 can be configured to charge the memory bank signal line, and the pull up transistor M 1 can also be controlled to drive the memory block signal line.
- the bridge circuit can be configured to pre-charge a pair of memory bank sign lines to a logic high, where the pair of memory bank sign lines includes a bit line and an inverse of the bit line of a row of bit cells to be read.
- a word line can be activated to control the row of bit cells, where each bit cell is coupled to a pair of memory bank sign lines, which are connected to a pair memory block sign lines via the bridge circuit.
- charges stored in the pre-charged pair of memory bank sign lines are shared with the capacitors (as shown in the signal line model of FIG. 1A ) of the pair memory block sign lines via the bridge circuit.
- This sharing of charges causes a deviation in voltage or current at the input node(s) of a sense amplifier.
- the sense amplifier is configured to sense the voltage differential or current differential of the pair of memory block signal lines caused by the shared charges from the pre-charged pair of memory bank sign lines to determine a value stored in the bit cell based on an outcome of the sense amplifier.
- FIG. 6E illustrates another exemplary bridge circuit according to aspects of the present disclosure.
- a bridge circuit may include a pull up transistor M 1 , a first switch S 1 , and a pull down transistor M 2 .
- the bridge circuit may further include a driver capacitor C D , a second switch S 2 , and a local sense amplifier 612 .
- the pull up transistor M 1 , drive capacitor C D , switch S, pull down transistor M 2 , driver capacitor C D , a second switch S 2 , and a local sense amplifier 612 are connected in the manner as shown in FIG. 6E .
- the pull-up transistor M 1 can be configured to charge the memory bank signal line, and the pull up transistor M 1 can also be controlled to drive the memory block signal line.
- the local sense amplifier 612 can be configured to control the second switch S 2 and to control the pre-charge of the driver capacitor C D .
- the bridge circuit can be configured to pre-charge the driver capacitor C D as well as to pre-charge a pair of memory bank sign lines to a logic high, where the pair of memory bank sign lines includes a bit line and an inverse of the bit line of a row of bit cells to be read.
- a word line can be activated to control the row of bit cells, where each bit cell is coupled to a pair of memory bank sign lines, which are connected to a pair memory block sign lines via the bridge circuit.
- charges stored in the driver capacitor C D and the pre-charged pair of memory bank sign lines are shared with the capacitors (as shown in the signal line model of FIG. 1A ) of the pair memory block sign lines via the bridge circuit.
- This sharing of charges causes a deviation in voltage or current at the input node(s) of a sense amplifier, such as sense amplifier 608 .
- the sense amplifier is configured to sense the voltage differential or current differential of the pair of memory block signal lines caused by the shared charges from the pre-charged pair of memory bank sign lines to determine a value stored in the bit cell based on an outcome of the sense amplifier.
- FIG. 6F illustrates an exemplary holding circuit according to aspects of the present disclosure.
- a holding circuit may include a NMOS transistor M 3 , a PMOS transistor M 4 , a first resistor R 1 , and a second resistor R 2 .
- the NMOS transistor M 3 , PMOS transistor M 4 , first resistor R 1 , and second resistor R 2 are connected in the manner as shown in FIG. 6F .
- the gate terminal of the NMOS transistor M 3 receives an input (labeled as Vin) from a memory block signal line and the drain terminal of the PMOS transistor M 4 produces an output (labelled as Vout) to an input of a sense amplifier.
- the disclosed SRAM designs of FIG. 6A-6F have a number of advantages over the conventional SRAM design of FIG. 1C .
- First, the disclosed SRAM design has eliminated the column multiplexors of the conventional design.
- Second, the sense amplifiers are shared among a larger set of memory banks as opposed to be shared among two memory banks. Both of the above improvements result in saving in area/size of the integrated circuit used in implementing the SRAM.
- the time delay before a sense amplifier can start sensing a high to low signal transition can be significantly reduced with the disclosed implementation, which can lead to significant improvement in performance of the SRAM.
- this improvement can result in power savings for the SRAM.
- FIG. 7A illustrates an exemplary method of reducing dynamic power in a SRAM according to aspects of the present disclosure.
- the method provides a plurality of memory blocks, where a memory block in the plurality of memory blocks includes a plurality memory banks, where a memory bank in the plurality of memory banks includes a plurality bit cells.
- the method provides a set of memory bank signal lines, where a bit cell in the memory bank is accessed through the set of memory bank signal lines.
- the method shares a set of memory block signal lines across the plurality of memory banks in the memory block.
- the method bridges, by a bridge circuit, between the set of memory bank signal lines and the set of memory block signal lines.
- the method shares a set of sense amplifiers among the plurality of memory banks in the memory block, where the set of sense amplifiers correspond to the set of memory block signal lines.
- the method controls, by a controller, an access of one or more bit cells in the plurality bit cells.
- FIG. 7B illustrates an exemplary implementation of bridging between a set of memory bank signal lines and a set of memory block signal lines according to aspects of the present disclosure.
- the method disconnects, by the bridge circuit, the set of memory bank signal lines and the set of memory block signal lines during a standby mode of the memory bank; and connects, by the bridge circuit, the set of memory bank signal lines and the set of memory block signal lines during an access mode of the memory bank.
- the set of memory bank signal lines includes a plurality pairs of differential bit lines.
- the methods performed in block 722 may further include the methods performed in block 724 .
- the method causes, by the bridge circuit, a deviation in voltage in the plurality pairs of differential bit lines; and detects, by the set of sense amplifiers, the deviation in voltage in the plurality pairs of differential bit lines, where the bridge circuit includes a set of pull up transistors configured to charge the set of memory bank signal lines and controlled to drive the set of memory block signal lines.
- the methods performed in block 724 may further include the methods performed in block 726 ; and the methods performed in block 726 may further include the methods performed in block 728 .
- the method detects, by a differential amplifier, a voltage differential from a pair of memory block signal lines.
- the method maintains, by a holding circuit, an input voltage at an input node of the sense amplifier.
- FIG. 7C illustrates an exemplary method of performing a read operation of a SRAM according to aspects of the present disclosure.
- the method pre-charges, by a bridge circuit, a pair of memory bank sign lines to a logic high, where the pair of memory bank sign lines includes a bit line and an inverse of the bit line of a bit cell to be read; upon completing pre-charging the pair of memory bank sign lines, activates, by a controller, a word line to control the bit cell, and connects, by the bridge circuit, the pair of memory bank sign lines to a pair memory block sign lines.
- the methods performed in block 732 may further include the methods performed in block 734 .
- the method shares, by the bridge circuit, charges stored in the pre-charged pair of memory bank sign lines with the pair memory block sign lines, senses, by the sense amplifier, the voltage differential of the pair of memory block signal lines caused by the shared charges from the pre-charged pair of memory bank sign lines, and determines, by the sense amplifier, a value stored in the bit cell based on a sensed outcome.
- FIG. 8A illustrates an exemplary implementation of controlling a bit cell in a SRAM according to aspects of the present disclosure.
- a SRAM may include a plurality of memory blocks.
- a memory block in the plurality of memory blocks may include a plurality memory banks.
- a memory bank in the plurality of memory banks may include a plurality bit cells.
- a bit cell 802 of a SRAM may be controlled by a controller 804 and a bias circuit 806 .
- a bias circuit 806 can be configured to couple to a circuit ground terminal of the bit cell 802 , where the bias circuit 806 can be configured to produce a bias voltage to the bit cell 802 .
- the controller 804 can be configured to control the bias circuit 806 to produce a first set of bias settings in an access mode; and control the bias circuit 806 with a second set of bias settings in a standby mode of the bit cell in the SRAM.
- FIG. 8B illustrates an exemplary implementation of generating a bias voltage to circuit ground of a bit cell in a SRAM according to aspects of the present disclosure.
- a bias circuit may include a first transistor M 1 and a second transistor M 2 , a reference voltage generator 808 .
- the first transistor M 1 is configured to pull down a bias voltage to the circuit ground of the bit cell during an access mode of the bit cell 802 .
- the second transistor M 2 is configured to pull up the bias voltage to a first reference voltage during a standby mode of the bit cell 802 .
- the reference voltage generator 808 is configured to generate the corresponding bias voltage during the access mode and during the standby mode of the bit cell 802 .
- the bias control is provided by controller 804 and the bias voltage is provided to the Vss (also referred to as circuit ground) line of the bit cell 802 .
- the first reference voltage can be programmable.
- FIG. 8C illustrates an exemplary implementation of controlling a word line of a bit cell in a SRAM according to aspects of the present disclosure.
- the word line control circuit may include a first transistor M 1 and a second transistor M 2 .
- the controller may be configured to control the word line by asserting the word line (pull up though M 2 ) to select the bit cell during an access mode, and de-assert the word line (pull down through M 1 ) to de-select the bit cell during a standby mode.
- FIG. 8D illustrates an exemplary plot of drain to source voltage of a NMOS transistor in a bit cell of a SRAM according to aspects of the present disclosure.
- a voltage differential across a bit cell may be controlled by setting a bias voltage V BIAS as described above in FIG. 8A-8C .
- the bias voltage is set to Vss.
- the voltage differential across a bit cell is Vdd ⁇ Vss.
- the bias voltage is set to V BIAS .
- the voltage differential across a bit cell is Vdd ⁇ V BIAS , which can be a desirable retention voltage V RETENTION for the bit cell when it is not in operation.
- the voltage differential across a bit cell can be switched back to Vdd ⁇ Vss for a subsequent access mode.
- the switching in and out of the standby mode can be done quickly with the controller 804 and bias circuit 806 .
- the performance of the system is improved.
- the conventional method described in FIG. 1F is done when the entire SRAM is not in operation, for example during a period when a user device is in sleep mode.
- the disclosed method of FIG. 8A-8D can be performed when certain particular section of the SRAM memory is not being accessed, or when a particular row of bit cells are not being accessed. As a result, the static power consumption of the SRAM memory is reduced.
- FIG. 9A illustrates an exemplary method of reducing static power in a SRAM according to aspects of the present disclosure.
- the method provides a plurality of memory blocks, where a memory block in the plurality of memory blocks includes a plurality memory banks, where a memory bank in the plurality of memory banks includes a plurality bit cells.
- the method produces, by a bias circuit, a bias voltage to a row of bit cells, where the bias circuit is coupled to a circuit ground terminal of the row of bit cells in the plurality of bit cells.
- the method controls, by a controller, the bias circuit to produce a first set of bias settings in an access mode and control the bias circuit to produce a second set of bias settings in a standby mode of the SRAM.
- the method of producing the bias voltage by the bias circuit includes pulling down, by a first transistor, the bias voltage to circuit ground during an access mode of the row of bit cells, and pulling up, by a second transistor, the bias voltage to a first reference voltage during a standby mode of the row of bit cells.
- the first reference voltage can be programmable.
- FIG. 9B illustrates an exemplary method of controlling a bias circuit according to aspects of the present disclosure.
- the method controls the bias circuit to pull down the bias voltage to circuit ground during the access mode of the row of bit cells, and controls the bias circuit to pull up the bias voltage to a first reference voltage during the standby mode of the row of bit cells.
- the methods performed in block 912 may further include the methods performed in blocks 914 , 916 , 918 , and 920 .
- the method controls a word line to select the row of bit cells during the access mode, and controls the word line to de-select the row of bit cells during the standby mode.
- the method controls the bias circuit to set a positive voltage differential between a source terminal and a substrate of one or more NMOS transistors in the row of bit cells to induce a body effect, where the body effect reduces leakage current in the one or more NMOS transistors.
- the method controls the bias circuit to reduce a voltage differential between a drain terminal and a source terminal of one or more NMOS transistors in the row of bit cells, where a reduction of the voltage differential between a drain terminal and a source terminal reduces leakage current in the one or more NMOS transistors.
- the method controls the bias circuit to set a negative voltage differential between a gate terminal and a source terminal of one or more NMOS transistors in the row of bit cells, where the negative voltage differential between a gate terminal and a source terminal reduces leakage current in the one or more NMOS transistors.
Landscapes
- Engineering & Computer Science (AREA)
- Microelectronics & Electronic Packaging (AREA)
- Computer Hardware Design (AREA)
- Power Engineering (AREA)
- Static Random-Access Memory (AREA)
Abstract
Description
- The present invention relates to the field of integrated circuits. In particular, the present invention relates to a circuit for reducing static power in SRAM and methods for using the same.
-
FIG. 1A illustrates a modeling of a signal line between a transmitter and a receiver. As shown inFIG. 1A , atransmitter 102 may be configured to transmit a signal through a signal line to areceiver 104. Thetransmitter 102 drives a transmitted signal atnode 103 and thereceiver 104 receives a received signal atnode 105. The signal line may be modelled as a series of resistor-capacitor networks, labelled as 106 a through 106 j. The propagation delay through the signal line depends on electrical characteristics of the signal line, which is a function of the series of resistor-capacitor networks. As the signal line becomes longer, the propagation delay of the transmitter signal becomes longer, which adversely affects the performance of a system that uses the signal line. One way to overcome the above issue is to employ a stronger driver at the transmitter, thus driving more current onto the signal line. However, this solution consumes more power. -
FIG. 1B illustrates a conventional method of signal communication among a transmitter and multiple receivers in a system. In the conventional method ofFIG. 1B , another way to overcome the problem ofFIG. 1A is shown. In this example, atransmitter 112 may be configured to transmit a signal through a signal line to a plurality of receivers, namely 114 a through 114 j. Similar toFIG. 1A , the signal line may be modelled as a series of resistor-capacitor networks, labelled as 116 a through 116 m. To reduce the propagation delay between thetransmitter 112 and a receiver such as receiver 2 (labelled as 114 b), a plurality of buffers, such as 118 a through 118 j, may be used to reduce propagation delay and ensure sufficient signal strength received at a receiver. There are a number of drawbacks with this conventional method. First, the communication betweentransmitter 112 and a receiver such asreceiver J 114 j becomes one directional. In other words, another set of signal line would be used to transmit a signal fromreceiver J 114 j totransmitter 112. As the number of receiver increases in a system with a 32 bit or wider system bus, the number of signal paths and the components and areas used to support such signal paths increase exponentially, which can significantly increase the cost and complexity of the system. -
FIG. 1C illustrates a conventional design of a subsection of a SRAM. In the conventional design ofFIG. 1C , a subsection of a SRAM may include memory bank 0 (122 a) and memory bank 1 (122 b).Memory bank 0 may be accessed through column multiplexor 0 (124 a) and similarlymemory bank 1 may be accessed through column multiplexor 1 (124 b). A set ofsense amplifiers 126 are shared by thememory bank 0 andmemory bank 1 through thecolumn multiplexor 0 andcolumn multiplexor 1, respectively. The set ofsense amplifiers 126 drive a set of global bit lines corresponding to local bit lines ofmemory bank 0 andmemory bank 1. This subsection of SRAM can be part of a SRAM memory block, where multiple SRAM memory blocks may form the memory core of the SRAM. -
FIG. 1D illustrates a conventional design of a SRAM memory bank ofFIG. 1C . As shown inFIG. 1D , a memory bank may include multiple rows and columns of bit cells (BCs), represented byBC 128. In this example, a 32-bit memory bank width is shown. In other implementations, different memory bank widths may be designed, such as memory bank width of 8, 16, 64, 128, 256 bits may be implemented. In this example, each row of the memory bank is controlled by a word line, for example WL0 controls the top row of the memory bank. Each column of the memory bank is controlled by a pair of differential signal lines, for example Bit0 and its inverse _Bit0 control the left column, and Bit31 and its inverse _Bit31 control the left column. - To increase the storage and to reduce the ratio of overhead area (for sense amplifiers and column multiplexors) to bit cells area, a typical conventional design would pack many rows of bit cells into a memory bank. To reduce power consumption, each bit cell in the typical conventional design would have a weak drive strength. These design criteria impose adverse impact on the performance and power consumption of the SRAM.
-
FIG. 1E illustrates a conventional design of a bit cell of a SRAM. In the conventional design ofFIG. 1E , a bit cell may include transistors M1, M2, M3, M4, M4, M5, and M6; and the transistors M1, M2, M3, M4, M4, M5, and M6 are connected in the manner shown inFIG. 1E . The transistors are controlled by a word line (for example WLJ) and by a pair of differential bit lines (for example Bit K and _BitK). The drain terminals of transistors M5, and M6 are typically pulled up to Vdd, and the source terminals of transistors M3 and M4 are coupled to Vss. - There are multiple paths that can produce leakage current in the bit cell when the word line is asserted/activated, even though a transistor may be considered as logically off. For example, one leakage path is through M1 and M3; another leakage path is through M5 and M3; yet another leakage path is through M6 and M4; and yet another leakage path is through M2 and M4. Note that the leakage through the NMOS transistors are larger compare to the leakage through the PMOS transistors, because the size of the NMOS transistors is larger than the size of the PMOS transistors.
- When the bit cell is in standby mode, one of the PMOS transistor M5 or M6 may still be on, which creates at least one leakage path. In other words, even during the standby mode, most of the bit cells may have at least one leakage path. To prevent leakage in the standby mode,
FIG. 1F illustrates a conventional method of static power reduction in a SRAM. As shown in the conventional method ofFIG. 1F , during the access mode, the drain terminals of transistors M5, and M6 are typically pulled up to Vdd, and the source terminals of transistors M3 and M4 are coupled to Vss. During the standby mode, the drain terminals of transistors M5, and M6 are typically pulled up to a retention voltage VRETENTION, which is lower than Vdd, and the source terminals of transistors M3 and M4 remain at Vss. In this manner, power leakage can be reduced. However, one drawback with the above conventional method is that there is a long delay to transition from the standby mode back to the access mode, because the adjustment of power supply is done to the entire SRAM or may be done to the entire circuit that includes the SRAM. The long delay makes frequent switching between the access mode and standby mode impractical when a device is being used. - Therefore, it is desirable to have a circuit for reducing static power in SRAM and methods for using the same in order to improve the performance and to reduce power consumption of the above conventional methods.
- A circuit for reducing static power in SRAM and methods for using the same are disclosed. In one embodiment, a circuit for reducing static power in SRAM includes a plurality of memory blocks, where a memory block in the plurality of memory blocks includes a plurality memory banks, where a memory bank in the plurality of memory banks includes a plurality bit cells. The circuit further includes a bias circuit configured to produce a bias voltage to a row of bit cells, where the bias circuit is coupled to a circuit ground terminal of the row of bit cells in the plurality of bit cells, and a controller configured to control the bias circuit to produce a first set of bias settings in an access mode and control the bias circuit to produce a second set of bias settings in a standby mode of the SRAM.
- In another embodiment, a method for reducing static power in SRAM includes providing a plurality of memory blocks, where a memory block in the plurality of memory blocks includes a plurality memory banks, where a memory bank in the plurality of memory banks includes a plurality bit cells; producing, by a bias circuit, a bias voltage to a row of bit cells, where the bias circuit is coupled to a circuit ground terminal of the row of bit cells in the plurality of bit cells; and controlling, by a controller, the bias circuit to produce a first set of bias settings in an access mode and control the bias circuit to produce a second set of bias settings in a standby mode of the SRAM.
- The aforementioned features and advantages of the disclosure, as well as additional features and advantages thereof, will be more clearly understandable after reading detailed descriptions of embodiments of the disclosure in conjunction with the non-limiting and non-exhaustive aspects of following drawings. Like numbers are used throughout the specification.
-
FIG. 1A illustrates a modeling of a signal line between a transmitter and a receiver. -
FIG. 1B illustrates a conventional method of signal communication among a transmitter and multiple receivers in a system. -
FIG. 1C illustrates a conventional design of a SRAM data signals. -
FIG. 1D illustrates a conventional design of a SRAM memory bank ofFIG. 1C . -
FIG. 1E illustrates a conventional design of a bit cell of a SRAM. -
FIG. 1F illustrates a conventional method of static power reduction in a SRAM. -
FIG. 2A illustrates an exemplary representation of signal communication between two transceivers of a system according to aspects of the present disclosure. -
FIG. 2B illustrates an exemplary representation of signal communication among multiple transceivers in a system according to aspects of the present disclosure. -
FIG. 3A illustrates an exemplary implementation of a transmitter ofFIG. 2A according to aspects of the present disclosure.FIG. 3B illustrates an exemplary implementation of a transmitter ofFIG. 2A according to aspects of the present disclosure. -
FIG. 3C illustrates another exemplary implementation of a transmitter ofFIG. 2A according to aspects of the present disclosure.FIG. 3D illustrates another exemplary implementation of a transmitter ofFIG. 2A according to aspects of the present disclosure. -
FIG. 3E illustrates an exemplary implementation of a receiver ofFIG. 2A according to aspects of the present disclosure. -
FIG. 3F illustrates another exemplary implementation of a receiver ofFIG. 2A according to aspects of the present disclosure. -
FIG. 3G illustrates yet another exemplary implementation of a receiver ofFIG. 2A according to aspects of the present disclosure. -
FIG. 4A illustrates exemplary node voltages at an output node of a transmitter and an input node of a receiver based on a conventional implementation of low to high signal transition. -
FIG. 4B illustrates exemplary node voltages at an output node of a transmitter and an input node of a receiver based on the implementation ofFIG. 2A for a low to high signal transition according to aspects of the present disclosure. -
FIG. 4C illustrates another exemplary node voltages at an output node of a transmitter and an input node of a receiver based on a conventional implementation of high to low signal transition. -
FIG. 4D illustrates another exemplary node voltages at an output node of a transmitter and an input node of a receiver based on the implementation ofFIG. 2A for a high to low signal transition according to aspects of the present disclosure. -
FIG. 4E illustrates yet another exemplary node voltages at an output node of a transmitter and an input node of a receiver based on the implementation ofFIG. 3C for a low to high signal transition. -
FIG. 4F illustrates yet another exemplary node voltages at an output node of a transmitter and an input node of a receiver based on the implementation ofFIG. 3C for a high to low signal transition. -
FIG. 4G illustrates an exemplary implementation of a data bus and its control signals according to aspects of the present disclosure. -
FIG. 5A illustrates an exemplary method of signal communication between a transmitter and a receiver according to aspects of the present disclosure. -
FIG. 5B illustrates another exemplary method of signal communication between a transmitter and a receiver according to aspects of the present disclosure. -
FIG. 5C illustrates an exemplary method of detecting a transmitted signal according to aspects of the present disclosure. -
FIG. 5D illustrates yet another exemplary method of signal communication between a transmitter and a receiver according to aspects of the present disclosure. -
FIG. 6A illustrates an exemplary implementation of a SRAM according to aspects of the present disclosure. -
FIG. 6B illustrates an exemplary implementation of SRAM signal lines according to aspects of the present disclosure. -
FIG. 6C illustrates another exemplary implementation of SRAM signal lines according to aspects of the present disclosure. -
FIG. 6D illustrates an exemplary bridge circuit according to aspects of the present disclosure. -
FIG. 6E illustrates another exemplary bridge circuit according to aspects of the present disclosure. -
FIG. 6F illustrates an exemplary holding circuit according to aspects of the present disclosure. -
FIG. 7A illustrates an exemplary method of reducing dynamic power in a SRAM according to aspects of the present disclosure. -
FIG. 7B illustrates an exemplary implementation of bridging between a set of memory bank signal lines and a set of memory block signal lines according to aspects of the present disclosure. -
FIG. 7C illustrates an exemplary method of performing a read operation of a SRAM according to aspects of the present disclosure. -
FIG. 8A illustrates an exemplary implementation of controlling a bit cell in a SRAM according to aspects of the present disclosure. -
FIG. 8B illustrates an exemplary implementation of generating a bias voltage to circuit ground of a bit cell in a SRAM according to aspects of the present disclosure. -
FIG. 8C illustrates an exemplary implementation of controlling a word line of a bit cell in a SRAM according to aspects of the present disclosure. -
FIG. 8D illustrates an exemplary plot of drain to source voltage of a NMOS transistor in a bit cell of a SRAM according to aspects of the present disclosure. -
FIG. 9A illustrates an exemplary method of reducing static power in a SRAM according to aspects of the present disclosure. -
FIG. 9B illustrates an exemplary method of controlling a bias circuit according to aspects of the present disclosure. - The following descriptions are presented to enable a person skilled in the art to make and use the disclosure. Descriptions of specific embodiments and applications are provided only as examples. Various modifications and combinations of the examples described herein will be readily apparent to those skilled in the art, and the general principles defined herein may be applied to other examples and applications without departing from the scope of the disclosure. Thus, the present disclosure is not intended to be limited to the examples described and shown, but is to be accorded the scope consistent with the principles and features disclosed herein. The word “exemplary” or “example” is used herein to mean “serving as an example, instance, or illustration.” Any aspect or embodiment described herein as “exemplary” or as an “example” in not necessarily to be construed as preferred or advantageous over other aspects or embodiments.
-
FIG. 2A illustrates an exemplary representation of signal communication between two transceivers of a system according to aspects of the present disclosure. As shown inFIG. 2A , afirst transceiver 202 may be configured to transmit a signal through asignal line 206 to asecond transceiver 204. Thefirst transceiver 202 may include afirst transmitter 203 and afirst receiver 205. Thesecond transceiver 204 may include asecond transmitter 207 and asecond receiver 209. The signal line may be modelled as a series of resistor-capacitor networks, and the propagation delay through the signal line depends on electrical characteristics of the signal line, which is a function of the series of resistor-capacitor networks. In this exemplary implementation, the communication between transceiver 1 (202) and transceiver 2 (204) is bi-directional. For example,transmitter 203 of transceiver 1 (202) may be configured to send a transmitted signal through thesignal line 206 and be received as a received signal atreceiver 209 of transceiver 2 (204). Similarly,transmitter 207 of transceiver 2 (204) may be configured to send a transmitted signal through thesignal line 206 and be received as a received signal atreceiver 205 of transceiver 1 (202). -
FIG. 2B illustrates an exemplary representation of signal communication among multiple transceivers in a system according to aspects of the present disclosure. In the example ofFIG. 2B , afirst transceiver TR 1 212 may be configured to communication with a plurality of transceivers, labelled asTR 2 214 a,TR 3 214 b, throughTR 1 214 j, etc. Similar toFIG. 2A , the communication betweenTR 1 212 and the plurality of transceivers (TR 2 214 a,TR 3 214 b, throughTR J 214 j) are bi-directional. -
FIG. 3A illustrates an exemplary implementation of a transmitter ofFIG. 2A according to aspects of the present disclosure. In the exemplary implementation ofFIG. 3A , a transmitter may include a driver capacitor CDRIVE, and a first transistors M1 configured to discharge the driver capacitor CDRIVE. The transmitter may further include a second transistor M2 configured to charge anode 303 at the transmitter end of a signal line such assignal line 206. The transmitter may further include a switch S1 configured to connect or separate the driver capacitor CDRIVE and the signal line during different phases of an operation. The driver capacitor CDRIVE, the first transistor M1, the second transistor M2, and the switch S1 are connected in the manner shown inFIG. 3A . - During operation, to drive a high to low signal transition onto the signal line, in phase one with switch S1 open, the driver capacitor CDRIVE is discharged to logic low through the control of the M1 transistor while
node 303 is charged to logic high through the control of M2 transistor. Then in phase two, switch S1 is closed, causing charge sharing among the driver capacitor CDRIVE and the capacitors of the signal line, resulting in a controlled high to low voltage transition at the receiver end of the receiving transceiver. The controlled high to low voltage transition may have a magnitude that is proportional to a ratio of capacitance between the driver capacitor CDRIVE and the capacitors of the signal line. In addition, the controlled low to high voltage transition can be captured and sensed by a sense amplifier to determine the high to low signal transition by the receiver at the receiving transceiver. The transistors M1, M2, and the switch S1 are controlled in the manner as described above for a high to low signal transition. -
FIG. 3B illustrates an exemplary implementation of a transmitter ofFIG. 2A according to aspects of the present disclosure. In the exemplary implementation ofFIG. 3B , a transmitter may include a driver capacitor CDRIVE, and a first transistors M1 configured to charge the driver capacitor CDRIVE. The transmitter may further include a second transistor M2 configured to discharge anode 313 at the transmitter end of a signal line such assignal line 206. The transmitter may further include a switch S1 configured to connect or separate the driver capacitor CDRIVE and the signal line during different phases of an operation. The driver capacitor CDRIVE, the first transistor M1, the second transistor M2, and the switch S1 are connected in the manner shown inFIG. 3B . - During operation, to drive a low to high signal transition onto the signal line, in phase one with switch S1 open, the driver capacitor CDRIVE is charged to logic high through the control of the M1 transistor while
node 313 is discharged to logic low through the control of M2 transistor. Then in phase two, switch S1 is closed, causing charge sharing among the driver capacitor CDRIVE and the capacitors of the signal line, resulting in a controlled low to high voltage transition at the receiver end of the receiving transceiver. The controlled low to high voltage transition may have a magnitude that is proportional to a ratio of capacitance between the driver capacitor CDRIVE and the capacitors of the signal line. In addition, the controlled low to high voltage transition can be captured and sensed by a sense amplifier to determine the low to high signal transition by the receiver at the receiving transceiver. The transistors M1, M2, and the switch S1 are controlled in the manner as described above for a low to high signal transition. -
FIG. 3C illustrates another exemplary implementation of a transmitter ofFIG. 2A according to aspects of the present disclosure. In the example shown inFIG. 3C , a transmitter may include a driver capacitor CDRIVE, and a first transistor M1 configured to discharge the driver capacitor CDRIVE. The transmitter may further include a second transistor M2 configured to charge anode 323 at the transmitter end of a signal line such assignal line 216. The transmitter may further include a switch S1 configured to connect or separate the driver capacitor and the signal line during different phases of an operation. The transmitter further includes apulse generator circuit 322 configured to generate control signals to control transistors M1, M2, and the switch S1. The driver capacitor CDRIVE, the first transistor M1, the second transistor M2, the switch S1, and thepulse generator circuit 322 are connected in the manner shown inFIG. 3C . - During operation, to drive a high to low signal transition onto the signal line, in phase one with switch S1 open, the driver capacitor CDRIVE is discharged to logic low through the control of the M1 transistor while
node 323 is charged to logic high through the control of M2 transistor. Then in phase two, switch S1 is closed, causing charge sharing among the driver capacitor CDRIVE and the capacitors of the signal line, resulting in a controlled high to low voltage transition at the receiver end of the receiving transceiver. The controlled high to low voltage transition may have a magnitude that is proportional to a ratio of capacitance between the driver capacitor CDRIVE and the capacitors of the signal line. In addition, the controlled low to high voltage transition can be captured and sensed by a sense amplifier to determine the high to low signal transition by the receiver at the receiving transceiver. Thepulse generator circuit 322 generates control signals configured to control transistors M1, M2, and the switch S1, in the manner as described above for a high to low signal transition. Thepulse generator circuit 322 is configured to generate a pulse onto the signal line based on the signal to be transmitted; and where a duration of the pulse is programmable. -
FIG. 3D illustrates another exemplary implementation of a transmitter ofFIG. 2A according to aspects of the present disclosure. In the example shown inFIG. 3D , a transmitter may include a driver capacitor CDRIVE, and a first transistor M1 configured to charge the driver capacitor CDRIVE. The transmitter may further include a second transistor M2 configured to discharge anode 325 at the transmitter end of a signal line such assignal line 216. The transmitter may further include a switch S1 configured to connect or separate the driver capacitor CDRIVE and the signal line during different phases of an operation. The transmitter further includes apulse generator circuit 324 configured to generate control signals to control transistors M1, M2, and the switch S1. The driver capacitor CDRIVE, the first transistor M1, the second transistor M2, the switch S1, and thepulse generator circuit 324 are connected in the manner shown inFIG. 3D . - During operation, to drive a low to high signal transition onto the signal line, in phase one with switch S1 open, the driver capacitor CDRIVE is charged to logic high through the control of the M1 transistor while
node 325 is discharged to logic low through the control of M2 transistor. Then in phase two, switch S1 is closed, causing charge sharing among the driver capacitor CDRIVE and the capacitors of the signal line, resulting in a controlled low to high voltage transition at the receiver end of the receiving transceiver. The controlled low to high voltage transition may have a magnitude that is proportional to a ratio of capacitance between the driver capacitor CDRIVE and the capacitors of the signal line. In addition, the controlled low to high voltage transition can be captured and sensed by a sense amplifier to determine the low to high signal transition by the receiver at the receiving transceiver. Thepulse generator circuit 324 generates control signals configured to control transistors M1, M2, and the switch S1, in the manner as described above for a low to high signal transition. Thepulse generator circuit 324 is configured to generate a pulse onto the signal line based on the signal to be transmitted; and where a duration of the pulse is programmable. -
FIG. 3E illustrates an exemplary implementation of a receiver ofFIG. 2A according to aspects of the present disclosure. As shown inFIG. 3E , a receiver may include transistors M1, M2, and M7, a pair of cross-coupled inverters implemented by M3, M4, M5 and M6, a latch implemented by a pair of cross-coupled Nand gates N1 and N2. The transistors M1, M2, M3, M4, M5, M6, M7, and the pair of cross-coupled Nand gates N1 and N2 are connected in the manner shown inFIG. 3E . The receiver receives an input signal (In) at transistor M2, and a control signal (En_In) configured to control transistors M1, M2, and M7. The receiver is configured to generate an output signal (Out). In some implementations, transistors M1, M2 may be replaced with switches. -
FIG. 3F illustrates another exemplary implementation of a receiver ofFIG. 2A according to aspects of the present disclosure. In the exemplary implementation ofFIG. 3F , a receiver may include transistors M1, M2, and M7, a pair of cross-coupled inverters implemented by M3, M4, M5 and M6, a latch implemented by a pair of cross-coupled Nand gates N1 and N2. The transistors M1, M2, M3, M4, M5, M6, M7, and the pair of cross-coupled Nand gates N1 and N2 are connected in the manner shown inFIG. 3F . The receiver receives an input signal (In) at transistor M1, and a control signal (En_In) configured to control transistors M1, M2, and M7. The receiver is configured to generate an output signal (Out). In some implementations, transistors M1, M2 may be replaced with switches. -
FIG. 3G illustrates yet another exemplary implementation of a receiver ofFIG. 2A according to aspects of the present disclosure. In the exampleFIG. 3G , the receiver includes transistors M1, M2, M3, and a receive capacitor CR. The transistors M1, M2, M3, receive capacitor CR, an input signal (In), and an output signal Out are connected in the manner shown inFIG. 3G . According to aspects of the present disclosure, a transmitter circuit, such as the transmitter circuit ofFIG. 3A or the transmitter circuit ofFIG. 3C , is configured to cause a deviation in current at a receiver node of the signal line. A receiver circuit, such as the receiver circuit ofFIG. 3E or the receiver circuit ofFIG. 3F , is configured to reproduce the transmitted signal using the deviation in current at the receiver node of the signal line. - In this exemplary implementation, the receiver circuit includes a current mirror, formed by M2 and M3, configured to multiply the deviation in current detected at the receiver node (input In) of the signal line, the receive capacitor CR is configured to store charges produced by the current mirror, where the charges represent a signal strength of the transmitted signal.
-
FIG. 4A illustrates exemplary node voltages at an output node of a transmitter and an input node of a receiver based on a conventional implementation of low to high signal transition. As shown inFIG. 4A ,line 402 represents a node voltage at an output node of a transmitter.Line 404 represents a node voltage at an input node of a receiver. The voltage at the output node of the transmitter can change quickly. However, due to the series of resistor-capacitor networks described above and the loading of the signal line, the voltage at the input node of the receiver can rise slowly. There can be a significant delay before the change in voltage at the input node of the receiver can be recognized. This delay is represented by the time T1. -
FIG. 4B illustrates exemplary node voltages at an output node of a transmitter and an input node of a receiver based on the implementation ofFIG. 2A for a low to high signal transition according to aspects of the present disclosure. In the example ofFIG. 4B ,line 406 represents a node voltage at an output node of a transmitter.Line 408 represents a node voltage at an input node of a receiver. The voltage at the output node of the transmitter can change quickly. Since the transmitter circuit is configured to cause a deviation in voltage at a receiver node of the signal line and the receiver circuit is configured to detect the deviation in voltage from a reference voltage at the receiver node of the signal line, the magnitude of the voltage deviation, represented by Vs1 with respect to Vss, to be sensed at the receiver end is significantly reduced. Thus, a steady voltage of Vs1 at the input node of the receiver can be reached in significantly shorter time, represented by T2. Note that the lines are shown for illustration purposes. In practice, the behavior of the node voltages can be non-linear and can vary in different design environments. Note that the lines are shown for illustration purposes and they are not drawn to scale. In practice, the behavior of the node voltages can be non-linear and can vary in different design environments. - Comparing
FIG. 4B andFIG. 4A , the time delay (T2 vs T1) before the receiver can start sensing a low to high signal transition on the signal line can be significantly reduced with the disclosed implementation, which can lead to significant improvement in performance of the system. In addition, comparing the amount of voltage change at the input node of the receiver (line 408 vs line 404), the amount of voltage swing on the signal line can be significantly reduced with the disclosed implementation, and this reduction in voltage swing can result in power savings for the system. -
FIG. 4C illustrates another exemplary node voltages at an output node of a transmitter and an input node of a receiver based on a conventional implementation of high to low signal transition. In the example shown inFIG. 4C ,line 412 represents a node voltage at an output node of a transmitter.Line 414 represents a node voltage at an input node of a receiver. The voltage at the output node of the transmitter can change quickly. However, due to the series of resistor-capacitor networks described above and the loading of the signal line, the voltage at the input node of the receiver can decrease slowly. There can be a significant delay before the change in voltage at the input node of the receiver can be recognized. This delay is represented by the time T3. -
FIG. 4D illustrates another exemplary node voltages at an output node of a transmitter and an input node of a receiver based on the implementation ofFIG. 2A for a high to low signal transition according to aspects of the present disclosure. As shown inFIG. 4D ,line 416 represents a node voltage at an output node of a transmitter.Line 418 represents a node voltage at an input node of a receiver. The voltage at the output node of the transmitter can change quickly. Since the transmitter circuit is configured to cause a deviation in voltage at a receiver node of the signal line and the receiver circuit is configured to detect the deviation in voltage from a reference voltage at the receiver node of the signal line, the magnitude of the voltage deviation, represented by Vs2 with respect to Vdd, to be sensed at the receiver end is significantly reduced. Thus, a steady voltage of Vs2 at the input node of the receiver can be reached in significantly shorter time, represented by T4. Note that the lines are shown for illustration purposes and they are not drawn to scale. In practice, the behavior of the node voltages can be non-linear and can vary in different design environments. - Comparing
FIG. 4D andFIG. 4C , the time delay (T4 vs T3) before the receiver can start sensing a high to low signal transition on the signal line can be significantly reduced with the disclosed implementation, which can lead to significant improvement in performance of the system. In addition, comparing the amount of voltage change at the input node of the receiver (line 418 vs line 414), the amount of voltage swing on the signal line can be significantly reduced with the disclosed implementation, and this reduction in voltage swing can result in power savings for the system. -
FIG. 4E illustrates yet another exemplary node voltages at an output node of a transmitter and an input node of a receiver based on the implementation ofFIG. 3C for a low to high signal transition. In the example shown inFIG. 4E ,line 422 represents a node voltage at an output node of a transmitter based on the exemplary implementationFIG. 3C for a low to high signal transition.Line 424 represents a node voltage at an input node of a receiver based on the exemplary implementationFIG. 3C for a low to high signal transition. With thepulse generator circuit 322, the duration of the signal being driven onto the signal line by the transmitter circuit can be further reduced, which can result in further savings in power consumption. In this example, the receiver may be configured to sense the voltage deviation Vs1 with respect to Vss at time T5. The transmitter circuit can stop driving the signal line after a duration controlled by the pulse generator circuit, for example after time T6. At T7, the input node at the receiver may be returned to a standby state, which is Vss in this example. Besides the differences introduced by thepulse generator circuit 322, the behavior of the transmission mechanism is similar to that ofFIG. 4B , and the description is not repeated here. -
FIG. 4F illustrates yet another exemplary node voltages at an output node of a transmitter and an input node of a receiver based on the implementation ofFIG. 3C for a high to low signal transition. As shown inFIG. 4F ,line 426 represents a node voltage at an output node of a transmitter based on the exemplary implementationFIG. 3C for a low to high signal transition.Line 428 represents a node voltage at an input node of a receiver based on the exemplary implementationFIG. 3C for a low to high signal transition. With thepulse generator circuit 322, the duration of the signal being driven onto the signal line by the transmitter circuit can be further reduced, which can result in further savings in power consumption. In this example, the receiver may be configured to sense the voltage deviation Vs2 with respect to Vdd at time T8. The transmitter circuit can stop driving the signal line after a duration controlled by the pulse generator circuit, for example after time T9. At T10, the input node at the receiver may be returned to a standby state, which is Vdd in this example. Besides the differences introduced by thepulse generator circuit 322, the behavior of the transmission mechanism is similar to that ofFIG. 4D , and the description is not repeated here. -
FIG. 4G illustrates an exemplary implementation of a data bus and its control signals according to aspects of the present disclosure. In the exemplary implementation shown inFIG. 4G , transceiver 1 (432) is configured to communicate with transceiver 2 (434) using a signal bus.Transceiver 1 receives a 32-bit input data (Data_in) and a control signal Enable_In (En_In), and produces a 32-bit output data (Data_Out), along with control signals (En_Out) and an optional pre-charge timing reference (PT_Ref). In some implementations, each line of the 32-bit output Data_Out is transmitted using the mechanism and/or circuits described in association withFIG. 2A-2B ,FIG. 3A-3E , andFIG. 4B, 4D-4F . In some implementations, the control signals En_out and PT_Ref may be communicated totransceiver 2 using a conventional buffering mechanism. The PT_Ref signal is configured to assist the receiver about the timing and state of charging and/or discharging of the output node at the transmitter oftransceiver 1, for example in implementations described in association withFIG. 4E andFIG. 4F . -
FIG. 5A illustrates an exemplary method of signal communication between a transmitter and a receiver according to aspects of the present disclosure. As shown inFIG. 5A , inblock 502, the method drives a transmitted signal onto a signal line. Inblock 504, the method detects the transmitted signal based on a deviation of a received signal from a reference signal on the signal line. Inblock 506, the method uses the received signal to communicate the transmitted signal. In some implementations, the signal line is configured to transmit bi-directional signals for different periods of signal transmission. An amount of deviation of the signal from the reference signal to be detected by the receiver is programmable. -
FIG. 5B illustrates another exemplary method of signal communication between a transmitter and a receiver according to aspects of the present disclosure. In the exemplary implementation ofFIG. 5B , inblock 512, the method causes, by the transmitter circuit, a deviation in voltage at a receiver node of the signal line; and detects, by the receiver circuit, the deviation in voltage from a reference voltage at the receiver node of the signal line. Inblock 514, the transmitter circuit includes a capacitor configured to share charges with the signal line; a pull down transistor configured to discharge the capacitor via control of a first switch and a second switch, or a pull up transistor configured to charge the capacitor via control of the first switch and the second switch; where the pull down transistor or the pull up transistor is controlled to drive the signal line. - According to aspects of the present disclosure, the methods performed in
blocks blocks block 516, for a low to high signal transition on the signal line, the method pre-charges the capacitor and drives the signal line using the pre-charged capacitor. Inblock 518, for a high to low signal transition on the signal line, the method discharges the capacitor and drives the signal line using the discharged capacitor. Inblock 520, the method generates, by a pulse generator, a pulse onto the signal line based on the transmitted signal; where the pulse generator is coupled to the pull down transistor and the pull up transistor, and where a duration of the pulse is programmable. -
FIG. 5C illustrates an exemplary method of detecting a transmitted signal according to aspects of the present disclosure. As shown in the example ofFIG. 5C , inblock 522, the method detects, by a sense amplifier, the deviation of the received signal from the reference signal on the receiver node of the signal line. In some implementations, the method performed inblock 522 may additionally and/or optionally include the method performed inblock 524. Inblock 524, the method maintains, by a holding circuit, an input voltage at the receiver node of the signal line. -
FIG. 5D illustrates yet another exemplary method of signal communication between a transmitter and a receiver according to aspects of the present disclosure. In the example ofFIG. 5D , inblock 532, the method causes, by the transmitter circuit, a deviation in current at a receiver node of the signal line; and reproduces, by the receiver circuit, the transmitted signal using the deviation in current at the receiver node of the signal line. In some implementations, the method performed inblock 532 may additionally and/or optionally include the method performed inblock 534. Inblock 534, the method multiplies, by a current mirror, the deviation in current detected at the receiver node of the signal line; and stores, by a capacitor, charges produced by the current mirror, where the charges represent a signal strength of the transmitted signal. -
FIG. 6A illustrates an exemplary implementation of a SRAM according to aspects of the present disclosure. A SRAM may include a plurality of memory blocks. A memory block in the plurality of memory blocks may include a plurality memory banks, such asmemory banks 0 through 3, memory bank L−1, memory bank L, memory bank W−1, and memory bank W, labelled 602 a through 602 h, respectively. A memory bank in the plurality of memory banks includes a plurality bit cells, such as the bit cell shown inFIG. 1E . In this particular implementation, two adjacent memory banks may share a bridge circuit. For example,memory bank 0 andmemory bank 1 share bridge circuit 1 (labeled as 604 a). - In the exemplary implementation of
FIG. 6A , a block of a SRAM may include a set of memory bank signal lines, labelled as 603 a through 603 h, and a corresponding bridge circuit. A bit cell in a memory bank can be accessed through the set of memory bank signal lines. The block of a SRAM may further include a set of memoryblock signal lines 606 shared across the plurality of memory banks in the memory block through bridge circuits (such as 604 a, 604 b, 6041 and 604 w) couple between the set of memory bank signal lines and the set of memory block signal lines 606. As described in association withFIG. 1D , the set of memory bank signal lines includes a plurality of differential bit lines. The block of a SRAM may further include a set ofsense amplifiers 608 corresponding to the set of memory block signal lines and additionally/optionally include holding circuits, such as holdingcircuit 610, where the set of sense amplifiers are shared among the plurality of memory banks in the memory block. A controller (not shown) may be used to control an access of one or more bit cells in the plurality bit cells. The controller may include an address decoder configured to generate control signals to the bridge circuit and the plurality bit cells. -
FIG. 6B illustrates an exemplary implementation of SRAM signal lines according to aspects of the present disclosure. In the example ofFIG. 6B , a signal path ofFIG. 6A is shown, for example from memory bank 0 (602 a) via bank signal lines (603 a), bridge circuit (604 a), memoryblock signal lines 606, to senseamplifiers 608. According to aspects of the present disclosure, thebridge circuit 604 a can be configured to disconnect the set of memorybank signal lines 603 a and the set of memoryblock signal lines 606 during a standby mode of a memory bank, such asmemory bank 0 ormemory bank 1. Thebridge circuit 604 a can be further configured to connect the set of memorybank signal lines 603 a and the set of memoryblock signal lines 606 during an access mode of a memory bank. In addition, thebridge circuit 604 a can be configured to cause a deviation in voltage in the plurality pairs of differential bit lines in a memory bank. The set ofsense amplifiers 608 can be configured to detect the deviation in voltage in the plurality pairs of differential bit lines, as described above in association withFIG. 3A-3E andFIG. 4A-4F . -
FIG. 6C illustrates another exemplary implementation of SRAM signal lines according to aspects of the present disclosure. Similar to the example toFIG. 6B , a signal path ofFIG. 6A is shown, for example frommemory bank 0 viabank signal lines 603 a, bridge circuit (604 a), a firstoptional holding circuit 610 a, memoryblock signal lines 606, a second optional holdingcircuits 610 b, to senseamplifiers 608. A holding circuit, such as 610 a or 610 b, can be configured to maintain a voltage at an output node of thebridge circuit 604 a or at an input node of asense amplifier 608. The set ofsense amplifiers 608 can be configured to detect the deviation in voltage in the plurality pairs of differential bit lines, as described above in association withFIG. 3A-3E andFIG. 4A-4F . In addition, a sense amplifier may include a differential amplifier configured to detect a voltage differential from a pair of memory block signal lines. -
FIG. 6D illustrates an exemplary bridge circuit according to aspects of the present disclosure. In the example ofFIG. 6D , an implementation of handling one bit of a signal line in the bridge circuit is shown. A bridge circuit may include a pull up transistor M1, a switch S, and a pull down transistor M2. The pull up transistor M1, switch S, and pull down transistor M2 are connected in the manner as shown inFIG. 6D . The pull-up transistor M1 can be configured to charge the memory bank signal line, and the pull up transistor M1 can also be controlled to drive the memory block signal line. - In some implementations, for a read operation, the bridge circuit can be configured to pre-charge a pair of memory bank sign lines to a logic high, where the pair of memory bank sign lines includes a bit line and an inverse of the bit line of a row of bit cells to be read. Upon completing pre-charging the pair of memory bank sign lines, a word line can be activated to control the row of bit cells, where each bit cell is coupled to a pair of memory bank sign lines, which are connected to a pair memory block sign lines via the bridge circuit.
- With the above implementation, charges stored in the pre-charged pair of memory bank sign lines are shared with the capacitors (as shown in the signal line model of
FIG. 1A ) of the pair memory block sign lines via the bridge circuit. This sharing of charges causes a deviation in voltage or current at the input node(s) of a sense amplifier. The sense amplifier is configured to sense the voltage differential or current differential of the pair of memory block signal lines caused by the shared charges from the pre-charged pair of memory bank sign lines to determine a value stored in the bit cell based on an outcome of the sense amplifier. -
FIG. 6E illustrates another exemplary bridge circuit according to aspects of the present disclosure. In the example ofFIG. 6E , an implementation of handling one bit of a signal line in the bridge circuit is shown. A bridge circuit may include a pull up transistor M1, a first switch S1, and a pull down transistor M2. The bridge circuit may further include a driver capacitor CD, a second switch S2, and alocal sense amplifier 612. - The pull up transistor M1, drive capacitor CD, switch S, pull down transistor M2, driver capacitor CD, a second switch S2, and a
local sense amplifier 612 are connected in the manner as shown inFIG. 6E . The pull-up transistor M1 can be configured to charge the memory bank signal line, and the pull up transistor M1 can also be controlled to drive the memory block signal line. Thelocal sense amplifier 612 can be configured to control the second switch S2 and to control the pre-charge of the driver capacitor CD. - In some implementations, for a read operation, the bridge circuit can be configured to pre-charge the driver capacitor CD as well as to pre-charge a pair of memory bank sign lines to a logic high, where the pair of memory bank sign lines includes a bit line and an inverse of the bit line of a row of bit cells to be read. Upon completing pre-charging the driver capacitor CD and the pair of memory bank sign lines, a word line can be activated to control the row of bit cells, where each bit cell is coupled to a pair of memory bank sign lines, which are connected to a pair memory block sign lines via the bridge circuit.
- With the above implementation, charges stored in the driver capacitor CD and the pre-charged pair of memory bank sign lines are shared with the capacitors (as shown in the signal line model of
FIG. 1A ) of the pair memory block sign lines via the bridge circuit. This sharing of charges causes a deviation in voltage or current at the input node(s) of a sense amplifier, such assense amplifier 608. The sense amplifier is configured to sense the voltage differential or current differential of the pair of memory block signal lines caused by the shared charges from the pre-charged pair of memory bank sign lines to determine a value stored in the bit cell based on an outcome of the sense amplifier. -
FIG. 6F illustrates an exemplary holding circuit according to aspects of the present disclosure. In the example ofFIG. 6F , an implementation of handling one bit of the signal line in the holding circuit is shown. A holding circuit may include a NMOS transistor M3, a PMOS transistor M4, a first resistor R1, and a second resistor R2. The NMOS transistor M3, PMOS transistor M4, first resistor R1, and second resistor R2 are connected in the manner as shown inFIG. 6F . Applying to the example of holdingcircuit 610 b inFIG. 6C , the gate terminal of the NMOS transistor M3 receives an input (labeled as Vin) from a memory block signal line and the drain terminal of the PMOS transistor M4 produces an output (labelled as Vout) to an input of a sense amplifier. - The disclosed SRAM designs of
FIG. 6A-6F have a number of advantages over the conventional SRAM design ofFIG. 1C . First, the disclosed SRAM design has eliminated the column multiplexors of the conventional design. Second, the sense amplifiers are shared among a larger set of memory banks as opposed to be shared among two memory banks. Both of the above improvements result in saving in area/size of the integrated circuit used in implementing the SRAM. Moreover, through the mechanism of charge sharing, the time delay before a sense amplifier can start sensing a high to low signal transition can be significantly reduced with the disclosed implementation, which can lead to significant improvement in performance of the SRAM. Furthermore, due to the low range of voltage swing at the inputs of the sense amplifiers, this improvement can result in power savings for the SRAM. -
FIG. 7A illustrates an exemplary method of reducing dynamic power in a SRAM according to aspects of the present disclosure. In the example shown inFIG. 7A , inblock 702, the method provides a plurality of memory blocks, where a memory block in the plurality of memory blocks includes a plurality memory banks, where a memory bank in the plurality of memory banks includes a plurality bit cells. Inblock 704, the method provides a set of memory bank signal lines, where a bit cell in the memory bank is accessed through the set of memory bank signal lines. Inblock 706, the method shares a set of memory block signal lines across the plurality of memory banks in the memory block. Inblock 708, the method bridges, by a bridge circuit, between the set of memory bank signal lines and the set of memory block signal lines. Inblock 710, the method shares a set of sense amplifiers among the plurality of memory banks in the memory block, where the set of sense amplifiers correspond to the set of memory block signal lines. Inblock 712, the method controls, by a controller, an access of one or more bit cells in the plurality bit cells. -
FIG. 7B illustrates an exemplary implementation of bridging between a set of memory bank signal lines and a set of memory block signal lines according to aspects of the present disclosure. As shown in the exemplary implementation ofFIG. 7B , inblock 722, the method disconnects, by the bridge circuit, the set of memory bank signal lines and the set of memory block signal lines during a standby mode of the memory bank; and connects, by the bridge circuit, the set of memory bank signal lines and the set of memory block signal lines during an access mode of the memory bank. The set of memory bank signal lines includes a plurality pairs of differential bit lines. - In some implementations, the methods performed in
block 722 may further include the methods performed inblock 724. Inblock 724, the method causes, by the bridge circuit, a deviation in voltage in the plurality pairs of differential bit lines; and detects, by the set of sense amplifiers, the deviation in voltage in the plurality pairs of differential bit lines, where the bridge circuit includes a set of pull up transistors configured to charge the set of memory bank signal lines and controlled to drive the set of memory block signal lines. - In some implementations, the methods performed in
block 724 may further include the methods performed inblock 726; and the methods performed inblock 726 may further include the methods performed inblock 728. Inblock 726, for each sense amplifier in the set of sense amplifiers, the method detects, by a differential amplifier, a voltage differential from a pair of memory block signal lines. Inblock 728, the method maintains, by a holding circuit, an input voltage at an input node of the sense amplifier. -
FIG. 7C illustrates an exemplary method of performing a read operation of a SRAM according to aspects of the present disclosure. As shown inFIG. 7C , inblock 732, for a read operation, the method pre-charges, by a bridge circuit, a pair of memory bank sign lines to a logic high, where the pair of memory bank sign lines includes a bit line and an inverse of the bit line of a bit cell to be read; upon completing pre-charging the pair of memory bank sign lines, activates, by a controller, a word line to control the bit cell, and connects, by the bridge circuit, the pair of memory bank sign lines to a pair memory block sign lines. - In some implementations, the methods performed in
block 732 may further include the methods performed inblock 734. Inblock 734, the method shares, by the bridge circuit, charges stored in the pre-charged pair of memory bank sign lines with the pair memory block sign lines, senses, by the sense amplifier, the voltage differential of the pair of memory block signal lines caused by the shared charges from the pre-charged pair of memory bank sign lines, and determines, by the sense amplifier, a value stored in the bit cell based on a sensed outcome. -
FIG. 8A illustrates an exemplary implementation of controlling a bit cell in a SRAM according to aspects of the present disclosure. A SRAM may include a plurality of memory blocks. A memory block in the plurality of memory blocks may include a plurality memory banks. A memory bank in the plurality of memory banks may include a plurality bit cells. As shown in the exemplary implementation ofFIG. 8A , abit cell 802 of a SRAM may be controlled by acontroller 804 and abias circuit 806. Abias circuit 806 can be configured to couple to a circuit ground terminal of thebit cell 802, where thebias circuit 806 can be configured to produce a bias voltage to thebit cell 802. Thecontroller 804 can be configured to control thebias circuit 806 to produce a first set of bias settings in an access mode; and control thebias circuit 806 with a second set of bias settings in a standby mode of the bit cell in the SRAM. -
FIG. 8B illustrates an exemplary implementation of generating a bias voltage to circuit ground of a bit cell in a SRAM according to aspects of the present disclosure. In the example ofFIG. 8B , a bias circuit may include a first transistor M1 and a second transistor M2, areference voltage generator 808. The first transistor M1 is configured to pull down a bias voltage to the circuit ground of the bit cell during an access mode of thebit cell 802. The second transistor M2 is configured to pull up the bias voltage to a first reference voltage during a standby mode of thebit cell 802. Thereference voltage generator 808 is configured to generate the corresponding bias voltage during the access mode and during the standby mode of thebit cell 802. The bias control is provided bycontroller 804 and the bias voltage is provided to the Vss (also referred to as circuit ground) line of thebit cell 802. According aspects of the present disclosure, the first reference voltage can be programmable. -
FIG. 8C illustrates an exemplary implementation of controlling a word line of a bit cell in a SRAM according to aspects of the present disclosure. InFIG. 8C , the word line control circuit may include a first transistor M1 and a second transistor M2. The controller may be configured to control the word line by asserting the word line (pull up though M2) to select the bit cell during an access mode, and de-assert the word line (pull down through M1) to de-select the bit cell during a standby mode. -
FIG. 8D illustrates an exemplary plot of drain to source voltage of a NMOS transistor in a bit cell of a SRAM according to aspects of the present disclosure. In the example ofFIG. 8D , a voltage differential across a bit cell may be controlled by setting a bias voltage VBIAS as described above inFIG. 8A-8C . During an access mode of a bit cell, the bias voltage is set to Vss. As a result, the voltage differential across a bit cell is Vdd−Vss. During a standby mode, where the word line of a bit cell is not asserted (i.e. the bit cell is not selected), the bias voltage is set to VBIAS. As a result, the voltage differential across a bit cell is Vdd−VBIAS, which can be a desirable retention voltage VRETENTION for the bit cell when it is not in operation. The voltage differential across a bit cell can be switched back to Vdd−Vss for a subsequent access mode. - Comparing the disclosed method of
FIG. 8A-8D to the convention method described inFIG. 1F , the switching in and out of the standby mode can be done quickly with thecontroller 804 andbias circuit 806. Thus, the performance of the system is improved. Moreover, the conventional method described inFIG. 1F is done when the entire SRAM is not in operation, for example during a period when a user device is in sleep mode. On the other hand, the disclosed method ofFIG. 8A-8D can be performed when certain particular section of the SRAM memory is not being accessed, or when a particular row of bit cells are not being accessed. As a result, the static power consumption of the SRAM memory is reduced. Furthermore, by biasing a bit cell with VBIAS when it is not being accessed, it results in: 1) the drain to source voltage of one or more transistors of the bit cell has been reduced; 2) the body effect of the one or more transistors of the bit cell has been increased; and most importantly 3) the gate to source voltage of the one or more transistors of the bit cell has been reduced and may become negative. All the above effects can contribute to the reduction of static power consumption in a SRAM. -
FIG. 9A illustrates an exemplary method of reducing static power in a SRAM according to aspects of the present disclosure. In the exemplary implementation ofFIG. 9A , inblock 902, the method provides a plurality of memory blocks, where a memory block in the plurality of memory blocks includes a plurality memory banks, where a memory bank in the plurality of memory banks includes a plurality bit cells. Inblock 904, the method produces, by a bias circuit, a bias voltage to a row of bit cells, where the bias circuit is coupled to a circuit ground terminal of the row of bit cells in the plurality of bit cells. Inblock 906, the method controls, by a controller, the bias circuit to produce a first set of bias settings in an access mode and control the bias circuit to produce a second set of bias settings in a standby mode of the SRAM. - In some implementations, the method of producing the bias voltage by the bias circuit includes pulling down, by a first transistor, the bias voltage to circuit ground during an access mode of the row of bit cells, and pulling up, by a second transistor, the bias voltage to a first reference voltage during a standby mode of the row of bit cells. The first reference voltage can be programmable.
-
FIG. 9B illustrates an exemplary method of controlling a bias circuit according to aspects of the present disclosure. As shown in the example ofFIG. 9B , inblock 912, the method controls the bias circuit to pull down the bias voltage to circuit ground during the access mode of the row of bit cells, and controls the bias circuit to pull up the bias voltage to a first reference voltage during the standby mode of the row of bit cells. In some implementations, the methods performed inblock 912 may further include the methods performed inblocks - In
block 914, the method controls a word line to select the row of bit cells during the access mode, and controls the word line to de-select the row of bit cells during the standby mode. Inblock 916, during the standby mode, the method controls the bias circuit to set a positive voltage differential between a source terminal and a substrate of one or more NMOS transistors in the row of bit cells to induce a body effect, where the body effect reduces leakage current in the one or more NMOS transistors. Inblock 918, during the standby mode, the method controls the bias circuit to reduce a voltage differential between a drain terminal and a source terminal of one or more NMOS transistors in the row of bit cells, where a reduction of the voltage differential between a drain terminal and a source terminal reduces leakage current in the one or more NMOS transistors. Inblock 920, during the standby mode, the method controls the bias circuit to set a negative voltage differential between a gate terminal and a source terminal of one or more NMOS transistors in the row of bit cells, where the negative voltage differential between a gate terminal and a source terminal reduces leakage current in the one or more NMOS transistors. - One skilled in the relevant art will recognize that many possible modifications and combinations of the disclosed embodiments may be used, while still employing the same basic underlying mechanisms and methodologies. The foregoing description, for purposes of explanation, has been written with references to specific embodiments. However, the illustrative discussions above are not intended to be exhaustive or to limit the invention to the precise forms disclosed. Many modifications and variations are possible in view of the above teachings. The embodiments were chosen and described to explain the principles of the invention and their practical applications, and to enable others skilled in the art to best utilize the invention and various embodiments with various modifications as suited to the particular use contemplated.
Claims (16)
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US16/399,942 US10832762B1 (en) | 2019-04-30 | 2019-04-30 | Static power reduction in SRAM |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US16/399,942 US10832762B1 (en) | 2019-04-30 | 2019-04-30 | Static power reduction in SRAM |
Publications (2)
Publication Number | Publication Date |
---|---|
US20200350002A1 true US20200350002A1 (en) | 2020-11-05 |
US10832762B1 US10832762B1 (en) | 2020-11-10 |
Family
ID=73016094
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
US16/399,942 Active US10832762B1 (en) | 2019-04-30 | 2019-04-30 | Static power reduction in SRAM |
Country Status (1)
Country | Link |
---|---|
US (1) | US10832762B1 (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20230317126A1 (en) * | 2022-03-30 | 2023-10-05 | Arm Limited | Circuits and Methods of Detecting at Least Partial Breakdown of Canary Circuits |
Family Cites Families (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US8675439B2 (en) * | 2011-10-12 | 2014-03-18 | Taiwan Semiconductor Manufacturing Co., Ltd. | Bit line voltage bias for low power memory design |
-
2019
- 2019-04-30 US US16/399,942 patent/US10832762B1/en active Active
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20230317126A1 (en) * | 2022-03-30 | 2023-10-05 | Arm Limited | Circuits and Methods of Detecting at Least Partial Breakdown of Canary Circuits |
Also Published As
Publication number | Publication date |
---|---|
US10832762B1 (en) | 2020-11-10 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US7450448B2 (en) | Semiconductor memory device | |
US5973972A (en) | Precharge system for a semiconductor memory device | |
JP3416062B2 (en) | Content addressable memory (CAM) | |
CA2479682A1 (en) | Low-power high-performance memory cell and related methods | |
US8779800B2 (en) | Control signal generation circuit and sense amplifier circuit using the same | |
US20210134358A1 (en) | Sram with burst mode operation | |
WO2019070355A1 (en) | Area efficient write data path circuit for sram yield enhancement | |
US7307901B2 (en) | Apparatus and method for improving dynamic refresh in a memory device | |
US7046565B1 (en) | Bi-mode sense amplifier with dual utilization of the reference cells and dual precharge scheme for improving data retention | |
US7738305B2 (en) | Read-out circuit for or in a ROM memory; ROM memory and method for reading the ROM memory | |
US20070140027A1 (en) | Precharge circuit and method employing inactive weak precharging and equalizing scheme and memory device including the same | |
US20160125934A1 (en) | Method of controlling auxiliary branches for sram cell | |
US6975532B1 (en) | Quasi-static random access memory | |
US6137732A (en) | Semiconductor memory device having voltage boosting circuit | |
KR101317874B1 (en) | Nonvolatile memory device and readout method therefor | |
KR100318321B1 (en) | A controll circuit for a bit line equalization signal in semiconductor memory | |
JP3753972B2 (en) | Semiconductor memory device | |
US10832762B1 (en) | Static power reduction in SRAM | |
US20200335151A1 (en) | Low-power memory | |
US10916298B2 (en) | Dynamic power reduction in SRAM | |
US10896723B2 (en) | Signal communication circuit implementing receiver and transmitter circuits | |
JP2003016785A (en) | Semiconductor memory and information device using it | |
US6898136B2 (en) | Semiconductor memory device, capable of reducing power consumption | |
US7054210B2 (en) | Write/precharge flag signal generation circuit and circuit for driving bit line isolation circuit in sense amplifier using the same | |
US6212120B1 (en) | Semiconductor memory device with less power consumption |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
FEPP | Fee payment procedure |
Free format text: ENTITY STATUS SET TO UNDISCOUNTED (ORIGINAL EVENT CODE: BIG.); ENTITY STATUS OF PATENT OWNER: SMALL ENTITY |
|
FEPP | Fee payment procedure |
Free format text: ENTITY STATUS SET TO SMALL (ORIGINAL EVENT CODE: SMAL); ENTITY STATUS OF PATENT OWNER: SMALL ENTITY |
|
AS | Assignment |
Owner name: AMBIENT SCIENTIFIC INC., CALIFORNIA Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNOR:SINGH, GAJENDRA PRASAD;REEL/FRAME:053850/0252 Effective date: 20200922 |
|
STCF | Information on status: patent grant |
Free format text: PATENTED CASE |
|
MAFP | Maintenance fee payment |
Free format text: PAYMENT OF MAINTENANCE FEE, 4TH YR, SMALL ENTITY (ORIGINAL EVENT CODE: M2551); ENTITY STATUS OF PATENT OWNER: SMALL ENTITY Year of fee payment: 4 |