Power management during wakeup

The present disclosure is related to power management during wakeup of a circuit.

In some approaches, when a circuit is powered up, many transistors are turned on at a same time. Powering up the circuit is also called waking up the circuit because the circuit is woken up for operation, from a sleep mode, for example. When many transistors are turned on during waking up, a large amount of current flows through corresponding power nodes used by the transistors and results in a current spike called a wakeup peak current.

Reducing the wakeup peak current of a circuit to be within a design specification increases a wakeup time of the circuit. For example, in an approach, a wakeup circuit includes a series of wakeup stages. Each wakeup stage includes a PMOS switch connected in a chain manner. A wakeup signal is sequentially fed to each PMOS switch to turn on the chained PMOS switches one by one in a sequential manner. By turning on each PMOS switch in a sequential manner, the wakeup peak current is reduced. In such a condition, the wakeup time increases, however. Logic gates are also used in the wakeup circuit and result in increased circuit areas.

The details of one or more embodiments of the disclosure are set forth in the accompanying drawings and the description below. Other features and advantages will be apparent from the description, drawings, and claims.

FIG. 1 is a diagram of a wakeup power-management circuit, in accordance with some embodiments.

FIG. 2 is a flowchart of a method of operating the circuit in FIG. 1 , in accordance with some embodiments.

FIG. 3 is a flowchart of a method of determining a time delay for use in the circuit of FIG. 1 , in accordance with some embodiments.

FIG. 4 is a diagram of another wakeup power-management circuit, in accordance with some embodiments.

FIG. 5 is a flowchart of a method of operating the circuit in FIG. 4 , in accordance with some embodiments.

FIG. 6 is a flowchart of a method of determining a time delay for use in the circuit of FIG. 4 , in accordance with some embodiments.

FIG. 7 is a diagram of a wakeup power-management circuit that includes features similar to those in the circuits in FIGS. 1 and 4 , in accordance with some embodiments.

Like reference symbols in the various drawings indicate like elements.

FIG. 8 is a diagram of a memory macro usable in conjunction with the circuit depicted in FIG. 1 , FIG. 4 , or FIG. 7 in accordance with some embodiments.

Embodiments, or examples, illustrated in the drawings are disclosed below using specific language. It will nevertheless be understood that the embodiments and examples are not intended to be limiting. Any alterations and modifications in the disclosed embodiments, and any further applications of the principles disclosed in this document are contemplated as would normally occur to one of ordinary skill in the pertinent art.

Some embodiments have one or a combination of the following features and/or advantages. In some embodiments, two banks of transistors are used. Transistors in a first transistor bank are smaller than transistors in a second transistor bank. The first transistor bank and the second transistor bank operate to reduce a wakeup peak current, without incurring a delay penalty of a sequential wakeup. The second transistor bank includes transistors that are sized for normal operation of the circuit and that are sized independently of the wakeup peak current limitation specified for the circuit. As a result, sizing the transistors in the second transistor bank is easier. In some embodiments, an RC delay of the first transistor bank is used to delay turning on transistors in the second transistor bank to meet the wakeup peak current limitation. In some embodiments, the RC delay tracks the word size of an SRAM macro. Compared with another approach that uses a daisy chain wakeup, some embodiments of the present disclosure use less transistors in the wakeup circuit.

Circuits

FIG. 1 is a diagram of a circuit 100 , in accordance with some embodiments. Circuit 100 is a wakeup power-management circuit. For example, during a wakeup mode, a peak current of circuit 100 is designed to be below a predetermined value specified as a maximum wakeup peak current. After the wakeup mode, circuit 100 provides a supply voltage VDD (not labeled) on a line VDDM to a line VDDI. In some embodiments in which circuit 100 is used to provide supply voltage VDD on line VDDI to circuits in a memory macro, the maximum wakeup peak current of circuit 100 is determined based on the peak current when the memory macro is under a read or a write operation. For example, the maximum wakeup peak current of circuit 100 is specified to be equal to the maximum peak current when the memory macro is under a read or a write operation.

Circuit or a transistor bank 120 includes a plurality of PMOS transistors P 20 , labeled as transistors P 20 - 1 to P 20 -M where M is an integer number. A line 155 couples gates of transistors P 20 - 1 to P 20 -M and includes nodes NO1 and NO2. Node NO1 represents a point on line 155 at a gate of the first transistor P 20 - 1 in the chain of transistors P 20 - 1 to P 20 -M. Node NO2 represents a point on line 155 at a gate of the last transistor P 20 -M in the chain of transistors P 20 - 1 to P 20 -M.

Sources of PMOS transistors P 20 - 1 to P 20 -M are coupled together and receive a supply voltage VDD (not labeled) on line VDDM. Drains of PMOS transistors P 20 are coupled together and provide supply voltage VDD from line VDDM to line VDDI. Gates of transistors P 20 - 1 to P 20 -M are coupled together by line 155 , and receive a signal SD_LS. When signal SD_LS is logically low, transistors P 20 - 1 to P 20 -M are turned on. As a result, supply voltage VDD at sources of transistors P 20 - 1 to P 20 -M is transferred to drains of corresponding transistors P 20 - 1 to P 20 -M. Effectively, supply voltage VDD on line VDDM is transferred to line VDDI. In some embodiments, in a wakeup mode, transistors P 20 - 1 to P 20 -M in circuit 120 are turned on before transistors P 30 - 1 to P 30 -M in circuit 130 are turned on. Explained in a different way, circuit 120 is turned on before circuit 130 is turned on. In some embodiments, a wakeup mode starts when signal SD at an input of an inverter 112 is activated with a transition from a high to a low logical value. The wakeup mode ends when a voltage on line VDDI is at a predetermined value with reference to a value of supply voltage VDD on line VDDM, and is ready for use by other circuits that use circuit 100 . For example, the voltage on line VDDI is about 90%, 95%, 99%, etc., of supply voltage VDD. Other values of the voltage on line VDDI to be considered as an end of the wakeup mode are within the contemplated scope of the present disclosure. In some embodiments, after transistor P 30 -M in circuit 130 is turned on, the wakeup mode has ended.

FIG. 8 is a diagram of a memory macro 800 usable in conjunction with the circuit 100 depicted in FIG. 1 in accordance with some embodiments. Memory macro 800 includes a plurality of memory cells MC arranged as an array with multiple rows and columns, such as columns COL- 1 , COL- 2 , and COL-M. Some details of memory macro 800 are not depicted in FIG. 8 . Memory macro 800 is configured to receive a supply voltage from node 810 and a reference voltage from node 820 . In some embodiments, node 810 is coupled with line VDDI in FIG. 1 , and node 820 is coupled with a reference voltage VSS or ground.

In some embodiments, each of transistors P 20 - 1 to P 20 -M corresponds to a local input-output (LIO) or a column COL- 1 , COL- 2 , or COL-M of a memory array of a memory macro 800 . Further, each of transistors P 20 - 1 to P 20 -M is sized such that, during a wakeup mode, the peak current of circuit 100 is less than a predetermined current specified as the maximum wakeup peak current for the memory macro.

In some embodiments, a portion or a line segment of line 155 between a gate of a first transistor P 20 and a gate of a second transistor P 20 in the chain of transistors P 20 - 1 to P 20 -M corresponds to a time delay T20. For illustration, a circuit 125 representing one time delay T20 between the gates of transistors P 20 - 1 and P 20 - 2 on line 155 is shown. Effectively, each of transistors P 20 - 1 to P 20 -M corresponds to a time delay T20. In some embodiments, time delay T20 corresponds to a resistor-capacitor (RC) constant of a resistor R20 and a capacitor C 20 that represent the corresponding effective resistance value and effective capacitance value of the line segment between two gates of two transistors P 20 . Because the portion of line 155 between node NO1 and node NO2 corresponds to M transistors P 20 , from transistor P 20 - 1 to transistor P 20 -M, the portion of line 155 between node NO1 and node NO2 has a time delay M*T20, which, for illustration, is called time delay T55 (not labeled). In some embodiments, time delay T55 is related to the number of columns in a memory array of a memory macro by a ratio RAT (not labeled), in which RAT is a mathematical real number. For example, when M is the number of columns of the memory array, time delay T55 is proportional to RAT*M.

Symbolically expressed,

T55αRAT*M

For example,

T55 α (0.25*M) or (0.5*M) or (1*M) or (2*M) or (3*M), etc.

Time delay T55 varies also based on a length of the portion of line 155 between nodes NO1 and NO2, which depends on the number of transistors P 20 in circuit 120 , the distance between gates of two transistors P 20 in the chain of transistors P 20 - 1 to P 20 -M, the RC constant of the line between two transistors P 20 , etc. In some embodiments, a value of time delay T55 is adjusted by adjusting a combination of a value of time delay T20, a value of ratio RAT, and/or a number of transistors P 20 in circuit 120 . A value of time delay T20 is adjusted by adjusting the RC constant value of resistor R20 and capacitor C 20 . In some embodiments, the RC constant value of resistor R20 and capacitor C 20 depends on a thickness of the line segment between two gates of two transistors P 20 . In some embodiments, based on a fixed number of transistors P 20 , such as M transistors P 20 - 1 to P 20 -M, in circuit 120 , after a value of time delay T20 is selected, a predetermined value of time delay T55 is obtained by varying a value of ratio RAT. For another example, after a ratio RAT is selected, a predetermined value of time delay T55 is obtained by adjusting a value of time delay T20. Obtaining a predetermined value of time delay T55 is used to obtain a time delay TDLY explained below with reference to FIG. 3 .

Because of the time delay T20 between two gates of two transistors P 20 , a second transistor P 20 succeeding a first transistor P 20 in the chain of transistors P 20 - 1 to P 20 -M is turned on a time delay T20 after the first transistor P 20 is turned on. For example, transistor P 20 - 2 is turned on a time delay T20 after transistor P 20 - 1 is turned on. Transistor P 20 - 3 is turned on a time delay T20 after transistor P 20 - 2 is turned on, etc. Transistor P 20 -M is then turned on last in the chain of transistors P 20 - 1 to P 20 -M.

For illustration, a voltage drop across a source and a drain of a transistor P 20 is called a voltage VSD20 (not labeled). For example, a voltage drop across a source and a drain of transistor P 20 - 1 is called VSD20-1, and a voltage drop across a source and a drain of transistor P 20 -M is called VSD20-M. A current flowing through a source and a drain of a transistor P 20 is called a current ISD20 (not labeled). A voltage drop across node VDDM and node VDDI is called voltage VDMMI.

A circuit or a transistor bank 130 includes a plurality of PMOS transistors P 30 , labeled as transistors P 30 - 1 to P 30 -M. Sources of PMOS transistors P 30 - 1 to P 30 -M are coupled together and receive supply voltage VDD on line VDDM. Drains of PMOS transistors P 30 - 1 to P 30 -M are coupled together and provide supply voltage VDD from line VDDM to line VDDI. Gates of PMOS transistors P 30 - 1 to P 30 -M are coupled together and receive a signal SD_ARRAY. When signal SD_ARRAY is logically low, transistors P 30 - 1 to P 30 M are turned on, and supply voltage VDD at sources of transistors P 30 - 1 to P 30 -M is transferred to drains of corresponding transistors P 30 - 1 to P 30 -M. Effectively, supply voltage VDD on line VDDM is transferred to line VDDI. In some embodiments, line VDDM that couples the sources of transistors P 30 - 1 to P 30 -M is coupled with line VDDM that couples the sources of transistors P 20 - 1 to P 20 -M in a grid structure. In some embodiments, during a wakeup mode, circuit 130 is turned on after circuit 120 is turned on. For example, after the last transistor P 20 -M in the chain of transistors P 20 - 1 to P 20 -M is turned on, transistors P 30 - 1 to P 30 -M are turned on in the sequence of transistors P 30 - 1 , P 30 - 2 , . . . , P 30 -M.

In some embodiments, each transistor P 30 - 1 to P 30 -M in circuit 130 corresponds to a transistor P 20 - 1 to P 20 -M in circuit 120 . Effectively, each transistor P 30 - 1 to P 30 -M also corresponds to a local input-output (LIO) and a column COL- 1 , COL- 2 , or COL-M of the memory macro 800 having the LIOs. In some embodiments, each transistor P 30 - 1 to P 30 -M is sized such that each transistor P 30 - 1 to P 30 -M provides sufficient current and voltage VDD to the drain of corresponding transistor P 30 - 1 to P 30 -M to run a particular application, a circuit, a device, etc. For example, when circuit 100 is used in a memory macro 800 , each of transistors P 30 - 1 to P 30 -M is sized such that transistor P 30 - 1 to P 30 -M provide sufficient current and voltage to perform a read or a write operation of the memory macro 800 .

A node NO3 represents a point at a gate of the first transistor P 30 - 1 in the chain of transistors P 30 - 1 to P 30 -M of circuit 130 .

For illustration, a voltage drop across a source and a drain of a transistor P 30 is called a voltage VSD30 (not labeled). For example, a voltage drop across a source and a drain of transistor P 30 - 1 is called VSD30-1, and a voltage drop across a source and a drain of transistor P 30 -M is called VSD30-M. A current flowing through a source and a drain of a transistor P 30 is called a current ISD30 (not labeled). When currents ISD20 of transistors P 20 - 1 to P 20 -N and currents ISD30 of transistors P 30 - 1 to P 30 -N start flowing at about the same time, a current spike occurs, resulting in a peak current. In some embodiments, the turn-on of circuits 120 and 130 is controlled to reduce the wakeup peak current of circuit 100 and of circuits that use circuit 100 .

A circuit 110 is used to control the turn-on and turn-off of circuits 120 and 130 . For illustration, when circuit 100 enters the wakeup mode, signal SD at an input of an inverter 112 transitions from a high to a low logical value. By operations of inverters 112 and 114 , a signal SD_LS at an output of inverter 114 also transitions from a high to a low logical value. Because signal SD_LS is also at gates of transistors P 20 - 1 to P 20 -M in circuit 120 , PMOS transistors P 20 - 1 to P 20 -M in circuit 120 are turned on.

In some embodiments, a signal SD_DEL is signal SD_LS delayed by time delay T55. As a result, signal SD_DEL also transitions from a high to a low logical value after time delay T55. Because signal SD_LS and signal SD_DEL transition from a high to a low logical value, by operation of a NOR gate 115 and inverters 116 and 118 , signal SD_ARRAY also transitions from a high to a low logical value, subject to a time delay as explained below. Consequently, PMOS transistors P 30 - 1 to P 30 -M in circuit 130 are also turned on. Effectively, circuit 130 is turned on after circuit 120 by a time delay as explained below.

Time Delays

A time delay TDLY includes a time delay for signal SD to travel through inverters 112 , 114 to node NO1, through transistors P 20 - 1 to P 20 -M to node NO2, a delay circuit 140 , OR gate 115 , inverter 116 , and inverter 118 to node NO3. Time delay T55 of line 155 between node NO1 and node NO2 is explained above.

Delay circuit 140 provides an adjustable time delay T40 (not labeled). In some embodiments, circuit 140 includes a resistor R10 and a capacitor C 10 to result in a time delay T40. In some embodiments, resistor R10 is adjustable, capacitor C 10 is adjustable, and time delay T40 is adjustable by tuning the resistance of resistor R10 and/or tuning the capacitance of capacitor C 10 . Different ways for circuit 140 to provide adjustable time delay T40 are within the contemplated scope of the present disclosure. For example, circuit 140 includes a plurality of pairs of resistors and capacitors that can be switched onto line 115 to increase the time delay and switched out of line 115 to decrease the time delay, etc. Circuit 140 shown between node NO2 and OR gate 115 is for illustration. Other locations of circuit 140 to cause a delay between node NO2 and node NO3 are within the contemplated scope of the present disclosure. For example, circuit 140 is in between inverter 118 and node NO3. In some embodiments, if time delay T40 is zero, circuit 140 is not included in circuit 100 .

Each of OR gate 115 , inverter 116 , and inverter 118 has a corresponding time delay. For simplicity, the time delay through OR gate 115 , inverter 116 , and inverter 118 is considered the same, and is called a gate time delay TGATE. A time delay between circuit 140 and node NO3 depends on the number of logic gates between circuit 140 and node NO3. As a result, additional gates are added between inverter 118 and node NO3 to increase time delay TDLY, for example. For another example, at least one pair of inverters is coupled in series, and is coupled between inverter 118 and node NO3 to increase the time delay for time delay TDLY.

Effectively, time delay TDLY includes time delays T55, T40, and corresponding time delays TGATEs of logic gates between circuit 140 and node NO3.

Time Delay and Wakeup Peak Current

As indicated above, the drains of transistors P 20 - 1 to P 20 -M and P 30 - 1 to P 30 -M are coupled together. In some embodiments, when circuit 100 is in a shutdown mode, the drains of transistors P 20 - 1 to P 20 -M and P 30 - 1 to P 30 -M are at a voltage VDSD (not labeled), for example. Without leakage currents of transistors P 20 - 1 to P 20 -M and/or other circuits that use circuit 100 , voltage VDSD is about a reference voltage VSS, which is some embodiments, is ground. Because of the leakage currents, however, voltage VDSD at the drains of transistors P 20 - 1 to P 20 -M and P 30 - 1 to P 30 -M is pulled away from voltage VSS a voltage amount depending on the leakage current. For illustration, the drains of transistors P 20 - 1 to P 20 -M and P 30 - 1 to P 30 -M are considered to be at voltage VSS.

In some embodiments, when voltage VSD30 dropped across a source and a drain of a transistor P 30 increases, corresponding current ISD30 of the same transistor P 30 increases. In contrast, when voltage VSD30 decreases, the corresponding current ISD30 decreases. In some embodiments, voltages VSD30 of transistors P 30 - 1 to P 30 -M are decreased so that corresponding currents IDS30 of the corresponding transistors are decreased. For example, voltage VSD30 of a transistor P 30 is decreased by raising a voltage at a drain of corresponding transistor P 30 . For another example, after transistor P 20 - 1 in circuit 120 is turned on, the drain of transistor P 20 - 1 instantly has voltage VDD from the source of transistor P 20 - 1 . Transistors P 20 - 2 , P 20 - 3 , . . . , P 20 -M and transistors P 30 - 1 , P 30 - 2 , . . . , P 30 -M are then each turned on sequentially. When some transistors in the chain of transistors P 20 - 1 to P 20 -M and P 30 - 1 to P 30 -M are turned on, the voltage at the drains of the transistors that are turned on instantly have voltage VDD from the corresponding sources of the turned-on transistors. The voltage VDD of the transistors that have been turned on propagates to the drains of transistors that have not been turned on. As a result, the voltage at the drains of transistors in the chain of transistors P 20 - 1 to P 20 -M and transistors P 30 - 1 to P 30 -M that have not been turned on is raised to be higher than voltage VSS. Effectively, voltage VSD dropped across the sources and the drains of transistors that have a raised voltage at the drain decreases. Similarly, currents ISD flowing through the sources and the drains of transistors that have a raised voltage at the drain also decreases.

In some embodiments, time delay TDLY is selected at a predetermined delay value such that when the low voltage value of signal SD_ARRAY reaches the gate of transistor P 30 - 1 to turn on transistors P 30 - 1 to P 30 -M, the peak current of circuit 100 is less than a predetermined current value. Explained in a different way, during the wakeup mode while transistors P 30 - 1 to P 30 -M are being turned on, the wakeup peak current is less than the predetermined current value, and is therefore compliant with a design specification for a particular application. Effectively, in circuit 100 , transistors P 20 - 1 to P 20 -M and P 30 - 1 to P 30 -M are turned on such that the wakeup peak current is less than the predetermined current value.

Methods

FIG. 2 is a flowchart 200 , in accordance with some embodiments. Flowchart 200 illustrates an operation of circuit 100 in FIG. 1 for a wakeup peak current to comply with a design specification.

In operation 205 , signal SD in FIG. 1 is caused to transition from a high to a low logical value. In some embodiments, in such a condition, circuit 100 is out of a shutdown mode and enters a wakeup or power-up mode.

In operation 210 , by operation of inverters 112 and 114 , signal SD_LS follows signal SD to transition from a high to a low logical value.

In operation 215 , transistors P 20 - 1 to P 20 -M are turned on in the sequence of transistors P 20 - 1 , P 20 - 2 , P 20 - 3 , etc. In some embodiments, during the time transistors P 20 - 1 to P 20 -M are being turned on, the peak current of circuit 100 is less than a predetermined value.

As more transistors P 20 in the chain of transistors P 20 - 1 to P 20 -M are turned on in combination with the elapse of time, the voltage at the drains of transistors in the chain of transistors P 20 - 1 to P 20 -M and P 30 - 1 to P 30 -M that have not been turned on is raised above reference voltage VSS.

In operation 220 , after a time delay TDLY from the time signal SD transitions to a low logical value, signal SD_ARRAY at node NO3 also transitions from a high to a low logical value. As a result, transistors P 30 - 1 to P 30 -M are turned on in the sequence of transistors P 30 - 1 , P 30 - 2 , P 30 - 3 , etc.

In operation 225 , after transistor P 30 -M is turned on, the drains of all transistors P 30 - 1 to P 30 -M have a supply voltage VDD. By the time transistor P 30 -M is turned on, the wakeup mode has ended. As a result, supply voltage VDD on line VDDI is ready for use by other circuits that use circuit 100 as a wakeup circuit.

After operation 220 above, by the time transistor P 30 - 1 is turned on, a time delay TDLY from the time signal SD transition to a low logical value has elapsed. Further, the voltage at the drains of various transistors P 30 - 1 to P 30 -M that have not been turned on has been raised above reference voltage VSS. As a result, voltage VSD30 dropped across the sources and the drains of transistors P 30 - 1 to P 30 -M that have not been turned on has decreased. Corresponding currents ISD30 also decreases. Effectively, currents ISD30 of transistors P 30 - 1 to P 30 -M decreases, and the peak current of circuit 100 decreases, compared with the situation when the voltage at the drains of the transistors P 30 that have not been turned on stays at reference voltage VSS. In some embodiments, time delay TDLY is configured such that the wakeup peak current is lower than a predetermined value to comply with a design specification that specifies a maximum wakeup peak current. Further, the wakeup time that includes the time when signal SD transitions from a high to a low logical value to the time the voltage on line VDDI is ready for use by other circuits is faster than a wakeup time of other approaches. As a result, various embodiments of the present disclosure are advantageous over other approaches. Explained in a different way, in other approaches, for the wakeup time to comply with the design specification, a wakeup time in the other approaches is longer that the wakeup time of the present disclosure.

FIG. 3 is a flowchart of a method 300 , in accordance with some embodiments. Method 300 illustrates operations to determine time delay TDLY for a wakeup peak current to comply with a design specification specifying a maximum wakeup peak current. The wakeup peak current in the illustration of FIG. 3 is with reference to supply voltage VDD.

In operation 305 , the number of transistors P- 20 in circuit 120 is determined. In some embodiments, M number of transistors P 20 is selected wherein M is the number of columns of a memory array in a memory macro. Other values of M are within the contemplated scope of the present disclosure.

In operation 310 , a size of each of transistors P 20 - 1 to P 20 -M is determined. In some embodiments, the size of each of transistors P 20 - 1 to P 20 -M is determined such that when transistors P 20 - 1 to P 20 -M are turned on in the sequence of transistors P 20 - 1 , P 20 - 2 , P 20 - 3 , etc., the peak current of circuit 100 is less than a predetermined value. In some embodiments, operation 310 is performed in conjunction with operation 305 , such that the peak current of circuit 100 is less than the predetermined value. For example, the number of transistors P 20 and/or the size of each transistor P 20 are selected so that the peak current is less than the predetermined value. For another example, the number of transistors P 20 is selected to be different from the number of columns of the memory array of the memory macro that uses circuit 100 as a wakeup circuit.

In some embodiments, the number of transistors P 20 and/or the size of each of transistors P 20 - 1 to P 20 -M are determined based on simulation. In some embodiments, the size of each of transistors P 20 - 1 to P 20 -M is the same as that of another transistor P 20 . Transistors P 20 having different sizes are within the contemplated scope of the present disclosure.

In operation 315 , a size of each of transistors P 30 - 1 to P 30 -M is determined. In some embodiments, the size of each transistor P 30 - 1 to P 30 -M is determined such that each of transistors P 30 - 1 to P 30 -M sufficiently provides voltage VDD and current ISD30 to run a particular application. For example, voltage VDD and current ISD30 each satisfy a predetermined value. In some embodiments, a size of each of transistors P 30 - 1 to P 30 -M is the same as that of another transistor P 30 . Transistors P 30 having different sizes are within the contemplated scope of the present disclosure. In various embodiments, the size of each of transistors P 30 - 1 to P 30 -M is determined independent of a specified maximum wakeup peak current.

In operation 320 , based on the size of each of transistors P 30 - 1 to P 30 -M, a current ISD30 of each of transistors P 30 - 1 to P 30 -M is determined.

In operation 325 , time delay TDLY is determined. In some embodiments, based on simulation, time delay TDLY is determined such that when transistors P 30 - 1 to P 30 -M are being turned on, the peak current of circuit 100 is less than a predetermined value specified as a maximum wakeup peak current. For example, if the peak current is to increase, time delay TDLY is decreased. But if the peak current is to decrease, time delay TDLY is increased. Time delay TDLY is adjusted by adjusting one or a combination of time delay T55, and time delay T40, and/or selecting different numbers of logic gates between inverter 118 and node NO3. For illustration, a time delay T40 is selected. The number of logic gates between inverter 118 and node NO2 is also selected. Time delay T55 is then adjusted as explained above, to adjust time delay TDLY.

Other Power Management Mechanisms

FIG. 4 is a diagram of a circuit 400 , in accordance with some embodiments. Circuit 400 is a wakeup power-management circuit. For example, during a wakeup mode, the peak current of circuit 400 is designed to below a predetermined value specified as a maximum reference wakeup peak current. At the end of the wakeup mode, circuit 400 provides reference supply voltage VSS (not labeled) on a line VSSM to a line VSSI for use by other circuits that use circuit 400 .

In some embodiments, memory macro 800 depicted in FIG. 8 is usable in conjunction with the circuit 400 depicted in FIG. 4 in accordance with some embodiments. In some embodiments, node 810 is coupled with supply voltage VDD, and node 820 is coupled with line VSSI in FIG. 4 .

In some embodiments, when circuit 400 is in the shutdown mode, the drains of transistors N 20 - 1 to N 20 -M and N 30 - 1 to N 30 -M are at a voltage value VSSD (not labeled), for example. The voltage value VSSD varies depending on leakage currents of transistors N 20 - 1 to N 20 -M and/or other circuits that use circuit 100 . In some embodiments, voltage value VSSD is about half of supply voltage VDD. Other voltage values of VSSD are within the contemplated scope of the present disclosure.

Compared with circuit 100 , in circuit 400 , circuits 410 , 420 , 430 , and 440 , correspond to circuits 110 , 120 , 130 , and 140 , respectively. Nodes NO41, NO42, and NO43 correspond to nodes NO1, NO2, and NO3, respectively. Signals SD_LSB, SD_DELB, SD_ARRAYB corresponds to signals SD_LS, SD_DEL, and SD_ARRAY, respectively.

In circuit 410 , inverters 412 , 414 , 416 , 418 and OR gate 415 correspond to inverters 112 , 114 , 116 , 118 and OR gate 115 , respectively. An inverter 419 is added to invert signal SD_DELB to signal SD_DEL.

In circuit 420 , transistors N 20 - 1 to N 20 -M are NMOS transistors, however, and are collectively referred to as transistors N 20 . Effectively, NMOS transistors N 20 - 1 to N 20 -M correspond to PMOS transistors P 20 - 1 to P 20 -M, respectively. A line 455 corresponds to line 155 . A time delay T420 corresponds to time delay T20. A line VSSM corresponds to line VDDM. A line VSSI corresponds to line VDDI.

Compared In circuit 430 , transistors N 30 - 1 to N 30 -M are also NMOS transistors, and are collectively referred to as transistors N 30 . Effectively, NMOS transistors N 30 - 1 to N 30 -M correspond to PMOS transistors P 30 - 1 to P 30 -M.

In circuit 440 , a resistor R40 and a capacitor C 40 correspond to resistor R10 and capacitor C 10 , respectively.

FIG. 5 is a flowchart 500 , in accordance with some embodiments. Flowchart 500 illustrates an operation of circuit 400 in FIG. 4 for a wakeup reference current to comply with a design specification.

In operation 505 , signal SD in FIG. 4 transitions from a high to a low logical value. In some embodiments, in such a condition, circuit 400 is out of a shutdown mode and enters a wakeup or power-up mode.

In operation 510 , by operation of inverter 412 , signal SD_LSB is an inverse of signal SD and transitions from a low to a high logical value.

In operation 515 , transistors N 20 - 1 to N 20 -M are turned on in the sequence of transistors N 20 - 1 , N 20 - 2 , N 20 - 3 , etc. In some embodiments, during the time transistors N 20 - 1 to N 20 -M are being turned on, the peak current of circuit 400 is less than a predetermined value.

As more transistors N 20 in the chain of transistors N 20 - 1 to N 20 -M are turned on in combination with the elapse of time, the voltage at the drains of transistors in the chain of transistors N 20 - 1 to N 20 -M and N 30 - 1 to N 30 -M that have not been turned on is lowered below voltage value VSSSD.

In operation 520 , after a time delay TDLYN from the time signal SD transitions from a high to a low logical value, signal SD_ARRAYB at node NO43 also transitions from a low to a high logical value. As a result, transistors N 30 - 1 to N 30 -M are turned on in the sequence of transistors N 30 - 1 , N 30 - 2 , N 30 - 3 , etc.

In operation 525 , after transistor N 30 -M is turned on, the drains of all transistors N 30 - 1 to N 30 -M have a reference supply voltage VSS from the sources of transistors N 30 - 1 to N 30 -M coupled with line VSSM. In other words, circuit 400 has completed the wakeup mode. Reference supply voltage VSS is therefore ready for used by other circuits that use circuit 400 as a wakeup circuit.

After operation 520 above, by the time transistor N 30 - 1 is turned on, a time delay TDLYN from the time signal SD transitions to a high logical value has elapsed. Further, the voltage at the drains of various transistors N 30 - 1 to N 30 -M that have not been turned on has been lowered below voltage value VSSD. As a result, a voltage VDS30N (not labeled) dropped across the drains and the sources of transistors N 30 - 1 to N 30 -M that have not been turned on has decreased. Corresponding currents IDS30N also decreases. Effectively, currents IDS30N of transistors N 30 - 1 to N 30 -M decreases, and the wakeup peak current of circuit 400 decreases, compared with the situation when the voltage at the drains of the transistors that have not been turned on stays at voltage level VSSD. In some embodiments, time delay TDLYN is configured such the wakeup peak current is lower than a predetermined value to comply with a design specification that specifies a maximum wakeup peak current with reference to reference supply voltage VSS. Further, the wakeup time that includes the time when signal SD transitions from a high to a low logical value to the time transistor N 30 -M is turned on is faster than a wakeup time of other approaches. As a result, various embodiments of the present disclosure are advantageous over other approaches. Explained in a different way, in other approaches, for the wakeup peak current to comply with the design specification, the wakeup time in those approaches is longer than the wakeup time of the present disclosure.

FIG. 6 is a flowchart of a method 600 , in accordance with some embodiments. Method 600 illustrates operations to determine time delay TDLYN for a wakeup peak current to comply with a design specification specifying a maximum wakeup peak current with reference to reference supply voltage VSS.

In operation 605 , the number of transistors N 20 in circuit 420 is determined. In some embodiments, M number of transistors N 20 is selected wherein M is the number of columns of a memory array in a memory macro. Other values of M are within the contemplated scope of the present disclosure.

In operation 610 , a size of each of transistors N 20 - 1 to N 20 -M is determined. In some embodiments, the size of each of transistors N 20 - 1 to N 20 -M is determined such that when transistors N 20 - 1 to N 20 -M are turned on in the sequence of transistors N 20 - 1 , N 20 - 2 , N 20 - 3 , etc., the peak current of circuit 400 is less than a predetermined value. In some embodiments, operation 610 is performed in conjunction with operation 605 , such that the peak current of circuit 400 is less than the predetermined value. For example, the number of transistors N 20 and/or the size of each transistor N 20 are selected so that the peak current is less than the predetermined value. For another example, the number of transistors N 20 is selected to be different from the number of columns of the memory array of the memory macro that uses circuit 400 as a wakeup circuit.

In some embodiments, the number of transistors N 20 and/or the size of each of transistors N 20 - 1 to N 20 -N are determined based on simulation. In some embodiments, the size of each of transistors N 20 - 1 to N 20 -M is the same as that of another transistor N 20 . Transistors N 20 having different sizes are within the contemplated scope of the present disclosure.

In operation 615 , a size of each of transistors N 30 - 1 to N 30 -M is determined. In some embodiments, the size of each transistor N 30 - 1 to N 30 -M is determined such that each of transistors N 30 - 1 to N 30 -M sufficiently provides reference voltage VSS and current ISD 30 N to run a particular application. For example, voltage VSS and current ISD 30 N each satisfy a predetermined value. In some embodiments, sizes of transistors N 30 - 1 to N 30 -M are each the same. Transistors N 30 - 1 to N 30 -M having different sizes are within the contemplated scope of the present disclosure. In various embodiments, the size of each of transistors N 30 - 1 to N 30 -M is determined independent of a specified maximum wakeup peak current with reference to reference supply voltage VSS.

In operation 620 , based on the size of each of transistors N 30 - 1 to N 30 -M, a current ISD30N of each of transistors N 30 - 1 to N 30 -M is determined.

In operation 625 , time delay TDLYN is determined. In some embodiments, based on simulation, time delay TDLYN is determined such that when transistors N 30 - 1 to N 30 -M are being turned on, the peak current of circuit 400 is less than a predetermined value specified as a maximum wakeup peak current with reference to reference supply voltage VSS. For example, if the peak current is to increase, time delay TDLYN is decreased. But if the peak current is to decrease, time delay TDLYN is increased. Time delay TDLYN is adjusted by adjusting one or a combination of time delay T455, and time delay T420, and/or determining the number of logic gates between inverter 418 and node NO43. For illustration, a time delay T440 is selected. The number of logic gates between inverter 418 and node NO43 is also selected. Time delay T455 is then adjusted as explained above, to adjust time delay TDLYN.

FIG. 7 is a diagram of a wakeup power-management circuit 700 , in accordance with some embodiments. Circuit 700 includes features of circuit 100 in FIG. 1 and circuit 400 in FIG. 4 . As a result, circuit 700 includes circuit elements similar to those of circuits 100 and 400 and other circuits to implement a combination of circuits 100 and 400 in circuit 700 . For simplicity, various elements of circuits 100 and 400 are not shown, nor labeled. For example, time delay 140 in FIG. 1 and time delay 440 are not shown in FIG. 7 . A circuit 710 includes circuit elements of circuit 110 in FIG. 1 and circuit 410 in FIG. 4 , and functions for circuits 100 and 400 to operate together.

For example, circuit 700 includes transistor banks 120 and 130 of circuit 100 and transistor banks 420 and 430 of circuit 400 . A circuit 710 includes inverters 112 , 114 , 116 , 118 of FIG. 1 , and includes inverter 419 of FIG. 4 . OR-gate 715 is used in place of OR-gate 115 and 415 .

In some embodiments, memory macro 800 depicted in FIG. 8 is usable in conjunction with the circuit 700 depicted in FIG. 7 in accordance with some embodiments. In some embodiments, node 810 is coupled with line VDDI in FIG. 7 , and node 820 is coupled with line VSSI in FIG. 7 .

In some embodiments, a circuit comprises a first set of first transistors and a second set of transistors. The first transistors are turned on in a sequential manner. The second transistors are turned on in a sequential manner after the first transistors are turned on. A transistor of the first set of first transistors corresponds to a first time delay. The first set of first transistors corresponds to a second time delay that is a multiple of the first time delay.

Some embodiments regard a method. In the method, transistors in a first set of transistors are turned on in a sequential manner, then transistors in a second set of transistors are turned in a sequential manner. A wakeup current value of a circuit using the first set of transistors and the second set of transistors is less than a predetermined value during turning on the transistors in the first set of transistors or during turning on the transistors in the second set of transistors. A delay time corresponding to the first set of transistors is related to a number of columns in an array of memory cells by a ratio that is mathematically a real number.

Some embodiments regard a method. In the method, one or a combination of a size of transistors in the first set of transistors and a number of transistors in the first set of transistor is adjusted such that a current flowing through the transistors in the first set of transistors is less than a predetermined value. A time delay is determined to turn on transistors in a second set of transistors after turning on a first transistor in the first set of transistors such that a current flowing through the transistors in the first set of transistors and the transistors in the second set of transistors is less than the predetermined value. The time delay is related to a number of columns of a memory array by a number that is mathematically a real number.

A number of embodiments have been described. It will nevertheless be understood that various modifications may be made without departing from the spirit and scope of the disclosure. For example, various transistors being shown as a particular dopant type (e.g., N-type or P-type Metal Oxide Semiconductor (NMOS or PMOS)) are for illustration purposes. Embodiments of the disclosure are not limited to a particular type. Selecting different dopant types for a particular transistor is within the scope of various embodiments. A low or high logical value of various signals used in the above description is also for illustration. Various embodiments are not limited to a particular logical value when a signal is activated and/or deactivated. Selecting different logical values is within the scope of various embodiments. In various embodiments, a transistor functions as a switch. A switching circuit used in place of a transistor is within the scope of various embodiments. In various embodiments, a source of a transistor can be configured as a drain, and a drain can be configured as a source.

The above illustrations include exemplary steps, but the steps are not necessarily performed in the order shown. Steps may be added, replaced, changed order, and/or eliminated as appropriate, in accordance with the spirit and scope of disclosed embodiments.