Cross-Reference to Related Application

[0001] This application claims priority to U.S. Nonprovisional Application No. 14/257,212, filed on April 21, 2014, which is herein incorporated by reference in its entirety.

Technical Field

[0002] This application relates to level-shifters, and more particularly to a wide- range level-shifter.

Background

[0003] As semiconductor technology has advanced into the deep submicron regime, the power supply voltage is scaled down in concert with the scaling down of transistor dimensions. For example, microprocessors are now manufactured with transistors powered by a sub-one volt power supply voltage. But these modern systems may need to interface with peripheral devices such as memories that operate on higher supply voltages. The signal flow from the low-voltage domain to the high-voltage domain requires a shift up in voltage. Conversely, the signal flow from the high-voltage domain to the low-voltage domain requires a shift down in voltage. A conventional level-shifter 100 is shown in Figure 1 that may perform such voltage level shifts between an input signal (IN) and an output signal (OUT). In this embodiment, the input signal is a low-voltage-domain (VDDL) signal whereas the output signal is a high- voltage-domain (VDDH) signal. However, level-shifter 100 is readily modified to instead shift down in voltage. [0004] The input signal drives a gate of an NMOS transistor MNl. If the input signal is low (ground or VSS), transistor MNl switches off, allowing a node Nl to float. The input signal also drives an inverter INV that produces an inverted input signal that in turn drives a gate of an NMOS transistor MN2. Inverter INV is powered by a power supply node providing the low-voltage-domain power supply voltage VDDL. Thus, inverter INV will charge the gate of transistor MN2 to VDDL when the input signal is low, which switches on transistor MN2 to pull node N2 to ground.

[0005] Node N2 couples to a gate of a PMOS transistor MPl that has its drain coupled to node Nl . Transistor MPl is cross-coupled with a PMOS transistor MP2. The input signal also drives a gate of a PMOS transistor MP3 in series with transistor MPl. When the input signal is low, both transistors MP3 and MPl will be on, which charges node Nl to a high-voltage-domain power supply voltage VDDH. Node Nl drives the gate of transistor MP2 coupled to node N2. Transistor MP2 will thus be off when the input signal is low. Another PMOS transistor MP4 that has its gate driven by the inverted input signal is in series with transistor MP3.

[0006] In response to the input signal switching high to VDDL, transistor MNl will switch on and transistor MN2 will switch off. Output node N2, which had been discharged while the input signal was low, must then float until transistor MP2 can be switched on. In turn, transistor MP2 can't switch on until transistor MNl can discharge node Nl. However, transistor MPl is still momentarily on and attempting to keep node Nl charged, which thus fights with transistor MNl discharging node Nl. Transistor MP3 is only weakly on because VDDL is effectively a weak zero with regard to VDDH. Transistor MP3 thus assists transistor MNl in terms of discharging node Nl by restricting the flow of charge to transistor MPl. Once node Nl is discharged, transistor MP2 will switch on. Since transistor MP4 will already be on due to the inverted input signal being driven low, the switching on of transistor MP2 will charge the output signal to VDDH. An analogous struggle occurs between transistors MN2 and MP2 when the inverted signal is driven to VDDL in response to the input signal transitioning low.

[0007] This fight between the NMOS and PMOS transistors in level-shifter 100 may be alleviated by weakening PMOS transistors MP3 and MPl (as well as transistors MP4 and MP2). But such a weakening adversely affects timing as each PMOS stack MP3 MP1 and MP4/MP2 must also pull up its corresponding node (Nl or N2, respectively) depending upon whether the input signal is high or low. There is thus a minimum amount of strength necessary for the PMOS stacks to meet desired timing requirements. Since the PMOS stacks must be left relatively strong, there is a limit to the voltage range for level-shifter 100. In that regard, as VDDL drops ever lower at the modern process nodes, transistor MPS turns on ever stronger with regard to VDDL functioning as an effective zero in keeping transistor MP3 on despite the input signal transitioning high to VDDL. This input range for level-shifter 100 is also affected by the process corner. Should level-shifter 100 be manufactured in a process corner that makes NMOS transistors relatively weak compared to the corresponding PMOS transistors, the PMOS/NMOS struggle with regard to discharging node Nl is aggravated.

[0008] Accordingly, there is a need in the art for level-shifters with improved input voltage range and operating speed.

Summary

[0009] A level-shifter is provided that includes a pair of cross-coupled PMOS transistors. The level-shifter also includes two PMOS transistor stacks. A first PMOS stack includes a first one of the cross-coupled PMOS transistors and a second one of the stacks includes a second one of the cross-coupled PMOS transistors. [0010] The first PMOS stack includes a first weak PMOS keeper transistor that has its gate driven by an input signal and is coupled in parallel with a first PMOS pull- up transistor that is relatively strong in comparison to the first weak keeper transistor. The level-shifter includes a control circuit that controls whether the first pull-up transistor is switched on or off. The second PMOS stack is analogous to the first PMOS stack. In that regard, the second PMOS stack includes a second weak keeper PMOS transistor that has its gate driven by an inverted version of the input signal and is coupled in parallel with a second pull-up PMOS transistor that is relatively strong with respect to the second weak keeper transistor. The control circuit also controls whether the second pull-up transistor is switched on or off.

[0011] The control circuit controls the first pull-up transistor to be switched on when the first PMOS stack functions to charge a first node to a first power supply voltage in response to the input signal transitioning from a second power supply voltage to ground. Conversely, the control circuit controls the first pull-up transistor to be switched off when the first node is discharged in response to the input signal transitioning from ground to the second power supply voltage. In this fashion, the level- shifter has the advantage of a weak first PMOS stack with regard to discharging the first node yet has the advantage of a strong first PMOS stack with regard to charging the first node to the first power supply voltage. The second power supply voltage may be less than or greater than the first power supply voltage.

[0012] The control circuit also controls the second pull-up transistor to be switched on when the second PMOS stack functions to charge a second node to the first power supply voltage in response to the input signal transitioning from ground to the second power supply voltage. Conversely, the control circuit controls the second pull- up transistor to be switched off when the second node is discharged in response to the input signal transitioning from the second power supply voltage to ground. In this fashion, the level-shifter has the advantage of a weak second PMOS stack with regard to discharging the second node yet has the advantage of a strong second PMOS stack with regard to charging the second node to the first power supply voltage. These

advantageous features may be better appreciated with reference to the following detailed description.

Brief Description of the Drawings

[0013] Figure 1 is a circuit diagram of a conventional level-shifter.

[0014] Figure 2 is a high-level circuit diagram of a wide-range level-shifter in accordance with an embodiment of the disclosure.

[0015] Figure 3 is a circuit diagram of a wide-range level-shifter in accordance with an embodiment of the disclosure.

[0016] Figure 4 is a timing diagram for various signals within the level-shifter of Figure 3 in response to the input signal transitioning from ground to a power supply voltage.

[0017] Figure 5 is a timing diagram for various signals within the level-shifter of Figure 3 in response to the input signal transitioning from a power supply voltage to ground.

[0018] Figure 6 is a flowchart for a method of operation for the level-shifter of Figure 2.

[0019] Embodiments of the disclosed level-shifter and their advantages are best understood by referring to the detailed description that follows. It should be appreciated that like reference numerals are used to identify like elements illustrated in one or more of the figures. Detailed Description

[0020] A level-shifter is provided that offers a PMOS stack that is selectively weakened with regard to an NMOS transistor pulling an internal node low with regard to a first binary transition in the input signal. The PMOS stack is also selectively strengthened when the PMOS stack must charge the internal node high with regard to a complementary second binary transition of the input signal.

[0021] This selective weakening and strengthening of the PMOS stacks may be better understood with reference to the level-shifter 200 shown in Figure 2. Transistors MNl, MN2, MPl, and MP2 as well as inverter INV operate as discussed with regard to level-shifter 100. Their operation will be reviewed again for additional clarity.

[0022] The input signal (IN) supplied on an input signal node that drives the gate of NMOS transistor MNl, which has its source coupled to ground and its drain coupled to node Nl . If the input signal is low (ground or VSS), transistor MNl switches off, allowing node Nl to float. The input signal also drives an inverter INV that in turn drives a gate of an NMOS transistor MN2 with an inverted version of the input signal. Transistor MN2 has its source coupled to ground and its drain coupled to node N2 that supplies the output signal. Inverter INV is powered by a power supply node providing the low-voltage-domain power supply voltage VDDL. Thus, inverter INV will charge the gate of transistor MN2 to VDDL when the input signal is low, which switches on transistor MN2 to pull node N2 to ground. Conversely, when the input signal is high (VDDL), transistor MNl will switch on to pull node Nl to ground and transistor MN2 will switch off, allowing node N2 to float.

[0023] PMOS transistors MPl and MP2 are cross-coupled with regard to their drains and gates. The drain of transistor MPl couples to node Nl that also couples to the gate of transistor MP2. Similarly, the drain of transistor MP2 couples to node N2 that also couples to the gate of transistor MP1. When node Nl is discharged, transistor MP2 will thus switch on. Conversely, transistor MP1 switches on responsive to node N2 discharging to ground. Transistors MP1 and MP2 are each part of a corresponding PMOS stack. In level-shifter 100, transistor MP1 was part of a PMOS stack that also included transistor MP3. However, transistor MP3 of level-shifter 100 is replaced by PMOS transistors MP3', MP3", and MP3 ¹ " in level-shifter 200. Similarly, transistor MP2 was part of a PMOS stack that also included transistor MP4. This transistor MP4 of level-shifter 100 is replaced by transistors MP4', MP4", and MP4"' in level-shifter 200.

[0024] With regard to the transistor MP1 PMOS stack, transistors MP3' and MP3" are in series between a power supply node providing VDDH and a source for transistor MP1. The source of transistor MP3 ¹ couples to a power supply node providing VDDH and its drain couples to a source for transistor MP3". The drain of transistor MP3" couples to a source of transistor MP 1. The input signal drives the gates of transistors MP3' and MP3" analogously to how the input signal drives the gate of transistor MP3 in level-shifter 100. Both transistors MP3 ¹ and MP3" are relatively weak in comparison to transistor MN1, which is relatively strong in comparison to transistors MP3' and MP3". By being coupled in series, transistors MP3' and MP3" are the equivalent of a single, even weaker PMOS transistor. Recall that one solution to extending the voltage range of level-shifter 100 was to weaken transistor MP3.

Transistors MP3' and MP3" thus function as this desirable weakening of transistor MP3 to prevent the node Nl discharge struggle discussed with regard to level-shifter 100. Despite this PMOS stack weakening, level-shifter 200 does not suffer the delays that would vex operation of level-shifter 100 if its transistor MPS were weakened to increase its input voltage range as will be explained further herein with regard to the operation of transistor MP3"\

[0025] Given the weakness of transistors MP3' and MP3", transistor MN1 in level-shifter 200 may thus discharge node Nl relatively quickly in response to the input signal transitioning high to VDDL ~ despite transistors MP3' and MP3" being switched weakly on. But level-shifter 200 does not suffer from timing delays due to the weakness of transistors MP3' and MP3" because of the relative strength of transistor MP3'", which is coupled in parallel to the series stack of transistors MP3' and MP3". The source of transistor MP3"' couples to a power supply node providing VDDH and its drain couples to the source of transistor MP1. A control signal CI drives the gate of transistor MP3'". A control circuit 205 discussed further below generates control signal CI responsive to binary transitions of the input signal such as sensed on nodes Nl and N2. Because transistor MP3'" functions to charge node Nl, it is also denoted herein as a pull-up transistor. In contrast, transistors MN1 and MN2 may be denoted as pulldown transistors since they function to discharge nodes N l and N2, respectively.

[0026] Because pull-up transistor MP3'" is relatively strong, control signal CI is generated by control circuit 205 so that pull-up transistor MP3'" is off when the input signal transitions high to VDDL. In that regard, control signal CI is a VDDH-domain signal such that pull-up transistor MP3"' is fully off when control signal CI is driven high to VDDH. In this fashion, pull-up transistor MP3'" does not struggle with transistor MN1 when transistor MN1 functions to pull node Nl to ground. Note that transistors MP3' and MP3" are weakly on when the input signal is high and transistor MP3'" is off. Transistors MP3' and MP3" function as weak keeper transistors that keep the drain of pull-up transistor MP3'" charged to VDDH. In this fashion, there is no leakage current through pull-up transistor MP3'" when transistor MP3'" is off. [0027] As will be explained further herein, control circuit 205 drives control signal CI low in response to the input signal transitioning high to VDDL. Pull-up transistor MP3'" then switches on so that it may then quickly charge node Nl to VDDH when cross-coupled transistor MPl is switched on responsive to a subsequent transition of the input signal low to ground. Cross-coupled transistor MPl is also relatively strong so that the serial combination or stack of pull-up transistor MP3"' and cross-coupled transistor MPl can quickly charge node Nl to VDDH. Control circuit 205 also drives control signal CI high to VDDH responsive to the input signal transitioning low. In this fashion, pull-up transistor MP3'" will be fully off and will not struggle with pulldown transistor MN1 at a subsequent high transition of the input signal.

[0028] Transistors MP4', MP4", and MP4'" function analogously. Transistor MP4'" may thus be denoted as a pull-up transistor whereas transistors MP4' and MP4" function as weak keeper transistors. The source of transistor MP4' couples to a power supply node for providing VDDH and its drain couples to a source of transistor MP4". The drain of transistor MP4" couples to a source for cross-coupled transistor MP2, The inverted version of the input signal from the inverter INV drives the gates of keeper transistors MP4' and MP4". When the input signal transitions low such that its inverted version transitions high to VDDL, keeper transistors MP4' and MP4" will be only weakly switched on. The serial combination of keeper transistors MP4' and MP4" is even weaker such that pull-down transistor MN2 can thus quickly discharge node N2 despite relatively-strong cross-coupled transistor MP2 being initially on at the transition of the input signal low since cross-coupled transistor MP2 is starved of charge due to the weakness of keeper transistors MP4' and MP4". Pull-up transistor MP4'" couples in parallel with the serial combination of keeper transistors MP4' and MP4". The source of pull-up transistor MP4'" thus couples to a power supply node providing VDDH whereas its drain couples to the source of cross-coupled transistor MP2. Control circuit 205 generates a control signal C2 that drives the gate of pull-up transistor MP4'" responsive to binary transitions of the input signal such as sensed on nodes Nl and N2.

[0029] Pull-up transistor MP4" ¹ is relatively strong in comparison to keeper transistors MP4' and MP4". Thus, control circuit 205 drives control signal C2 high to VDDH responsive to the input signal transitioning high to VDDL to prevent pull-up transistor MP4'" from struggling with pull-down transistor MN2 with regard to an eventual discharge of node N2 in response to a subsequent low transition of the input signal (which charges the inverted input signal to VDDL). While pull-up transistor MP4'" is off, keeper transistors MP4' and MP4" function as a weak keeper device that keeps the drain of pull-up transistor MP4'" charged to VDDH so as to eliminate leakage current across pull-up transistor MP4'" while it is off. Control circuit 205 drives control signal C2 low to switch on pull-up transistor MP4'" in response to the input signal transitioning low so that pull-up transistor MP4'" may assist in rapidly charging node N2 to VDDH when cross-coupled transistor MP2 is switched on at a subsequent high transition of the input signal. Cross-coupled transistor MP2 is also relatively strong so that the serial combination of cross-coupled transistor MP2 and pull-up transistor MP4'" may quickly charge node N2 to VDDH. In this fashion, level-shifter 200 has the advantage of weak PMOS stacks when discharging either of nodes Nl and N2 yet has strong PMOS stacks with regard to charging these nodes.

[0030] A level-shifter 300 of Figure 3 shows an example embodiment for control circuit 205. In particular, control circuit 205 comprises PMOS transistors MP5, MP6, MP7, and MP8 as well as NMOS transistors MN3 and MN4. The following discussion will also reference Figures 4 and 5. Figures 4 diagrams the response of node Nl, node N2, control signal C I, and control signal C2 in level-shifter 300 for a transition of the input signal from low (VSS) to high (VDDL). Conversely, Figure 5 illustrates the same signals for level-shifter 300 in response to a transition of the input signal from high to low. The structure of control circuit 205 will first be described followed by a discussion of its function with regard to these binary transitions of the input signal.

[0031] Node Nl couples to the gate of transistors MN4 and MP8. The source of transistor MN4 couples to ground and its drain couples to the drain of transistor MP8. The source of transistor MP8 couples to the drain of transistor MP5, which in turn has its source coupled to a power supply node providing VDDH and has its gate driven by control signal CI . The voltage at the drain of transistor MN4 (as well as the drain of transistor MP8) functions as control signal C2 that drives the gates of pull-up transistor MP4'" and transistor MP6. As will be explained further herein, transistor MN4 functions to discharge control signal C2 and thus is also designated herein as a pulldown transistor. Conversely, transistor MP8 functions to charge control signal C2 to VDDH and thus is designated herein as a pull-up transistor.

[0032] Transistors MP7 and MN3 have their gates driven by node N2. The source of transistor MN3 couples to ground and its drain couples to the drain of transistor MP7. In turn, the source of transistor MP7 couples to the drain of transistor MP6. The drain voltage for transistors MP7 and MN3 functions as control signal CI, which not only drives the gate of transistor MP5 but also drives the gate of pull-up transistor MP3"'. Analogous to transistor MN4, transistor MN3 is also designated herein as a pull-down transistor in that it functions to discharge control signal CI . Similarly, transistor MP7 is also designated herein as a pull-up transistor in that it functions to charge control signal CI to VDDH. Transistor MP6 is analogous to transistor MP5 in that transistor MP6 has its source tied to the VDDH power supply node and its gate is driven by control signal C2. Both transistors MP5 and MP6 function as switches that are controlled by control signals CI and C2, respectively. Node Nl couples to the gate of cross-coupled transistor MP2. Similarly, node N2 couples to the gate of cross-coupled transistor MP1.

[0033] Given this structure for control circuit 205, the transition of the input signal from low to VDDL will first be discussed as diagrammed in Figure 4. Prior to the high transition of the input signal, node Nl would have been charged high to VDDH. This high voltage on node Nl thus switches on pull-down transistor MN4 such that its drain and control signal C2 are grounded prior to the high transition of the input signal. The output signal (which is the same as the node N2 voltage) was low prior to the high transition of the input signal. Thus, pull-up transistor MP7 is on at that time. Since control signal C2 is also low prior to the high transition of the input signal, pull- up transistor MP7 and transistor MP6 are both on while the input signal is low, which charges control signal CI high. Thus, pull-up transistor MP3'" is off prior to the input signal transitioning to VDDL. Conversely, pull-up transistor MP4" ¹ is on prior to the high transition of the input signal.

[0034] When the input signal transitions high, the switching on of pull-down transistor MN1 readily discharges node Nl since pull-up transistor MP3'" is off. Cross- coupled transistor MP2 is then switched on and pull-down transistor MN2 switched off. Since cross-coupled transistor MP2 and pull-up transistor MP4'" are then both switched on, node N2 and the output signal are quickly charged to VDDH as node Nl is discharged. In turn, the charging of node N2 switches on pull-down transistor MN3 and switches off pull-up transistor MP7. Control signal CI is then discharged in anticipation of a subsequent transition of the input signal low as discussed below with regard to Figure 5. The discharge of node Nl switches on pull-up transistor MP8 and switches off pull-down transistor MN4. Pull-up transistor MP8 and transistor MP5 are then both conducting such that control signal C2 is driven high to VDDH, which shuts off pull-up transistor MP4"\ In this fashion, pull-up transistor MP4'" won't struggle with pull-down transistor MN2 for a subsequent low transition of the input signal.

[0035] When the input signal transitions low as diagrammed in Figure 5, pulldown transistor MN2 is switched on and pull-down transistor M l switched off. Since pull-up transistor MP4'" is off at this time, node N2 is quickly discharged to ground. Cross-coupled transistor MP1 then switches on. Since pull-up transistor MP3'" was on prior to the discharge of node N2, switched-on pull-up transistor MP3 ¹ " and cross- coupled transistor MP1 quickly charge node Nl to VDDH. The charging of node Nl switches on pull-down transistor MN4 and switches off pull-up transistor MP8 so that control signal C2 is discharged. Conversely, the discharge of node N2 switches on pull- up transistor MP7 and switches off pull-down transistor MN3. Transistor MP6 is switched on responsive to the discharge of control signal C2 so that the combination of transistor MP6 and pull-up MP7 then charge control signal CI high to VDDH. The high state of control signal CI then shuts off pull-up transistor MP3'".

[0036] When the input signal subsequently switches high to VDDL, pull-down transistor MNl can thus quickly discharge node Nl since this discharge is countered only by the weakly-switched-on combination of keeper transistors MP3' and MP3". In that regard, transistors MN3, MN4, MP5, MP6, MP7, and MP8 that comprise control circuit 205 are all relatively strong. The discharge of node Nl causes transistors MP2 and MP8 to switch on and also causes transistor MN4 to switch off. Transistors MP5 and MP8 then charge control signal C2 high to VDDH, which shuts off transistors MP4'" and MP6. [0037] After the low transition of the input signal, control circuit 205 thus operates to keep control signals CI and C2 in their previous state until node Nl is charged and node N2 discharged. Control circuit 205 then responds to the binary transitions of the node Nl and N2 voltages to flip the binary state of the control signals CI and C2. Control signal C2 is thus discharged so that node N2 can be quickly charged at the subsequent high transition of the input signal. Conversely, control signal CI is charged to VDDH so that pull-up transistor MP3'" will not struggle with pulldown transistor MN1 when pull-down transistor MN1 discharges node Nl at the subsequent transition of the input signal to VDDH.

[0038] Control circuit 205 operates in an analogous fashion during the high transition of the input signal. Control signals CI and C2 thus are momentarily maintained in their previous state until node Nl is discharged and node N2 is charged. After the discharge of node Nl, control circuit 205 charges control signal C2 so that pull-down transistor MP4'" is then turned off. But transistor MP4'" was on long enough after the high transition of the input signal to quickly charge node N2 to VDDH.

Control circuit 205 then responds to the charging of node N2 by discharging control signal CI so that pull-up transistor MP3'" is then turned on so that it may assist in a charging of node l responsive to a subsequent low transition of the input signal. But this switching on of pull-up transistor MP3'" occurs after node Nl is already discharged so that there is no struggle with regard to the discharge of node Nl. The net result is that level-shifter 300 of Figure 3 achieves the conflicting goals of having weak PMOS stacks with regard to the discharging of nodes Nl and N2 yet also having strong PMOS stacks with regard to the charging of nodes Nl and N2. In this fashion, level-shifter 300 may have a wide input signal voltage range such that the VDDL may be dropped to low levels (for example, .6 V or less) yet high-speed and accurate operation is achieved. Moreover, this performance may be guaranteed across all expected process corners.

[0039] A method of operation for level-shifter 300 will now be discussed.

Figure 6 is a flowchart of an example method of operation. A step 600 comprises discharging an output signal responsive to an input signal transitioning from a first power supply voltage to ground. The discharging of the output signal as discussed with regard to Figure 5 is an example of such an act. A step 605 comprises switching on a first pull-up transistor responsive to the input signal transitioning from the first power supply voltage to ground. The switching on of pull-up transistor MP4 ¹ " after the input signal transitions low is an example of such an act. A step 610 comprises using the switched-on first pull-up transistor to charge the output signal to a second power supply voltage responsive to the input signal transitioning from ground to the first power supply voltage. The discussion above with regard to pull-up transistor MP4'" charging the output signal to VDDH (in conjunction with cross-coupled transistor MP2) is an example of such an act. Finally, the method includes a step 615 of, after the charging of the output signal, switching off the first pull-up transistor such that the discharging of the output signal responsive to the input signal transitioning to ground is not opposed by the first pull-up transistor. The charging of control signal C2 through pull-up transistor MP8 in response to the discharging of node Nl so as to switch off pull-up transistor MP4'" as discussed with regard to level-shifter 300 is an example of such an act.

[0040] As those of some skill in this art will by now appreciate and depending on the particular application at hand, many modifications, substitutions and variations can be made in and to the materials, apparatus, configurations and methods of use of the devices of the present disclosure without departing from the spirit and scope thereof. In light of this, the scope of the present disclosure should not be limited to that of the particular embodiments illustrated and described herein, as they are merely by way of some examples thereof, but rather, should be fully commensurate with that of the claims appended hereafter and their functional equivalents.