Integrated circuit charge pumps having control circuits therein that inhibit parasitic charge injection from control signals

The present invention relates to integrated circuit devices and, more particularly, to integrated circuit driver devices and charge pumps.

Phase-locked loop (PLL) circuits are frequently used to generate highly accurate internal clock signals on an integrated circuit substrate. Typical PLL circuits include charge pumps that selectively source current to and sink current from a loop filter which controls an oscillator stage of the PLL circuit. This oscillator stage may be a voltage-controlled oscillator (VCO) stage. The amount of charge (Q) provided by the sourcing or sinking currents may be proportional to the duration of an active period of a control signal provided to the charge pump.

One example of a conventional charge pump that may be used in a PLL circuit is described in U.S. Pat. No. 6,124,741 to Arcus. In particular, FIGS. 3-6 and 9 of the '741 patent illustrate various charge pumps that include current sources defined by a plurality of MOS transistors. These MOS transistor are arranged as a totem pole of upper PMOS and lower NMOS transistors that are connected in series (source-to-drain) between a power supply line (Vdd) and a ground reference line (Vss).

Control circuitry is also provided to drive the MOS transistors in a manner that matches the up/down current pumping by reducing Vds variations. PLL circuits are also described in U.S. Pat. No. 6,222,895 to Larsson and in published U.S. application Ser. No. 2001/0052806 A1 to Gu et al. U.S. Pat. No. 6,388,499 to Tien et al., assigned to the present assignee, also discloses a totem pole arrangement of MOS transistors that can be used in a signal driver.

Integrated circuit charge pumps according to embodiments of the present invention reduce parasitic charge injection from signals that drive control inputs of the charge pumps. This reduction in parasitic charge injection can be utilized to lower phase error when the charge pumps are used in phase-locked loops (PLLs). In some embodiments, an integrated circuit charge pump includes an output current source and a control circuit that drives the output current source with control signals in a preferred manner. The output current source may include a totem pole driver therein. This totem pole driver includes at least an upper PMOS supply transistor and a lower PMOS current source transistor that are disposed in series in a pull-up path extending between an output of the driver and a power supply line (e.g., Vdd). The totem pole driver also includes at least an upper NMOS current source transistor and a lower NMOS supply transistor. These NMOS transistors are disposed in series in a pull-down path extending between the output of the device and a reference line (e.g., Vss). The control circuit may include a pull-up control circuit and a pull-down control circuit. The pull-up control circuit is configured to drive a gate of the upper PMOS supply transistor with a PMOS turn-on voltage in response to a leading edge of a pull-up control signal. The pull-up control circuit is further configured to simultaneously drive a gate of the upper PMOS supply transistor with a PMOS turn-off voltage and a source of the lower PMOS current source transistor with a first non-zero turn-off reference voltage (e.g., PVT), in response to a trailing edge of the pull-up control signal.

In some additional embodiments, the pull-up control circuit includes a first output that is electrically coupled to the gate of the upper PMOS supply transistor and a second output that is electrically coupled to a source of the lower PMOS current source transistor. This second output of the pull-up control circuit is preferably disposed in a high-impedance state in response to the leading edge of the pull-up control signal. In the event the pull-up path includes only two PMOS transistors, then the second output of the pull-up control circuit may be electrically coupled to a source of the lower PMOS current source transistor and a drain of the upper PMOS supply transistor. A gate of the lower PMOS current source transistor may also be held at a first bias voltage (e.g., VP 1 ).

The pull-down control circuit is preferably configured to drive a gate of the lower NMOS supply transistor with an NMOS turn-on voltage in response to a leading edge of a pull-down control signal. The pull-down control circuit is further configured to simultaneously drive a gate of the lower NMOS supply transistor with an NMOS turn-off voltage and a source of the upper NMOS current source transistor with a second turn-off reference voltage (e.g., NVT<Vdd) in response to a trailing edge of the pull-down control signal. Moreover, the pull-down control circuit may include a first output that is electrically coupled to the gate of the lower NMOS supply transistor and a second output that is electrically coupled to a source of the upper NMOS current source transistor. This second output of the pull-down control circuit is preferably disposed in a high-impedance state in response to the leading edge of the pull-down control signal. In the event the pull-down path includes only two NMOS transistors, then the second output may also be electrically coupled to a drain of the lower NMOS supply transistor. A gate of the upper NMOS current source transistor may also be held at a second bias voltage (e.g., VN 1 ).

According to preferred aspects of these embodiments of the present invention, the pull-up control circuit includes a PMOS access transistor. This PMOS access transistor has a first current carrying terminal (e.g., source) that is electrically coupled to the source of the lower PMOS current source transistor and a second current carrying terminal (e.g., drain) that is held at the first turn-off reference voltage. The PMOS access transistor is preferably sized to match the upper PMOS supply transistor in width and length, to reduce switching noise and charge injection caused by capacitive coupling in the pull-up path. The pull-down control circuit also includes an NMOS access transistor. This NMOS access transistor has a first current carrying terminal (e.g., drain) electrically coupled to the source of the upper NMOS current source transistor and a second current carrying terminal (e.g., source) that is held at the second turn-off reference voltage. The NMOS access transistor is preferably sized to match the lower NMOS supply transistor in width and length, to reduce switching noise and charge injection caused by capacitive coupling in the pull-down path.

According to still further preferred aspects of these embodiments, the pull-up control circuit is configured to include first and second delay-balanced inverter strings that drive the gate of the upper PMOS supply transistor and a gate of the PMOS access transistor, respectively, in response to the pull-up control signal. These delay-balanced inverter strings may include an uneven number of inverters, with one or more delay elements providing an equivalent inverter delay. The pull-down control circuit is also configured to include a pair of delay-balanced inverter strings that drive the gate of the lower NMOS supply transistor and a gate of the NMOS access transistor in response to the pull-down control signal.

FIG. 1 is a block diagram of a differential charge pump according to an embodiment of the present invention.

FIG. 2 is an electrical schematic of a charge pump according to another embodiment of the present invention.

FIG. 3 are timing diagrams that illustrate operation of the charge pump of FIG. 2 when a power supply voltage (Vdd) of 2.5 volts is used.

The present invention now will be described more fully herein with reference to the accompanying drawings, in which preferred embodiments of the invention are shown. This invention may, however, be embodied in many different forms and should not be construed as being limited to the embodiments set forth herein; rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the invention to those skilled in the art. Like reference numerals refer to like elements throughout and signal lines and signals thereon may be referred to by the same reference characters.

Referring now to FIG. 1, a differential charge pump 10 according to an embodiment of the present invention includes a pair of charge pumps 12 a and 12 b that generate a pair of complementary output signals OUT 1 and OUT 2 in response to a pair of primary control signals UP and DOWN and a plurality of reference signals, shown as PVT, VP 1 , VN 1 and NVT, which may be fixed-voltage reference signals. As described more fully hereinbelow, the reference signal PVT may be set to a first turn-off reference voltage and the reference signal NVT may be set to a second turn-off reference voltage. The reference signal VP 1 may be set to a first bias voltage and the reference signal VN 1 may be set to a second bias voltage.

The primary control signals UP and DOWN are illustrated as active high signals that are provided to a pair of inverters 16 a and 16 b . The primary control signals UP and DOWN may be made active during non-overlapping time intervals. Alternatively, to prevent dead-band conditions when a phase error in a PLL is small, the primary control signals UP and DOWN may be made to overlap. As explained more fully hereinbelow with respect to FIG. 2, when the primary control signal UP is switched to an active high level, the input DOWNB of the first charge pump 12 a and the input UPB of the second charge pump 12 b are switched to active low levels. At these active low levels, the first charge pump 12 a sinks current from the first output signal line OUT 1 and the second charge pump 12 b sources current to the second output signal line OUT 2 . The duration of the active period of the primary control signal UP controls the amount of “sinking” charge (Q down1 ) received by the first charge pump 12 a from the first output signal line OUT 1 and the amount of “sourcing” charge (Q up2 ) supplied by the second charge pump 12 b to the second output signal line OUT 2 . The pull-down path of the first charge pump 12 a and the pull-up path of the second charge pump 12 b may be matched so that the magnitude of the sinking charge equals the magnitude of the sourcing charge (i.e., Q down1 =Q up2 ).

Similarly, when the primary control signal DOWN is switched to an active high level, the input DOWNB of the second charge pump 12 b and the input UPB of the first charge pump 12 b are switched to active low levels. At these active low levels, the second charge pump 12 b sinks current from the second output signal line OUT 2 and the first charge pump 12 a sources current to the first output signal line OUT 1 . The duration of the active period of the primary control signal DOWN controls the amount of “sinking” charge (Q down2 ) received by the second charge pump 12 b and the amount of “sourcing” charge (Q up1 ) supplied by the first charge pump 12 a . The pull-up path of the first charge pump 12 a and the pull-down path of the second charge pump 12 b may be matched so that the magnitude of the sinking charge equals the magnitude of the sourcing charge (i.e., Q down2 =Q up1 ).

The charge pump 12 of FIG. 2 includes a pull-up control circuit 14 a that is responsive to an active low pull-up control signal UPB and a pull-down control circuit 14 b that is responsive to an active low pull-down control signal DOWNB. The pull-up control circuit 14 a also receives reference signal PVT, which is held at a first turn-off reference voltage, and reference signal VP 1 , which is held at a first bias voltage. The pull-down control circuit 14 b also receives reference signal NVT, which is held at a second turn-off reference voltage, and reference signal VN 1 , which is held at a second bias voltage. Typical |NVT| and |PVT| values are 900 mV with respect to the source voltage of the corresponding supply FET and typical |VN 1 | and |VP 1 | values are 800 mV with respect to the source voltage of the corresponding supply FET. Thus, for a power supply voltage (Vdd) of 2.5 Volts, the first turn-off reference voltage (PVT) may be set to about 1.6 Volts and the first bias voltage (VP 1 ) may be set to about 1.7 Volts. Likewise, the second turn-off reference voltage (NVT) may be set to about 0.9 Volts and the second bias voltage (VN 1 ) may be set to about 0.8 Volts.

The charge pump 12 also includes an output current source. This output current source is illustrated as including a totem pole driver. This driver is illustrated as including a plurality of PMOS pull-up transistors and a plurality of NMOS pull-down transistors. In some embodiments, two PMOS pull-up transistors are provided in a pull-up path and two NMOS pull-down transistors are provided in a pull-down path. A greater number of transistors may also be provided in each pull-up and pull-down path. The pull-up path extends between an output OUTn of the driver and a power supply line (Vdd) and the pull-down path extends between the output OUTn and a reference line (Vss). The PMOS transistors include at least an upper PMOS supply transistor P 2 and a lower PMOS current source transistor P 1 . The NMOS transistors include at least an upper NMOS current source transistor N 1 and a lower NMOS supply transistor N 2 .

According to a preferred aspect of the charge pump 12 , the pull-up control circuit 14 a is configured to drive a gate of the upper PMOS supply transistor P 2 with a PMOS turn-on voltage (i.e., logic 0 voltage), in response to a leading edge (e.g., falling edge) of the pull-up control signal UPB. The pull-up control circuit 14 a is also configured to simultaneously drive a gate of the upper PMOS supply transistor P 2 with a PMOS turn-off voltage (i.e., logic 1 voltage) and a source of the lower PMOS current source transistor P 1 with the first turn-off reference voltage PVT in response to a trailing edge (e.g., rising edge) of the pull-up control signal UPB. The pull-up control circuit 14 a includes a first output that is electrically coupled to the gate of the upper PMOS supply transistor P 2 by signal line PS 1 . The pull-up control circuit 14 a also includes a second output that is electrically coupled to a source of the lower PMOS current source P 1 and a drain of the upper PMOS supply transistor P 2 by signal line SP.

The pull-up control circuit 14 a may also include a PMOS access transistor P 3 having one current carrying terminal (e.g, source) that is connected to the signal line SP and another current carrying terminal (e.g., drain) that receives the first turn-off reference voltage PVT. The pull-up control circuit 14 a includes first and second delay-balanced inverter strings that are configured to drive the gate of the upper PMOS supply transistor P 2 at signal line PS 1 and a gate of the PMOS access transistor P 3 at signal line PS 2 , respectively, in response to the pull-up control signal UPB. The first inverter string is illustrated as including a normally-on transmission gate TG 1 and inverters I 1 and I 2 . The second inverter string is illustrated as including an odd number of inverters I 3 , I 4 and I 5 . The transmission gate TG 1 and the inverters I 1 -I 5 are configured to provide equivalent signal delays. Accordingly, a leading edge of the active low pull-up control signal UPB (i.e., high-to-low edge) will simultaneously cause the upper PMOS supply transistor P 2 to turn on and the PMOS access transistor P 3 to turn off and thereby isolate the source of the lower PMOS current source P 1 from the first turn-off reference voltage PVT.

According to a preferred aspect of the pull-up control circuit 14 a , the PMOS access transistor P 3 matches the upper PMOS supply transistor P 2 in width and length. In some embodiments, this matching between the PMOS transistors P 2 and P 3 may be present in those situations where the leading and trailing edges of the complementary output signals generated on signal lines PS 1 and PS 2 are matched and a finite source-to-well voltage of the PMOS access transistor P 3 can be ignored. The matching of the PMOS transistors P 2 and P 3 may also be present in other situations.

This matching of the sizes of the PMOS transistors P 2 and P 3 causes the gate-to-source capacitance of the PMOS access transistor P 3 to match the gate-to-drain capacitance of the upper PMOS supply transistor P 2 . In this manner, the capacitive couplings between signal lines PS 1 and SP and between signal lines PS 2 and SP (and the source of the lower PMOS current source transistor P 1 ) automatically cancel each other when the signal lines PS 1 and PS 2 are simultaneously switched to opposite logic levels by the first and second delay-balanced inverter strings. This cancellation of the capacitive coupling to the source terminal of the lower PMOS current source transistor P 1 operates to minimize the signal swing at this source terminal when the pull-up path is switched on and off and thereby inhibit charge injection from the pull-up control signal UPB to the source of the lower PMOS current source transistor P 1 . In addition, the magnitude of the first turn-off reference voltage PVT can be carefully controlled to optimize pull-up performance. However, in alternative embodiments, the PMOS access transistor P 3 and the PMOS supply transistor P 2 may not match each other in size. Such intentional mismatch may be advantageous to optimize performance for a particular fabrication process, for example.

The pull-down control circuit 14 b is configured to drive a gate of the lower NMOS supply transistor N 2 with an NMOS turn-on voltage (e.g., logic 1 voltage), in response to a leading edge of the pull-down control signal DOWNB. The pull-down control circuit 14 b is also configured to simultaneously drive a gate of the lower NMOS supply transistor N 2 with an NMOS turn-off voltage and a source of the upper NMOS current source transistor N 1 with the second turn-off reference voltage NVT in response to a trailing edge of the pull-down control signal DOWNB. The pull-down control circuit 14 b includes a first output that is electrically coupled to the gate of the lower NMOS supply transistor N 2 by signal line NS 1 . The pull-down control circuit 14 b also includes a second output that is electrically coupled to a source of the upper NMOS current source N 1 and a drain of the lower NMOS supply transistor N 2 by signal line SN.

The pull-down control circuit 14 b also includes an NMOS access transistor N 3 that may, in some embodiments, match the lower NMOS supply transistor N 2 in width and length, as described above. The pull-down control circuit 14 b includes first and second delay-balanced inverter strings that are configured to drive the gate of the lower NMOS supply transistor N 2 and a gate of the NMOS access transistor N 3 at signal line NS 2 , respectively, in response to the pull-down control signal DOWNB. The first inverter string is illustrated as including a normally-on transmission gate TG 2 and inverters I 6 and I 7 . The second inverter string is illustrated as including an odd number of inverters I 8 , I 9 and I 10 . The transmission gate TG 2 and the inverters I 6 -I 10 are configured to provide equivalent signal delays. Accordingly, a leading edge of the active low pull-down control signal DOWNB (i.e., high-to-low edge) will simultaneously cause the lower NMOS supply transistor N 2 to turn on and the NMOS access transistor N 3 to turn off and thereby isolate the source of the upper NMOS current source N 1 from the second turn-off reference voltage NVT.

According to a preferred aspect of the pull-down control circuit 14 b , the NMOS access transistor N 3 matches the lower NMOS supply transistor N 2 in width and length. This matching of the sizes of the NMOS transistors N 2 and N 3 causes the gate-to-drain capacitance of the NMOS access transistor N 3 to match the gate-to-drain capacitance of the lower NMOS supply transistor N 2 . In this manner, the capacitive couplings between signal lines NS 1 and SN and between signal lines NS 2 and SN (and the source of the upper NMOS current source transistor N 1 ) automatically cancel each other when the signal lines NS 1 and NS 2 are simultaneously switched to opposite logic levels by the first and second delay-balanced inverter strings within the pull-down control circuit 14 b . This cancellation of the capacitive coupling to the source terminal of the upper NMOS current source transistor N 1 operates to minimize the signal swing at this source terminal when the pull-down path is switched on and off and thereby inhibit charge injection from the pull-down control signal DOWNB to the source of the upper NMOS current source transistor N 1 . In addition, the magnitude of the second turn-off reference voltage NVT can be carefully controlled to optimize pull-down performance. In alternative embodiments, the NMOS access transistor N 3 and the NMOS supply transistor N 2 may not match each other in size. Such intentional mismatch may be advantageous to optimize performance for a particular fabrication process, for example.

Referring now to FIG. 3, a plurality of timing diagrams are provided that illustrate operation of the charge pump 12 of FIG. 2 at a power supply voltage of 2.5 Volts. The voltage levels for signals PVT, NVT, VP 1 and NV 1 are 1.6, 0.9, 1.7 and 0.8 Volts, respectively. In particular, the first timing diagram illustrates the switching of the active low pull-up control signal UPB and the active low pull-down control signal DOWNB from a high voltage level equal to Vdd to a low voltage level of Vss. The second timing diagram illustrates the synchronous switching of signal lines NS 1 and NS 2 in response to the pull-down control signal DOWNB. The third timing diagram illustrates the synchronous switching of signal lines PS 1 and PS 2 in response to the pull-up control signal UPB. The fourth timing diagram illustrates the upward and downward movement of node SP (source of PMOS transistor P 1 ) from between 1.6 Volts and 2.5 volts. This movement of node SP is responsive to the switching of signal lines PS 1 and PS 2 . The fifth timing diagram illustrates the upward and downward movement of node SN (source of NMOS transistor N 1 ) from between 0.9 Volts and 0 volts. This movement of node SN is responsive to the switching of signal lines NS 1 and NS 2 .

In the drawings and specification, there have been disclosed typical preferred embodiments of the invention and, although specific terms are employed, they are used in a generic and descriptive sense only and not for purposes of limitation, the scope of the invention being set forth in the following claims.