The present invention relates generally to electronic circuits, and in particular to charge pump circuits for use in phase locked loops, and the like.
In electronic systems, good clock distribution can be very important to the overall performance of the product. Unwanted clock skew and jitter are two phenomena that may result from poor clock distribution, thus causing problems in the design and operation of the electronic system. Techniques have been developed using phase lock loops (PLLs) to successfully manage these problems and reduce both to manageable levels. However, current conventional solutions are not without some disadvantages.
Conventional PLL solutions can have certain advantages, including the fact that PLLs can suppress skew in digital systems (for example, clock to data out delay), can generate multiple phases of output clocks, and can be used to multiply or divide clock signals. In an exemplary application, a PLL may operate as a clock multiplier, where an input clock of 10 MHz may be multiplied by the PLL to yield 1000 Mhz. In an ideal embodiment, a clock multiplication function can result in an output CLK that is in perfect phase alignment with the input CLK.
A common implementation of a conventional PLL is that of a charge-pump based PLL. In this topology, a phase-frequency detector (PFD) is used to determine the relative phase and frequency of the PLL output and a reference clock input. The outputs of the PFD are coupled to a charge pump. The charge pump is used to add or subtract charge from a loop filter in response to PFD output signals by switching on or off electric currents of predetermined values for a duration determined by the PFD signals. The loop filter integrates charge from the charge pump, converting it to a voltage input to a voltage controlled oscillator (VCO). The VCO converts the voltage into the PLL output frequency. This output frequency can be connected back to the PFD input in some fashion.
Inputs received by a charge pump from a PFD can instruct a PLL to go “up” or “down” in frequency. These “up” and “down” signals have a duration that is proportional to the phase difference of the reference frequency and PLL output frequency, particularly when the two frequencies become very close. This phase difference is referred to as “phase error”. When a PLL is locked in both phase and frequency, this phase error is driven to a minimum that is determined by natural mismatches in the PFD, charge pump and other PLL circuitry, along with the precision with which the up and down signals are switched. This minimum phase error is referred to as static phase error.
To better understand various features of the disclosed embodiments, a conventional charge pump will be described with reference to
Referring still to
Source current source 606 can source current (i.e., provide current) to an output node 614, while sink current source 608 can sink current (i.e., draw current) from output node 614, to thereby generate an output current lout.
Conventional charge pump 600 can operate in the following manner: When input signal Dn is high or input signal UpM is low, current flows either to or from output node 614. More particularly, current should only flow from output node 614 to ground (when signal Dn high) or from VCC to output node 614 (when signal UpM low). Current sources (606 and 608) can be a single transistor or several transistors (e.g., a cascade configuration). A buffer (610 or 612) may operate as a delay to improve transistor switching and minimize charge injection. The upper n-type transistor N60 and lower p-type transistor P62 can be present to effect a precise turn off of charge pump 600, which can minimize PLL static phase error.
Disadvantages of the conventional solution can include that when Dn is low and UpM is high, the charge pump is ideally off. However, in a conventional case like that of
In addition, an unintended reverse current through current sources (such as biased current mirrors), can increase as the geometry of the devices used in such circuits decreases, as is the case in advanced manufacturing processes. This can be true even when the charge pump devices are operating in the sub-threshold region of operation. For smaller geometry devices, sub-threshold currents in the weak inversion region of operation are typically much higher. This can further reduce the usable range of loop filter voltage. Compounding this problem, small geometry processes must also be operated at lower voltage ranges in order to protect these small devices from overstress.
Various embodiments of the present invention will now be described in detail with reference to a number of drawings. The embodiments show charge pump circuits and methods that can greatly reduce if not essentially eliminate reverse currents through current sources.
A charge pump circuit according to a first embodiment is shown in a block schematic diagram in
A source section 102 can include a source enable switch 110, a first blocking switch 112, a source disable switch 114, and a source timing circuit 116. A source enable switch 110 can include a controllable impedance path connected between a high power supply node 118 and a source output node 120 that can be enabled (e.g., provide a low impedance) or disabled (e.g., provide a high impedance), in response to a signal EN1.
A first blocking switch 112 and source disable switch 114 can be connected in series with one another between source output node 120 and a low power supply node 122. First blocking switch 112 can be enabled or disabled in response to a signal BLK1. Source disable switch 114 can be enabled or disabled in response to a signal DIS1.
A source timing circuit 116 can provide signals EN1, BLK1 and DIS1 in response to an input signal UP received at a first input node 124. A source timing circuit 116 can be conceptualized as including timing paths 116-0 to 116-2 between first input signal node 124 and source enable switch 110, first blocking switch 112, and source disable switch 114, respectively. In response to a transition in signal UP, any of switches 102, 104, 106 can be disabled or enabled.
In this way, a source section 102 can either source current to a source output node 120, or disable such a current sourcing operation.
A sink section 104 can have the same general construction as a source section 102, but with respect to a low power supply node 122. Thus, a sink section 104 can include a sink enable switch 126, a second blocking switch 128, a sink disable switch 130, and a sink timing circuit 132. A sink enable switch 126 can be connected between low power supply node 122 and a sink output node 134, and second blocking switch 128 and sink disable switch 130 can be connected in series with one another between sink output node 134 and high power supply node 118.
A sink timing circuit 132 can provide signals EN2, BLK2 and DIS2 in response to an input signal DN received at a second input node 136. Sink timing circuit 132 can be conceptualized as including timing paths 132-0 to 132-2 between second input signal node 136 and switches 126, 128 and 130. Thus, in response to a transition in signal DN, any of switches 126, 128, 130 can be disabled or enabled.
In this way, a sink section 104 can either sink current from sink output node 134, or disable such a current sinking operation.
A source current source 106 can be connected to provide a current path from source output node 120 to a control output node 138. A sink current source 108 can be connected to provide a current path from control output node 138 to sink output node 134.
A disable operation for charge pump circuit 100 of
Referring now to
At time t0, or thereafter, the current paths for both a blocking switch and the corresponding disable switch (IBLK and IDIS) can be enabled. Thus, in the case of a source section 102, a source output node 120 can be connected to a low power supply node 122, while in the case of sink section 104, a sink output node 134 can be connected to a high power supply node 118. In this way, an output node (i.e., 120 or 134) can be “pulsed” to a disable voltage.
At time t1, a current path for a blocking switch (IBLK) can be disabled, while a current path for the corresponding disable current path (IDIS) can remain enabled. This can result in a blocking device (i.e., 112 or 128) blocking current flow in either direction (reverse or otherwise).
Note that unlike a conventional case, like that of
In this way, a charge pump circuit can prevent a reverse current through current source and sink paths when such paths are disabled.
Referring now to
In the particular arrangement of
Like the source timing circuit 216, a sink timing circuit 232 can include a second buffer B22 and a second inverter I22. In such an arrangement, a first timing path 232-0′ can include buffer B22, a second timing path 232-1′ can include buffer B22 and inverter 122, and a third timing path 232-2′ can include a wiring path. An input signal Dn can be connected to an input of second buffer B22. An output of second buffer B22 can be connected to an input of second inverter INV22 and to a gate of n-channel transistor 226′. The output of second inverter INV22 can be connected to a gate of p-channel transistor 228′. An input signal Dn can be further connected to a gate of p-channel transistor 230′. Transistors 226′, 228′ and 230′ can form another stack from a high power supply (VCC) to a low power supply (VSS). Sink output node 234 between transistors 228′ and 226′ can be connected to sink current source 208.
The embodiment of
Selected operations of the circuit 200 will now be described with reference to
First, a circuit 200 can generate a pulse on a deactivating edge of an input signal (Up or Dn) by applying normal and time delayed inverted signals to series connected transistors (212′/214′ or 228′/230′). In particular, at time t0, signals Up_BUF and Up_INV can both be high, pulsing source output node 210 to a low power supply level (VSS). Also, at time t2, signals Dn_BUF and Dn_INV can both be low, pulsing sink output node 234 to a high power supply level (VCC). Activating edges/levels of input signals Up and Dn can remain unchanged, as compared to conventional approaches.
Second, such a pulse generation circuit maintains proper timing during switching (Up and Dn charge/discharge correctly and charge injection minimized).
Third, signals Up and Dn can cause a tri-stating when the charge pump circuit is off (after a short delay). In particular, at time t1, signal Up_INV can transition low, turning off transistor 212′. Also, at time t3, signal Dn_INV can transition high, turning off transistor 228′. Such a turning off of these transistors can result in the essential elimination of reverse current paths, extending the operating range of voltage on a control output node 238. This can offer a significant advantage as manufacturing processes scale to lower operating voltages and smaller device sizes, which can lead to higher leakage in devices.
It is understood that the particular timing arrangement of
However, alternate embodiments can include different timing arrangements.
While the embodiments have shown charge pump circuits, other embodiments can include such charge pump circuits included in larger circuits for advantageous performance. One particular example of such an embodiment is shown in
A PFD 402 can be a conventional circuit designed to generate two control signals (UP and DN) for indicating a difference in phase/frequency between an input signal Fin and a feedback signal Ffdbk. A charge pump 404 can take the form of any of the above embodiments, or equivalents, and can source current to, sink current from, or provide a high impedance (i.e., little or no reverse current) at output node lout. A loop filter 406 can provide a control voltage Vctrl based on currents (e.g., either sourced or sunk) provided by charge pump 404. A VCO 408 can provide an output signal Fout that oscillates at a frequency according to control voltage Vctrl.
An input frequency multiplier/divider 410 can be provided to frequency multiply or divide an input signal Fin. Similarly, a feedback multiplier/divider 412 can frequency multiple or divide an output signal Fout to generate a feedback signal Ffdbk. Of course, such multiplier/dividers are optional depending upon application.
In this way, a PLL can include a charge pump circuit that has little or no reverse current draw, thus providing a wider usable control voltage range and reduction in long term jitter.
Referring now to
Of course, a PLL circuit could include other types of loop filters.
Advantages of the various embodiments can be that a reverse current path for a charge pump, like that shown in
While the above embodiments have shown arrangements that might imply a single ended architecture (e.g., single ended PLL), the present invention can be included in differential architectures.
Of course, embodiments can include any combination of the individually mentioned circuits, whether combining all elements or only some, not limited to any specific combination or set of combinations. For example, it may be possible to have more than one PFD. In addition, a pre-charge circuit can be included, or a PLL can have a second VCO. It is also possible to use a lock detector to power down the fast lock circuitry once phase lock has been achieved. Further, other embodiments can include a phase detector in lieu of a PFD.
Embodiments of the present invention are well suited to performing various other steps or variations of the steps recited herein, and in a sequence other than that depicted and/or described herein. In one embodiment, such a process is carried out by processors and other electrical and electronic components, e.g., executing computer readable and computer executable instructions comprising code contained in a computer usable medium.
For purposes of clarity, many of the details of the improved solution and the methods of designing and manufacturing the same that are widely known and are not relevant to the present invention have been omitted from the following description.
It should be appreciated that reference throughout this specification to “one embodiment” or “an embodiment” means that a particular feature, structure or characteristic described in connection with the embodiment is included in at least one embodiment of the present invention. Therefore, it is emphasized and should be appreciated that two or more references to “an embodiment” or “one embodiment” or “an alternative embodiment” in various portions of this specification are not necessarily all referring to the same embodiment. Furthermore, the particular features, structures or characteristics may be combined as suitable in one or more embodiments of the invention.
Similarly, it should be appreciated that in the foregoing description of exemplary embodiments of the invention, various features of the invention are sometimes grouped together in a single embodiment, figure, or description thereof for the purpose of streamlining the disclosure aiding in the understanding of one or more of the various inventive aspects. This method of disclosure, however, is not to be interpreted as reflecting an intention that the claimed invention requires more features than are expressly recited in each claim. Rather, as the following claims reflect, inventive aspects lie in less than all features of a single foregoing disclosed embodiment. Thus, the claims following the detailed description are hereby expressly incorporated into this detailed description, with each claim standing on its own as a separate embodiment of this invention.
This application claims the benefit of U.S. Provisional Patent Application Ser. No. 60/764,589, filed on Feb. 1, 2006, the contents of which are incorporated by reference herein.
Number | Name | Date | Kind |
---|---|---|---|
4287480 | Swift et al. | Sep 1981 | A |
4388596 | Yamashita | Jun 1983 | A |
4668922 | Crawford et al. | May 1987 | A |
4692718 | Rosa et al. | Sep 1987 | A |
4812783 | Honjo et al. | Mar 1989 | A |
4814726 | Byrd et al. | Mar 1989 | A |
4817199 | Wallraff | Mar 1989 | A |
4884042 | Menon et al. | Nov 1989 | A |
4949051 | Viola | Aug 1990 | A |
4972442 | Steierman | Nov 1990 | A |
4980899 | Troost et al. | Dec 1990 | A |
5008635 | Hanke et al. | Apr 1991 | A |
5059925 | Weisbloom | Oct 1991 | A |
5075639 | Yaya | Dec 1991 | A |
5101117 | Johnson et al. | Mar 1992 | A |
5119043 | Brown et al. | Jun 1992 | A |
5121086 | Srivastava | Jun 1992 | A |
5126690 | Bui et al. | Jun 1992 | A |
5157355 | Shikakura et al. | Oct 1992 | A |
5164685 | Niemio | Nov 1992 | A |
5194767 | Chao | Mar 1993 | A |
5194828 | Kato et al. | Mar 1993 | A |
5239455 | Fobbester et al. | Aug 1993 | A |
5256989 | Parker et al. | Oct 1993 | A |
5329252 | Major | Jul 1994 | A |
5331295 | Jelinek et al. | Jul 1994 | A |
5332930 | Volk | Jul 1994 | A |
5343167 | Masumoto et al. | Aug 1994 | A |
5357216 | Nguyen | Oct 1994 | A |
5363066 | Chen et al. | Nov 1994 | A |
5375148 | Parker et al. | Dec 1994 | A |
5412349 | Young et al. | May 1995 | A |
5424689 | Gillig et al. | Jun 1995 | A |
5444744 | Yamamoto et al. | Aug 1995 | A |
5446867 | Young et al. | Aug 1995 | A |
5455840 | Nakauchi et al. | Oct 1995 | A |
5457428 | Alder et al. | Oct 1995 | A |
5485125 | Dufour | Jan 1996 | A |
5495207 | Novof | Feb 1996 | A |
5525932 | Kelkar et al. | Jun 1996 | A |
5550493 | Miyanishi | Aug 1996 | A |
5550515 | Liang et al. | Aug 1996 | A |
5576647 | Sutardja et al. | Nov 1996 | A |
5596614 | Ledda | Jan 1997 | A |
5619161 | Novof et al. | Apr 1997 | A |
5625306 | Tada | Apr 1997 | A |
5631591 | Bar-Niv | May 1997 | A |
5657359 | Sakae et al. | Aug 1997 | A |
5663686 | Tada | Sep 1997 | A |
5684844 | Bouzidi et al. | Nov 1997 | A |
5699016 | Federspiel et al. | Dec 1997 | A |
5703511 | Okamoto | Dec 1997 | A |
5724007 | Mar | Mar 1998 | A |
5734279 | Bereza | Mar 1998 | A |
5736880 | Bruccoleri et al. | Apr 1998 | A |
5737589 | Doi et al. | Apr 1998 | A |
5740213 | Dreyer | Apr 1998 | A |
5801578 | Bereza | Sep 1998 | A |
5815042 | Chow et al. | Sep 1998 | A |
5825640 | Quigley et al. | Oct 1998 | A |
5898336 | Yamaguchi | Apr 1999 | A |
5949264 | Lo | Sep 1999 | A |
6043695 | O'Sullivan | Mar 2000 | A |
6067336 | Peng | May 2000 | A |
6079161 | Takano et al. | Jun 2000 | A |
6084479 | Duffy et al. | Jul 2000 | A |
6172571 | Moyal et al. | Jan 2001 | B1 |
6204705 | Lin | Mar 2001 | B1 |
6229345 | Kirkland et al. | May 2001 | B1 |
6242956 | McCollough et al. | Jun 2001 | B1 |
6275116 | Abugharbieh et al. | Aug 2001 | B1 |
6320435 | Tanimoto | Nov 2001 | B1 |
6466078 | Stiff | Oct 2002 | B1 |
6472915 | Moyal et al. | Oct 2002 | B1 |
6954511 | Tachimori | Oct 2005 | B2 |
6963233 | Puccio et al. | Nov 2005 | B2 |
6980038 | Boerstler et al. | Dec 2005 | B2 |
7030688 | Dosho et al. | Apr 2006 | B2 |
7589594 | Liu | Sep 2009 | B2 |
20030025538 | Bisanti et al. | Feb 2003 | A1 |
20040160281 | McDonald, II et al. | Aug 2004 | A1 |
20040239386 | Lim et al. | Dec 2004 | A1 |
20050134391 | Kimura et al. | Jun 2005 | A1 |
20050280453 | Hsieh | Dec 2005 | A1 |
Number | Date | Country | |
---|---|---|---|
60764589 | Feb 2006 | US |