The present invention pertains in general to phase lock loops (PLLs) and, more particularly, to the operation of the charge pump and loop filter portion of the PLL.
This application is related to U.S. Provisional Patent Application Ser. No. 60/640,488, filed on Dec. 30, 2004 entitled “INTEGRATED PLL LOOP FILTER AND CHARGE PUMP.”
Phase lock loops utilize a phase detector for comparing the phase of a reference clock with that of an output clock that utilizes a voltage controlled oscillator (VCO) to generate a phase error that varies the control voltage on the input to the VCO. By adjusting this voltage, the phase of the VCO can be locked the phase of the reference clock. Typically, some type of loop filter is disposed between the phase detector and the VCO. In a charge pump PLL, a typical phase detector generates control voltages for controlling a charge pump circuit which is operable to selectively pump charge to a node for increasing a voltage level or pulling charge from the node to provide a decreasing voltage level. To increase the voltage level, charge is sourced from a supply voltage and, to decrease the voltage level, charge is sinked to a ground reference. When the relative phase between the VCO and the reference clock are either lagging or leading, then either the sourcing or sinking of a charge pump is controlled.
This charge pump is typically facilitated with two current sources that are switched to the voltage input to the VCO. When charge is being sourced to the node, the phase of the VCO will change from either a lagging or leading to a leading or lagging phase, such that the phase detector will then cause the charge pump to sink current. When the PLL is locked, the phase error should be substantially at a zero phase error which should result in no current being sourced to or sinked from the voltage control input of the VCO. However, conventional charge pumps are fabricated with two transistor switches, one for sourcing current and one for sinking current, that are switched to either a conducting state or a non-conducting state. However, the current source is a function of the voltage on the VCO input. As the voltage changes, the characteristics of the switch and the associated current source will also change. Therefore, if the voltage changes, i.e., it is not constant, there is a possibility that the currents will not be balanced. If they are not balanced, then a phase error can result at phase lock, which could cause jitter in the clock. Thus, it is desirable that the currents are balanced for all possible voltages input to the VCO over the entire range required during the operation thereof.
The present invention disclosed and claimed herein, in one aspect thereof, comprises a charge pumped phase locked loop circuit (PLL). The PLL includes a phase detector for detecting the phase error between a reference clock and an output clock to generate a phase error signal. A charge pump is provided that is controlled by the phase error signal to either source current to an intermediate control node or to sink current therefrom. An isolation circuit maintains the intermediate control node at a virtual AC reference voltage such that it remains at substantially the same voltage during the sourcing of current thereto or sinking of current therefrom, the isolation circuit generating a control voltage on the output thereof to control the frequency of the output clock. A loop filter is provided for filtering said control voltage.
For a more complete understanding of the present invention and the advantages thereof, reference is now made to the following description taken in conjunction with the accompanying Drawings in which:
a illustrates a schematic diagram of an alternate embodiment of a prior art charge pump and loop filter;
Referring now to the drawings, and more particularly to
The phase detector 104 is operable to generate a phase error on a line 114. This phase error provides a control signal for a charge pump 116. The charge pump 116 is operable to either increase the charge on a node 118 or decrease the charge thereon, such that the voltage thereon will be controlled as a function of the output of the phase detector 104. This is filtered by a loop filter 120 to provide a voltage VO for input to the VCO 108. This voltage VO provides a control voltage to control the frequency of the VCO 108. This is a conventional operation.
Referring now to
The loop filter 120 is comprised of a resistor 212 connected between node 118 and one plate of a capacitor 214. The other plate of capacitor 214 is connected to ground. Resistor 212 has a value of R2 and capacitor C has a value of C′. The voltage VO is defined as follows:
where: s=jω
This results in the following relationship:
The absolute value of VO/icp has a frequency response as set forth in the plot of
Referring now to
The node 406 is also connected to the gate of a p-channel transistor 420, the source/drain path thereof connected between VDD and a node 422. Node 422 is a switched node which is switched between an output node 424 and a bias node 426 with switches 425 and 427, respectively. It is switched between an output node 424 and a bias node 426. Node 426 is connected to a bias voltage VB, which could be the same bias voltage disposed on the negative input of the amplifier 416, or a different bias voltage. The node 414 is connected to the gate of an n-channel transistor 428, the source/drain thereof connected between ground and a node 430. Node 430 is switched between node 426 and node 424 with switches 432 and 434, respectively.
The switch 425 is controlled by a clock signal U and switch 427 is controlled by a clock signal U-Bar. Switch 434 is controlled by a clock signal D and switch 432 is controlled by the inverse thereof, a clock signal D-Bar.
In operation, the current through transistor 404 is mirrored over to the mirror leg comprised of transistors 408 and 412. This controls the bias on the transistor 428, such that the current through transistor 420 will be icp and the current through transistor 428 will be icp, since the current through transistor 408 is icp and the current through transistor 412 is icp. If switch 425 is closed, connecting node 422 to node 424, then current will be sourced by transistor 420 to node 424. This current will be icp. However, if the node 430 is disconnected from node 424 and is left floating, then no current will flow through transistor 428 until it is connected. Once connected by switch 434 (with switch 425 open), then current will be sinked from node 424. However, if transistor 420 is completely off prior to switch 434 turning on, this will require the capacitance on node 430 to be charged up. To prevent the situation, switch 432 will be connected to node 426 when switch 434 is open and switch 425 is closed. In this manner, current will flow through transistor 424, maintaining a current of icp, Thus, when switch 425 opens and switch 434 closes, and switch 432 opens, then the VDS of the transistor will not have to charge up and, therefore, will not affect the voltage on node 424 due to the change in state of transistor 428. Therefore, when either of the switches 425 or 434 are open, the complimentary side thereof, switch 427 or 432, will be connected to the bias voltage 426, such that current flows there through. In the prior art embodiment of
Referring now to
With the isolator 440 in the embodiment of
Referring now to
Referring now to
Node 502 is connected to the negative input of an amplifier 614, the positive input thereof connected to the bias voltage VB. The negative input of the amplifier 614 represents a virtual ground, such that the node 502 will always be connected to virtual ground. The nodes 502 and 506 are intermediated control voltage nodes. The negative input of the amplifier 614 is connected through a series resistor 616, labeled R2, to the output thereof on a node 618. Resistor 610 is labeled R1 and resistor 616 is labeled R2. Therefore, the voltage on node 606 will be amplified by amplifier 614, an inverting amplifier, by the following relationship for a voltage component Vi+:
where R1>R2. This will be summed at the output node with the voltage associated with −icp. This will result in the voltage component Vi− as follows:
Vi−=R2(−icp)
The combined voltage VO on the output node 618 will be Vi++Vi− with the following relationship:
where:
where: β=R1/R2.
It can be seen that the value of β represents the capacitance multiplication. As such, for a βof “10” the capacitance value can be reduced by a value of “10.” This results in a lower required capacitor value and consequently, a lower amount of area on the semiconductor surface that is required to realize such a capacitor.
Although the preferred embodiment has been described in detail, it should be understood that various changes, substitutions and alterations can be made therein without departing from the scope of the invention as defined by the appended claims.
Number | Name | Date | Kind |
---|---|---|---|
4987387 | Kennedy et al. | Jan 1991 | A |
6111468 | Tanishima | Aug 2000 | A |
6344772 | Larsson | Feb 2002 | B1 |
6711229 | Harada | Mar 2004 | B1 |
6768359 | Hsu | Jul 2004 | B2 |
6774679 | Nogami | Aug 2004 | B2 |
7030688 | Dosho et al. | Apr 2006 | B2 |
7154345 | Moyal et al. | Dec 2006 | B2 |
Number | Date | Country | |
---|---|---|---|
20060145770 A1 | Jul 2006 | US |
Number | Date | Country | |
---|---|---|---|
60640488 | Dec 2004 | US |