Patent Application
20060125535
1. Technical Field
The present invention relates generally to phase-locked loop circuitry and more particularly to phase-locked loop circuitry using charge pumps with current mirror circuitry.
2. Description of Related Art
A phase-locked loop (PLL) is an electronic circuit with a voltage- or current-driven oscillator that is adjusted to match in phase (and thus lock on) the frequency of an input signal. In addition, PLLs are used to generate a signal, modulate or demodulate a signal, reconstitute a signal with less noise, and multiply or divide a frequency. PLLs are frequently used in wireless communications, particularly where signals are carried using amplitude modulation (AM), frequency modulation (FM) and phase modulation (PM). PLLs are more commonly used for digital data transmission, but can also be designed for analog information. Examples of applications for PLLs include frequency synthesizers for digitally-tuned radio receivers and transmitters, recovery of small signals that otherwise would be lost in noise lock-in amplifier, recovery of clock timing information from a data stream such as from a disk drive, clock multipliers, and dual-tone multi-frequency (DTMF) decoders, modems, and other tone decoders, for remote control and telecommunications.
The loop filter circuitry 135 filters the charge pump output signal 130 and generates a filtered control signal 140. Voltage controlled oscillator (VCO) circuitry 145 generates an output signal 150 whose frequency is determined by the voltage of the filtered control signal 140. The PLL circuitry 100 loops the output signal 150 back to the PDF circuitry 110 as the feedback signal 160. Optionally, frequency divider circuitry 155 is placed in the feedback path of the loop to generate the feedback signal 160 and to allow the output signal 150 to be a multiple of the input signal 105.
One problem that arises with the PLL circuitry 100 is that the magnitude of the current pulses that charge the loop filter circuitry 135 (e.g., current pulses of the charge pump output signal 130) is dependent on a variety of voltage sources. Some examples upon which the current pulses depend are power supply voltages and the voltage of the filtered control signal 140. Also, if circuitry in the PLL circuitry 100 is terminated to voltages not equidistant from the voltage of the filtered control signal 140, the current pulses for the charge pump output signal 130 may not have equal magnitude. Having unequal magnitude in the current pulses, either from voltage variations or termination mismatch, results in static phase offsets in the output of the PLLs.
The magnitude of the independent current pulses may also be adjusted to set the desired loop bandwidth. The loop bandwidth is the measure of the ability of the PLL circuitry 100 to lock onto the input signal 105 and to handle jitter. A high loop bandwidth provides a fast lock time and tracks jitter on the input signal 105, passing the jitter through to the output signal 150. A low loop bandwidth filters out the jitter of the input signal 105, but increases the lock time of the PLL circuitry 100. Typically, for a given capacitor in the loop filter 140, a smaller current for the charge pump output signal 130 produces a lower loop bandwidth and a larger current for the charge pump output signal 130 produces a higher loop bandwidth.
When determining the ideal loop bandwidth, noise performance is an important consideration. In most PLLs, the two primary noise sources that exist are noise from the VCO circuitry 145 (VCO noise) and reference noise. Each source of noise has conflicting loop bandwidth requirements to minimize the effects of the noise. VCO noise is due to thermal and shot noise in the VCO circuitry 145 and affects the output signal 150. VCO noise is usually dominant and is reduced by increasing the loop bandwidth (i.e., increasing the current output from the charge pump circuitry 125) which allows the PLL circuitry 100 to track low frequency noise (i.e., noise below the loop bandwidth) and compensate for the effect of the low frequency noise on the output signal 150. Typically, VCO noise drops off rapidly at higher frequencies, so the noise remaining above the loop bandwidth generally has little effect on the output signal 150.
Reference noise has multiple contributors such as jitter on the input signal 105, thermal noise in the charge pump circuitry 125, and supply noise associated with the voltage of the filtered control signal 140. In integer PLLs with a clean input signal 105 (i.e., no jitter), increasing the loop bandwidth reduces the effect of the thermal noise in the charge pump circuitry 125, which reduces reference noise. However, if the input signal 105 is noisy, a high loop bandwidth allows more of the reference noise to pass through to the output signal 150. Similarly, in certain types of PLLs, known as a delta-sigma PLLs, the value of the feedback divider circuitry 155 may be varied dynamically, which produces noise similar to reference noise on a noisy input signal 105. The reference noise can dominate the VCO noise and therefore the minimum possible loop bandwidth is desired rather than increasing the loop bandwidth.
Two exemplary ways to minimize the loop bandwidth are to increase the size of capacitors used in the loop filter circuitry 135 and to decrease the magnitude of the current of the charge pump output signal 130. Because area is a major concern in many designs and large capacitors increase the area requirements, decreasing the magnitude of the current is usually chosen. To achieve loop bandwidths on the order of 100 kHz using integrated capacitors of a reasonable size, currents with magnitudes ranging down to 10 s of nanoamps may be necessary. Reducing the magnitude of the current may reduce noise (e.g., reference noise), however, many other challenges arise when attempting to generate very low currents, especially in deep submicron technologies.
In particular, with nanoampere currents, such as those used in the delta-sigma PLLs, current mismatch due to the Early effect (i.e., reduction of the width of the base in bipolar transistor due to the widening of the base-collector junction with increasing base-collector voltage) can be significant depending on the value of the voltage for the filtered control signal. Moreover, the magnitude of device leakage in deep submicron technologies is often significantly larger than the actual signals being generated. Any mismatch between the currents in the up signal 115 and the down signal 120 caused by the device leakage results in the static phase offset between the input signal 105 and the output signal 160. Additionally, the mismatch generally requires one of the up signal 115 and the down signal 120 to be “on” for more than the minimum required time which allows more noise to be injected into the loop.
The invention addresses the above problems by providing a system and method for performing phase-locked loop. The system includes phase frequency detector circuitry, charge pump circuitry having a first current mirror circuitry and a second current mirror circuitry, loop filter circuitry, and voltage controlled oscillator circuitry. The phase frequency detector circuitry generates an up signal and a down signal based on the phase difference of an input signal and a feedback signal. The charge pump circuitry includes the first current mirror circuitry and the second mirror circuitry and generates a charge pump output signal based on the up and down signals. The loop filter circuitry generates a filtered control signal based on the charge pump output signal. The voltage controlled oscillator circuitry generates the feedback signal with a repeating waveform based on the filtered control signal. Advantageously, the system and method provide accurate current pulses to reduce static phase offset and provide a good resolution for tracking the input signal in the feedback signal. Another advantage is that the system and method provide good resolution for tracking the input signal at low power in submicron technologies.
To provide the accurate current pulses, the second current mirror circuitry of the charge pump circuitry may mirror a reference current of the first current mirror circuitry. Additionally, the charge pump circuitry may generate current pulses of substantially equal magnitude. Further, the system may include biasing circuitry that generates a voltage bias in the second current mirror circuitry. The biasing circuitry may comprise an operational amplifier. The biasing circuitry may generate the voltage bias based on a voltage for the first current mirror circuitry and a voltage for the charge pump output signal. Generating the voltage bias can provide current pulses of substantially equal magnitude at low currents.
In some embodiments, one of the first current mirror circuitry and the second current mirror circuitry further comprises current mirror output circuitry having a plurality of current outputs of differing magnitude. The plurality of current outputs may provide one or more loop bandwidths. Programmable loop bandwidth circuitry may select one of the plurality of current outputs to determine the loop bandwidth. Further, leakage compensation circuitry may reduce off-state leakage from the current mirror output circuitry to the charge pump output signal. The leakage compensation circuitry may also include a third current mirror circuitry configured to receive the off-state leakage and transfer the off-state leakage to the current mirror output circuitry.
In further embodiments, current pulse circuitry generates up current pulses and down current pulses in the charge pump output signal from the first current mirror circuitry and the second current mirror circuitry, respectively. Pulse leakage isolation circuitry may reduce off-state leakage from the pulse circuitry to the charge pump output signal. In one aspect, reducing the off-state leakage comprises producing a first voltage across the pulse circuitry substantially equal to a second voltage for the charge pump output signal. Additionally, charge compensation circuitry may reduce charge transfer from the pulse circuitry to the charge pump output signal.
In some embodiments, power-on circuitry generates a turn-on voltage for the filtered control signal. Power-on leakage isolation circuitry may reduce off-state leakage from the power-on circuitry to the filtered control signal. Advantageously, the power-on circuitry is large enough to start the system with little or no leakage during normal operation.
The embodiments discussed herein are illustrative of one example of the present invention. As these embodiments of the present invention are described with reference to illustrations, various modifications or adaptations of the methods and/or specific structures described may become apparent to those skilled in the art. All such modifications, adaptations, or variations that rely upon the teachings of the present invention, and through which these teachings have advanced the art, are considered to be within the scope of the present invention. Hence, these descriptions and drawings should not be considered in a limiting sense, as it is understood that the present invention is in no way limited to only the embodiments illustrated.
A circuit implementation of the principles disclosed may be implemented using PMOS transistors alone, NMOS transistors alone, parallel combinations of PMOS and NMOS transistors, or other types of transistors. In some embodiments, the parallel combinations may be preferred to result in improved charge compensation. Additionally, current mirror circuitry may have multiple possible transistor implementations.
The PFD circuitry 210 comprises any device, component, or circuitry configured to generate the up signal 215 and the down signal 220 based on the phase difference of the input signal 205 and the feedback signal 260. The replica current charge pump circuitry 225 comprises any device, component, or circuitry configured to replicate a reference current in first current mirror circuitry and second current mirror circuitry and generate the charge pump output signal 230 based on the up signal 215 and the down signal 220. One example of the replica current charge pump circuitry 225 is described below in
The loop filter circuitry 235 comprises any device, component, or circuitry configured to filter the charge pump output signal 230 and generate a filtered control signal 240. The VCO circuitry 245 comprises any device, component, or circuitry configured to generate the output signal 250 with a repeating waveform based on the voltage of the filtered control signal 240. The frequency divider circuitry 255 comprises any device, component, or circuitry configured to multiply and/or divide the frequency of the output signal 250 and generate the feedback signal 260.
The first current mirror circuitry 305 receives a reference current 340 and has a first output 345 linked to the replica switch 320. The first current mirror circuitry 305 also has a second output 350 linked to the up signal switch 330. The second current mirror circuitry 310 is linked to the biasing circuitry 315 via an input 360 and has a first output 365 linked to the replica switch 325. The second current mirror circuitry 310 also has a second output 370 linked to the down signal switch 335. The replica switch 320 is linked to the replica switch 325. The up signal switch 330 is linked to the down signal switch 335. The biasing circuitry 315 is further linked to the connection between the replica switches 320 and 325, and to the line 355. The gates of the replica switches 320 and 325 are enabled (i.e., the switch is always closed). The up signal 215 (
The first current mirror circuitry 305 comprises any device, component, or circuitry configured to replicate a reference current on one or more current mirror outputs. For example, the first current mirror circuitry 305 replicates the reference current 340 on the output 345 and the output 350. The second current mirror circuitry 310 comprises any device, component, or circuitry configured to replicate a reference current on one or more current mirror outputs. For example, the second current mirror 310 replicates the current received from the biasing circuitry 315 on the output 365 and the output 370. The biasing circuitry 315 comprises any device, component, or circuitry configured to generate a voltage bias in current mirror circuitry. One example of the biasing circuitry 315 is an operational amplifier.
During operation, the biasing circuitry 315 forces the first output 365 and the second output 370 of the second current mirror circuitry 310 to the same voltage of the charge pump output signal 230 on the line 355. The voltage bias produced by the biasing circuitry 315 results also in the first current mirror circuitry 305 having an identical and/or similar voltage bias condition as the charge pump output signal 230 on the line 355. Since there is negligible loading on the up signal switch 330 and the down signal switch 335, all the current from the reference current 340 flows through the second current mirror circuitry 310. The second output 370, if enabled by the down signal switch 335, will have an identical current as the second output 350 of the first current mirror circuitry 305 for any given voltage.
Therefore, the replica current charge pump circuitry 225 provides equal magnitude current pulses for the charge pump output signal 230 when enabled by the up signal 215 and the down signal 220. The voltage bias provided by the biasing circuitry 315 removes the effect of the voltage of the charge pump output signal 230 on the accuracy of the first current mirror circuitry 305 and the second current mirror circuitry 310 to generate the current pulses. The voltage bias provided by the biasing circuitry 315 further allows the first current mirror circuitry 305 and the second current mirror circuitry 310 to accurately replicate the reference current 340 at low power thereby providing the current pulses of equal magnitude at low currents. Accurately producing current pulses of equal magnitude prevents noise from propagating from the replica current charge pump circuitry 225 to the loop filter 235, and eventually to the VCO 245 where the noise causes offsets in the output signal 250. Additionally, replicating the current pulses with equal magnitude reduces mismatch. Furthermore, any mismatch between currents pulses for the up signal 215 and the down signal 220 is now determined by the basic mismatch between the first current mirror circuitry 305 and the second current mirror circuitry 310 under essentially identical voltage bias conditions.
In some embodiments, a large current in the charge pump output signal 230 is required relative to the magnitude of the reference current 340. For example, large divider values in the feedback divider circuitry 255 (
The converter outputs 420, 430, and 440 comprise any device, component, or circuitry configured to provide currents of differing magnitude from the reference current 480. In one example, the PCMDAC 400 receives the reference current 480 from current mirror circuitry (e.g., via the second output 350 of the first current mirror circuitry 305). The digital input to the converter selectors 460, 470 and 480 is a binary value M where M=[magnitude of converter output 430×20]+[magnitude of converter output 440×21]+[magnitude of converter output 450×22]. The converter outputs 420, 430, and 440 are binary weighted multiples of the reference current 480 in series with the converter selectors 450, 460, and 470, respectively. Therefore, the converter output signal 490 is a multiple of M times the reference current 480. This principle may be extended to large numbers of bits with the accuracy of the PCMDAC 400 limited by the matching of the individual converter outputs and converter selectors which have similar bias conditions.
The input switch 410 in the PCMDAC 400 may be a single unit (N=1), such that the converter output signal 490 equals M times the reference current 480. The input switch 410 may provide a larger unit such that the converter output signal 490 equals M/N times the reference current 490. In some embodiments, generating arbitrary functions are possible for the converter output signal 490 where the converter outputs 420, 430, and 440 are not binary weighted.
Used in the replica current charge pump circuitry 225 (
In alternative embodiments of the PLL circuitry 200 having more than one VCO (with different voltages for the filtered control signal 240), a third PCMDAC 400 may be used. The converter output 490 of the third PCMDAC 400 is controlled by the logical values selecting the VCO in use. As a result, the third PCMDAC 400 provides a constant loop bandwidth, regardless of the individual loop parameters (e.g., voltage of the filtered control signal 230). The concatenation of several PCMDACs further enables the construction of programmable PLL circuitry where the loop bandwidth is thus independently programmable. The programmable PLL circuitry facilitates usage in other circuits where a constant loop bandwidth is desirable.
Moreover, in the programmable PLL circuitry, a wide range of loop bandwidth settings are desirable in combination with a wide range of settings for the feedback divider circuitry 255 (
To minimize noise and mismatch effects, the PCMDAC 400 is preferably dimensioned with the largest transistor drive voltages allowed by the minimum operational supply voltage desired. The largest transistor drive voltage therefore results in a low voltage on the input switch 410 when large output currents are programmed. The PCMDAC 400 may also work with currents down to the nanoampere range for other loop bandwidth, VCO voltage settings, and frequency divider settings. In some embodiments, the converter selectors 450, 460, and 470 provide a very low resistance to allow sufficient voltage headroom for the wide range of currents.
In deep submicron technologies, however, low on-resistance corresponds to high leakage in the off-state. The leakage through the converter selectors 450, 460, and 470 for the converter outputs 420, 430, and 440 having a higher magnitude current may exceed the desired output signal for the converter output signal 490 (e.g., for a 7 bit DAC, the leakage may be 1-2 orders of magnitude larger than the signal). The leakage may also increase with each additional bit (and therefore each higher magnitude converter output) in the PCMDAC 400.
The converter output 510 is linked by the off-state switch 525 to the converter selector 540. The converter output 515 is linked by the off-state switch 530 to the converter selector 545. The converter output 520 is linked by the off-state switch 535 to the converter selector 550. The converter selectors 540, 545, and 550 are linked to the current mirror circuitry 510 via line 555. The current mirror circuitry 510 is linked to the PCMDAC 400 at the converter output 490.
Leakage currents in the PCMDAC 400 are replicated in the leakage compensated PCMDAC 500 which is nearly identical to the PCMDAC 400. In the leakage compensated PCMDAC 500, however, the converter outputs 510, 515, and 520 are always off or disabled. The off-state switches 525, 530, and 535 (i.e., which may be “low-leakage” switches) are placed in series with each converter output 510, 515, and 520. The converter selectors 540, 545, and 550 are only enabled when the corresponding converter selectors 450, 460, and 470, respectively in the PCMDAC 400 are disabled. For example, when converter output 420 is disabled by the converter selector 450, the converter selector 540 is enabled. Every inactive branch (converter output) in the PCMDAC 400 corresponds to an active branch in the leakage compensated PCMDAC 500 and identical leakage currents flow through both.
In some embodiments, high on-resistance in the converter selectors 540, 545, and 550 is acceptable because the currents through the converter selectors 540, 545, and 550 are so small (only leakage) that the voltage drop across the converter selectors 540, 545, and 550 is negligible even with a relatively high resistance. Additionally, the current mirror circuitry 505, used to mirror the leakage compensation current, is able to operate with low supply voltages. The leakage current can be mirrored and then fed back into the PCMDAC 400 leaving only the intended signal on the converter output 490. One advantage of the leakage compensated PCMDAC 500 is that the topology allows the use of short (low on-resistance) switch devices, enabling efficient layout and a smaller silicon area than would otherwise be possible.
In another aspect of reducing leakage in the replica current charge pump 225 (
The biasing circuitry 630 comprises any device, component, or circuitry configured as a voltage follower to reduce the potential another device, component, or circuitry. For example, the biasing circuitry 630 reduces the potential across the second switch 620. One example of the biasing circuitry 630 is an operational amplifier. During startup, both the first switch 610 and the second switch 620 are enabled (e.g., closed). Once the voltage of the filtered control signal 240 reaches a level where the VCO 245 begins running, both the first switch 610 and the second switch 620 may be turned off (e.g., opened). Because the potential across the switch 620 is zero, no leakage current will flow through the second switch 620 to the loop filter 235 via the line 660.
In some embodiments, the biasing circuitry 630 may not need to be turned off because the current through the second switch 620 dominates. In other embodiments, the biasing circuitry 630 only needs to be strong enough to supply the leakage through the second switch 620, which is off for normal operation. An alternative to using the biasing circuitry 630 with a weak buffer is to turn off the biasing circuitry 630 when the first switch 610 and the second switch 620 are closed.
The largest current and voltage levels, for the PLL circuitry 200 operating with small supply voltages, determine the size of the up and down signal switches 720, 730, 740, and 750. However, with small output currents, the leakage through the up and down signal switches 720, 730, 740, and 750 may result in a leakage current flowing through the up and down signal switches 720, 730, 740, and 750 to the charge pump output signal 230 (and into the loop filter 235). To reduce the leakage current, the biasing circuitry 710 generates a voltage bias equal or close to the voltage of the filtered control signal 240 (e.g., the voltage across the loop filter circuitry 235). When the up and down signal switches 720, 730, 740, and 750 are off, the leakage currents are diverted to the biasing circuitry 710. The biasing circuitry 710 ensures that the voltage across the up and down signal switches 720, 730, 740, and 750 is small. The small voltage reduces the leakage currents from the up and down switches 720, 730, 740, and 750 to the charge pump output signal 230.
During operation of the current pulse leakage isolation circuitry 700 (
The above description is illustrative and not restrictive. Many variations of the invention will become apparent to those of skill in the art upon review of this disclosure. The scope of the invention should, therefore, be determined not with reference to the above description, but instead should be determined with reference to the appended claims along with their full scope of equivalents.
This application claims the priority of U.S. Provisional Patent Application No. 60/635,849 entitled “Systems and Method for Phase Locked Loop Using Charge Pumps” filed on Dec. 13, 2004, which is hereby incorporated by reference.
| Number | Date | Country | |
|---|---|---|---|
| 60635849 | Dec 2004 | US |
