Oscillators – Solid state active element oscillator – Transistors
Reexamination Certificate
2001-08-01
2004-01-20
Callahan, Timothy P. (Department: 2816)
Oscillators
Solid state active element oscillator
Transistors
C331S1170FE
Reexamination Certificate
active
06680655
ABSTRACT:
FIELD OF THE INVENTION
This invention relates to voltage controlled oscillators (VCOs) and, more particularly, to a novel circuit topology which provides automatic gain control for a VCO.
BACKGROUND OF THE INVENTION
Due to the emergence of the mobile telecommunications market, the need for small, inexpensive, and low-power RF circuit components is paramount. By integrating more and more functions on the same die, single-chip transceivers are only now becoming a reality.
One of the major challenges in the design of an inexpensive transceiver system is frequency synthesis of the local oscillator signal. Frequency synthesis is usually done using a phase-locked loop (PLL). A PLL typically contains a phase detector, amplifier and a voltage-controlled oscillator (VCO). The feedback action of the loop causes the output frequency to be some multiple of a supplied reference frequency. This reference frequency is generated by the VCO whose output rate is variable over some range according to an input control voltage.
Despite numerous advances in the art, VCOs still remain as one of the most critical design components in RF transceivers. The most important parameters of a VCO are phase noise, power consumption and frequency tuning range. Specifically, it is of major importance to build low-power, low-phase-noise oscillators. This has only been possible with oscillators based on the resonance frequency of an inductor-capacitor (LC) tank circuit.
A perfectly lossless resonant circuit is very nearly an oscillator, but lossless elements are difficult to realize. Overcoming the energy loss implied by the finite Q of practical resonators with the energy supplying action of active elements is one potentially attractive way to build practical oscillators. The basic ingredients in all LC feedback oscillators, then, are simple: one transistor plus a resonator. In theory, there is no limit to the number of ways to combine a resonator with a transistor or two to make an oscillator.
In order to guarantee oscillation, the net resistance across the LC tank of an oscillator must be negative. This negative resistance is used to offset the positive resistance of all practical resonators, thereby overcoming the damping losses and reinforcing the oscillation. The negative transconductance (−G
m
) oscillator uses a cross-coupled differential pair to synthesize the negative resistance.
The basic topology for the −G
m
oscillator is shown in FIG.
1
. The oscillator may be viewed as consisting of three parts: an LC resonant tank, a negative resistance generation (positive feedback) network and a biasing network. Transistors Q
1
and Q
2
form a negative resistance generator in parallel with the LC tank that sets the frequency of oscillation. From the point of view of the LC tank, the active circuitry (transistors Q
1
and Q
2
) cancels the losses due to the finite Q of the LC resonator tank. Varactors C
var
are used in place of fixed capacitors to provide a tuning scheme. The supply voltage V
cc
is fed into the circuit through the center tap of the symmetric differential inductor L. A single inductor is usually preferred in this implementation because of the superior quality factor Q and reduced chip area compared with using two inductors connected in series. Finally, in order to reduce the common-mode gain and maintain a constant total emitter current shared equally by the two collector circuits, a large emitter resistance must be realized. A constant current source It is typically used to simulate this high resistance (since large resistances are hard to fabricate on IC chips). The tail current It aids in controlling the swing of the oscillator and may be implemented with any of the common current generator circuits already in existence.
An important element reflecting the performance of VCOs is phase noise: instantaneous variations in the frequency of oscillation as determined by loss elements such as series resistances in the circuit. Variations in the amplitude of oscillation correspond to variations in phase noise as the total energy stored in the tank includes signal energy and noise energy. VCO operation is, therefore, susceptible to variations in the load resistance and component tolerances, as well as variations in the values of the inductor, the capacitor and their parasitic resistances, caused by semiconductor process variations.
The noise contributed by the active drive circuitry in
FIG. 1
can be minimised by operating the LC tank circuit at a maximum attainable oscillation amplitude. This is because the total energy stored in the tank (which is a function of the peak voltage amplitude) includes both signal energy and noise energy. As the oscillation amplitude increases, the percentage of total energy attributed to noise from the drive circuit decreases, thereby reducing the phase noise of the oscillator. Therefore, it is desirable to produce the largest voltage swing possible across the tank.
However, the circuit in
FIG. 1
cannot be used to generate large voltage swings across the tank circuit because the bases of transistors Q
1
and Q
2
are connected to the collectors of Q
2
and Q
1
, respectively. If the voltage swing across the tank becomes too great, Q
1
and Q
2
are driven into saturation during peak voltage excursions of the tank thereby introducing additional phase noise into the oscillator. If the amplitude is increased further, the emitters of Q
1
and Q
2
might even breakdown. Therefore, trying to increase the amplitude of oscillation of the circuit in
FIG. 1
above a certain bound is counterproductive because the saturation of the transistors introduces additional phase noise.
The limited voltage swing available from the oscillator of
FIG. 1
aggravates its susceptibility to variations in load resistance because the circuit also requires a minimum tail current to start oscillation. The minimum tail current needed to start the oscillation is referred to as the critical factor. If the tail current is set to exactly the critical factor, the oscillator just barely begins to oscillate and then stops. Therefore, to obtain a sustainable oscillation, it is necessary for the tail current to be greater than the critical value. Depending on the value of the tail current selected, the oscillation amplitude settles to a steady state value. However, the amplitude must not be allowed to grow too large or the transistors will saturate as discussed above, thereby increasing the phase noise of the system. The tail current must be chosen within a small window between the critical factor and some maximum value, otherwise the oscillator will either fail to start or the transistors Q
1
and Q
2
will saturate.
In a board-level design, manual bias adjustments, although costly, can be made to drive the circuit so as to set the drive conditions to assure start-up and prevent saturation of the drive transistors. However, in a monolithic implementation, there is no opportunity for manual intervention. The oscillator must consistently meet its difficult specifications, not only over a wide range of supply voltage and temperature, but also in the presence of numerous parametric variances associated with the production of commercial integrated circuits, and perhaps for a variety of possibly different external (board-level) components. More importantly, in a wide tuning range oscillator, the loss in a particular component may vary at different frequencies of oscillation. A change in the loss of a component and associated change in the loss in the tank necessitates a change in the ideal current source for optimal biasing to be achieved. In other words, if the loss in the tank changes over the frequency of operation of the VCO, then so does the ideal biasing current. Without any feedback, there is no way to adjust the biasing current to an optimal value for all frequencies of operation.
If the load resistance and all of the other factors that affect the oscillation amplitude were known in advance, it would be possible to select a value of tail current that is greater than the critical value s
Callahan Timothy P.
Cox Cassandra
Sige Semiconductor Inc.
LandOfFree
Automatic gain control for a voltage controlled oscillator does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Automatic gain control for a voltage controlled oscillator, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Automatic gain control for a voltage controlled oscillator will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3221480