Miscellaneous active electrical nonlinear devices – circuits – and – Specific identifiable device – circuit – or system – Unwanted signal suppression
Reexamination Certificate
1997-12-16
2002-01-01
Le, Dinh T. (Department: 2816)
Miscellaneous active electrical nonlinear devices, circuits, and
Specific identifiable device, circuit, or system
Unwanted signal suppression
C327S562000, C327S103000, C330S109000, C330S258000
Reexamination Certificate
active
06335655
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to a filter circuit used as a basic circuit when processing analog signals in a MOS integrated circuit (IC).
2. Description of the Related Arts
In recent years, due to the increase in digital devices and advances in digital signal processing technology, CMOS integrators applied for digital signal processing account for a large portion of the semiconductor market.
However, since video and audio signals have analog input/output, they can be more easily processed by analog processing. Even when video and audio signals are digitally processed, analog circuits are required for A/D and D/A conversion and filtering carried out before or after conversion and in a clock-generating oscillator and such like. Conventionally, bipolar technology has been regarded as suitable for analog circuits, whereas CMOS technology is regarded as unsuitable except for some circuits such as analog switches, sample holders or such like.
However, bipolar and BiCMOS processing are rather expensive, and in view of the strong demand for a 1-chip CMOS structure achieved by digital/analog consolidation, there has been an increase in development of circuits for processing analog signals with CMOS circuitry.
Analog signal processing features an important function known as an “active filter” which has a high frequency of use and exerts considerable influence on total performance. Conventionally, discrete-time processing filters such as switched capacity filters (SCF) or sample data filters have been the main active filters used in CMOS analog technology. While these filters have the advantage of high precision, since frequencies are fixed precisely according to a clock, and have low sensitivity to capacitor variations originating during manufacturing, these filters have the following disadvantages:
1. continuous-time filters are required before and after due to the existence of “aliasing”;
2. since a frequency band several times greater than frequency used for operational amplifier and sample-and-hold (S/H) circuit is required, the filter cannot be used for frequencies higher than the video band.
3. circuit scale is large and therefore not economical. Therefore, it is not possible to produce a simple and inexpensive filter which can be used at high frequencies. This problem cannot easily be resolved since it derives directly from the fact that the filter is a discrete-time filter. Recently, attempts are under way to develop a high-performance continuous-time CMOS filter. The most popular continuous-time filter is a “biquad circuit”, consisting of two integrators each comprising a transconductor (Gm) circuit and a capacitor, in which a second-order filter is multistage-connected in order to obtain desired filtering properties. In a bipolar technology, transconductor (voltage-current conversion characteristic) is linearized its characteristic using a resistor and a “gain cell structure of transistors”. However, when the above method is used in CMOS, since element Gm (transconductance) is small, many elements of enormous size are required, which is extremely uneconomical. Transconductance properties are therefore obtained by performing voltage-to-current conversion using a differential pair-transistor direct-coupled to the source.
However, in a MOS transistor, gate-source voltage versus drain current is a square relationship and the integration property of capacitor current versus voltage is linear, a single output of the integrator has second-order distortion. It is therefore necessary to cancel second-order distortion using a fully differential input and output signals. In such fully differential processing, since the output DC voltage cannot usually be determined, a bias circuit must be added to determine the DC operating point. Generally, therefore, output DC voltage is detected and DC feedback technique is applied to the bias of the transconductor input. This method is called DC feedback (or common-mode feedback or such like).
FIG. 6
shows a conventional example of this type of fully differential biquadratic circuit provided with DC feedback. This circuit is made up of integrators in two stages, each stage consisting of two transconductors sharing a common output terminal and a capacitor connected to said output terminal. Lower transconductors Gm
1
+ and Gm
2
+ correspond to + input in a single configuration; upper transconductors Gm
1
− and Gm
2
− correspond to − input (feedback input) in a single configuration.
Two-stage integrators are cascade-connected as above to form a low-pass filter or such like. The output DC voltage is controlled to a predetermined voltage by monitoring the output of DC feedbacks 1 and 2 at each integrator and controlling the bias current of the output terminal in each stage. In addition to an LPF, other filter types such as BPF and HPF can be created by modifying the signal input position and output signal extraction position.
FIG. 7
shows a concrete example of integrators in the stages forming the filter in
FIG. 6
realized using CMOS circuitry. A pair of source-coupled differential transconductors are used to produce transconductance. The differential circuit comprising M1, M2 and I1 corresponds to transconductor Gm
1
+ or Gm
2
+; the differential circuit comprising M3, M4 and I2 corresponds to transconductor Gm
1
− or Gm
2
−. The current outputs are summed at the output point. Output is differentially extracted by biasing with two current supplies I3, I4 which join at GND. With respect to differential input bias currents I1 and I2, these current supplies I3, I4 must be precisely (I1+I2)/2 respectively. If this relationship is evenly slightly disrupted, since the DC impedance of the output terminal of each stage is extremely high, the unbalance between the upper and lower current supplies causes the output DC voltage to be greatly disrupted, leading to instability.
“DC feedback” is a circuit designed to counter this problem, ensuring stability by fixing output DC potential at a certain voltage. One of the resistor terminals is connected to the output terminal in FIG.
7
. The other resistor terminal is now connected to each other, and is compared by the operational amplifier with the intended voltage Vref. When all output signals are fully differential and the two resistors are equal values, the DC potential of the output signal can be extracted from the center point and compared with Vref. When the DC potential is higher than Vref, the current from the current supply is increased and the common-mode voltage of the output is lowered. Conversely, when the DC potential is lower than Vref, the current from the current supply is decreased, raising the common-mode voltage of the output.
Thus current is controlled so that the common-mode voltage of the output signal equals Vref. The circuit is biased so that the DC voltages of differential outputs in each stage of the filter equal Vref, but the number of circuit elements is likely to be increased. Not only does the operational amplifier itself require a considerable number of elements, but a buffer circuit and such like must be provided as shown in the diagram in order to prevent the resistors which detect center point voltage from influencing the high-impedance integrator output terminals. Moreover, a DC feedback circuit is required for each integrator, thereby taking up a large portion of the overall area of the filter.
In the circuit shown in
FIG. 7
, for instance, the essential portion of the integrator within the dotted line on the left side comprises 4 MOS transistors, 4 current supplies and 1 capacitor, thus requiring approximately 10~15 elements. By contrast, a DC feedback circuit for setting bias has as many as 20~30 elements, thereby taking up two thirds of the area. Since the filter is formed simply by assembling these elements, it follows that the DC feedback circuit takes up roughly two thirds of the total filter area. This need for a DC feedback has resulted in increased costs f
Le Dinh T.
Pillsbury & Winthrop
LandOfFree
Filter circuit does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Filter circuit, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Filter circuit will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2864281