Integrated programmable continuous time filter with programmable capacitor arrays

Abstract

A programmable capacitor array including a plurality of user-selectable, numerically weighted capacitors, each of which includes at least one fixed capacitor and one manufacturer-controlled trim capacitor, advantageously provides a variety of selectable capacitance values for a programmable analog integrated circuit. When coupled to a memory, for example a static memory, switches can be controlled that determine whether a particular fixed capacitor (user-selectable) or trim capacitor (manufacturer-selectable) is electrically coupled into the circuit. User access to those portions of memory controlling switches associated with the trim capacitors can be restricted via an I/O interface and security command. Such programmable capacitor arrays allow efficient implementation of user-programmable filter circuits where the user can conveniently program or reprogram a variety of filter parameters.

Description

BACKGROUND OF THE INVENTION

1. Field of the Invention

This invention relates to analog signal processing, and more particularly to programmable integrated circuits that perform analog processing, and include both user and manufacturer programmable capacitor arrays.

2. Description of the Related Art

Typically, analog integrated circuits are designed to receive one or more analog input signals, and process those signals by performing specific functions such as amplification, attenuation, filtering, integration, addition and subtraction. These functions usually dictate the topology of the analog integrated circuit. For example, the topologic arrangement of operational amplifiers and resistors are adjusted to provide either inverting or non-inverting gain. Every topology has specific noise, distortion and offset voltage sensitivities. Changing an analog circuit's function often requires a change in the topology of the analog circuit, which in turn changes the noise, distortion and offset voltage characteristics of the circuit.

An analog integrated circuit with a programmable analog circuit block architecture permits change in a function of the analog circuit without changing the topology of the circuit elements, thereby reducing changes in voltage offset and distortion created by changes in topologies. Examples of such analog integrated circuit architectures can be found in U.S. Pat. No. 5,574,678, entitled “Continuous Time Programmable Analog Block Architecture,” by James L. Gorecki, which is incorporated herein by reference in its entirety.

In addition to such programmable analog circuits, another way to change an analog circuit's function or performance characteristics is to vary the values of components in the circuit. In the case of an analog filter circuit, it is well known that the performance characteristics of the filter (e.g, the cutoff or corner frequency f c , the stop frequency f s , the passband ripple, and the stopband attenuation) depend generally on the RC product of the filter, and more particularly on the values of various resistors and capacitors in the filter. Thus, tunable or programmable filter circuits depend on the ability to adjust resistor and/or capacitor values.

When implementing integrated or monolithic programmable analog circuits, and particularly filters, processing and temperature variations can cause the values of various components in the circuit to vary by as much as 50%, which in turn can cause important circuit parameters, such as RC product, to vary by even greater amounts from their nominal design value. One conventional solution is to use integrated amplifiers in conjunction with high-precision, external resistors and capacitors. However, this approach is counter-productive to the goal of creating an integrated filter circuit, and increases both the size and the cost of the resulting filter.

Another solution is to provide for a fully integrated filter with an RC product that can be tuned to account for various process and temperature variations. One method for doing this, is to replace passive resistors with active resistors made from metal oxide field effect transistors (MOSFETs). For example, a feedback circuit can be implemented which continuously measures the RC product of the filter with reference to, for example, a clock or an external high-precision resistor. The outputs of the tuning circuit are then continuously applied to the MOSFET devices to set their resistances. A disadvantage to this technique is that the tuning circuits are very complex, sometimes more complex than the filter itself. Additionally, such tuning circuits usually require substantial amounts of power, and the variable range of the MOS resistors may not be sufficient to provide desired filter tunability. As an alternative to a tuning circuit and MOSFET resistors, arrays of passive resistors can be used with analog switches to afford programmability. However such an approach suffers from signal distortion due to the analog switches.

Still another alternative is to construct tunable filters by using tunable capacitor arrays and linear, passive resistors. Unfortunately, the value of capacitance per unit area in ordinary integrated circuit manufacturing processes is usually not well controlled. Therefore, if an accurate value of capacitance is needed for an integrated circuit capacitor, the manufacturer must “trim” the capacitor as needed to compensate for the large variations inherent in ordinary integrated circuit manufacturing processes. One well known way of trimming the value of an integrated circuit capacitor is to provide an array in which a main capacitor of somewhat smaller than nominal capacitance is initially in parallel with a number of smaller “trim” capacitors, which then can be removed from parallel connection with the main capacitor by trimming, for example by using a laser to cut electrical connections which directly or indirectly remove one or more of the trim capacitors from parallel connection to the main capacitor. This technique has the significant drawback that it requires that there be a capacitor trimming operation, which can add significantly to the cost of the product. Another problem with this technique is that it is not well suited to use in a user-programmable capacitor array because each user-programmable capacitance value has to be separately trimmed. Compounding this problem is the desire to have a wide range of programmable capacitances with high resolution (i.e., small steps between available capacitances).

Accordingly, it is desirable to have a user-programmable capacitor array that has high resolution and a wide range of available capacitances, while at the same time having an efficient and effective mechanism for a manufacturer to carry out trim operations. Additionally, it is desirable to have such a user-programmable capacitor array for use in programmable analog integrated circuits, particularly programmable filter circuits.

SUMMARY OF THE INVENTION

It has been discovered that a programmable capacitor array including a plurality of user-selectable, numerically weighted capacitors, each of which includes at least one fixed capacitor and one manufacturer-controlled trim capacitor, advantageously provides a variety of selectable capacitance values for a programmable analog integrated circuit. When coupled to a memory, for example a static memory, switches can be controlled that determine whether a particular fixed capacitor (user-selectable) or trim capacitor (manufacturer-selectable) is electrically coupled into the circuit. User access to those portions of memory controlling switches associated with the trim capacitors can be restricted via an I/O interface and security command. Such programmable capacitor arrays allow efficient implementation of user-programmable filter circuits where the user can conveniently program or reprogram a variety of filter parameters.

Accordingly, one aspect of the present invention provides a programmable capacitor array including a plurality of user-selectable capacitors. Each of the plurality of user-selectable capacitors is coupled in series with an associated user-controlled switch. Each of the plurality of user-selectable capacitors and associated user-controlled switches is coupled in parallel with each other of the plurality of user-selectable capacitors and associated user-controlled switches. At least one of the user-selectable capacitors includes at least one trim capacitor coupled in series with a manufacturer-controlled switch, and at least one fixed capacitor coupled in parallel with the at least one trim capacitor and the manufacturer-controlled switch.

In another aspect of the invention, a programmable analog circuit includes an amplifier having at least one input terminal and at least one output terminal, and a programmable capacitor array coupled between the at least one input terminal and the at least one output terminal. The programmable capacitor array includes a plurality of user-selectable capacitors. Each of the plurality of user-selectable capacitors is coupled in series with an associated user-controlled switch. Each of the plurality of user-selectable capacitors and associated user-controlled switches is coupled in parallel with each other of the plurality of user-selectable capacitors and associated user-controlled switches. At least one of the user-selectable capacitors includes at least one trim capacitor coupled in series with a manufacturer-controlled switch, and at least one fixed capacitor coupled in parallel with the at least one trim capacitor and the manufacturer-controlled switch.

BRIEF DESCRIPTION OF THE DRAWINGS

The present invention may be better understood, and its numerous objects, features, and advantages made apparent to those skilled in the art by referencing the accompanying drawings.

FIG. 1 is a block diagram of a programmable continuous time 5th order low-pass filter.

FIG. 2 shows circuit details of the 5th order low-pass filter of FIG. 1 .

FIG. 2A illustrates a common mode feedback stage.

FIG. 3 illustrates portions of a programmable capacitor array.

FIG. 4 illustrates a portion of programmable capacitor C 1 including logic and memory to facilitate programmability.

FIG. 5 illustrates a simplified circuit topology used for testing programmable capacitor C 1 .

FIG. 6 illustrates a calibration circuit used in conjunction a portion the filter shown in FIG. 2 .

DETAILED DESCRIPTION

FIG. 1 is a block diagram of a programmable continuous time 5th order low-pass filter integrated circuit 100 . Differential inputs 110 provide input signals to differential input instrumentation amplifier 120 . Although only one instrumentation amplifier 120 is illustrated, an alternative implementation would have dual input instrumentation amplifiers and provide, for example, signal summation. Instrumentation amplifier 120 is implemented as a programmable gain amplifier with, for example, four gain values (1, 2, 5, and 10) that are user-programmable. Noise signals that are common to both inputs, i.e., common-mode signals, typically pass through amplifier 120 un-amplified. The effectiveness of amplifier 120 at rejecting common-mode signals is usually measured as a function of differential voltage gain to the common-mode voltage gain called the common-mode rejection ratio (CMRR). Consequently, a common-mode feedback topology (not shown) is usually implemented with or within instrumentation amplifier 120 .

Output signals from instrumentation amplifier 120 are sent to 5th order low-pass filter 130 . Filter 130 is just one example of a programmable analog block that can be selectively coupled into integrated circuit 100 . For example, integrated circuit 100 can include additional programmable analog blocks, for example a high-pass filter that can be programmably selected by a user and thereby coupled into a signal path. For a more general description of programmable analog integrated circuit architectures, see the aforementioned U.S. Pat. No. 5,574,678. Signals then pass from filter 130 to output amplifier 140 , and to differential outputs 150 .

Memory 170 and 180 , which is preferably part of integrated circuit 100 , store user-programmable values that specify particular filter parameters as illustrated, i.e, configuration A and configuration B. As indicated above, and further described below, the values stored typically represent settings for user selectable capacitor values, a particular combination of capacitor values yielding a corresponding set of filter parameters. Memory 170 and 180 can be implemented using static read only memory, dynamic random access memory, static random addressable memory, a serial string of shift registers, electronically erasable (E 2 ) memory, or any other memory. Multiplexer 160 allows one of the two stored configuration to be programmed into filter 130 , depending on, for example, a user command selecting one of the configurations.

Note that throughout FIGS. 1 and 2 , the circuit paths are represented as being single-ended for convenience. The circuits illustrated are implemented with fully differential circuit pathways in their preferred embodiments, although single-ended operation is possible by design, by programming, or via conversion circuits at the input and output nodes. Differential architecture substantially increases dynamic range as compared to single-ended I/O, while affording improved performance with regard to circuit specifications such as common mode rejection and total harmonic distortion. Moreover, differential operation affords added immunity to variations in the circuit's power supply.

FIG. 2 illustrates details of the 5th order low-pass filter circuit 130 . Filter 130 is constructed using a ladder architecture, a filter architecture well known to those having ordinary skill in the art of filter design and use. The ladder architecture is based on passive LC components cascaded, much in the same way as a typical RC filter structures. Due to the lossless nature of their reactive LC components, doubly terminated passive ladder filters minimize sensitivity of the filter's transfer function to changes in element values due to temperature and manufacturing tolerances. This low sensitivity, particularly in the passband, to component tolerance makes the ladder architecture very advantageous. Active ladder filters such as filter 130 are based upon topologies that use RC equivalents to realize a passive LC ladder filter. Two common approaches to realizing and active inductor are component simulation using general immittance converters to model inductors, and operational simulation to simulate the branch currents of a network as voltages. In the latter approach, RC operational amplifier integrators are used with time constants proportional to the inductors and capacitors in the original, simulated filter.

Note that filter 130 is a low-pass filter, but other filter types (e.g., high-pass and band-pass) could similarly be implemented, as well as different filter topologies (e.g, cascaded active RC filters, bi-quadradic filters).

Amplifiers A 1 , A 2 , A 3 , A 4 , and A 5 are important to the overall performance of filter 130 . Although the ladder architecture tolerates finite bandwidth effects better than other filter topologies, amplifier unity-gain bandwidth should be approximately 10-20 MHz so that the ladder filter can achieve good performance across a wide range of corner frequencies up to 500 kHz. Amplifiers A 1 , A 2 , A 3 , and A 4 drive internal nodes of the filter, while amplifier A 5 must drive output loads as well. Note that in a typical implementation, amplifier A 5 integrates the functions of the final amplifier stage of the ladder filter, and output amplifier 140 , thereby eliminating the need for a separate output buffer. Also in a typical implementation, amplifiers A 1 and A 5 can include special trim features. For example, amplifier A 1 can include an analog trim cell for residual differential offset of the entire filter. The differential offset correction is preferably made in the first amplifier of the ladder so that added stress, for example pushing the amplifier to a rail, is not placed on succeeding amplifiers in the filter. Additionally, amplifier A 5 can include circuitry to implement automatic calibration of the filter, and to provide differential analog trim of the common mode output voltage.

As noted above, integrated circuit 100 can include a common-mode feedback circuit implemented in conjunction with instrumentation amplifier 120 . FIG. 2A illustrates a common mode feedback (CMFB) stage 190 typically located after instrumentation amplifier 120 in order to determine the CMFB of the input front end. CMFB stage 190 is an inverting gain stage including an amplifier 195 having feedback resistors ( 193 and 194 ) and input resistors ( 191 and 192 ). Since the CMFB is only on the output of the amplifier, the input common mode voltage of amplifier 195 moves around. For example, if the gain of the CMFB stage 190 is −1, the input nodes of amplifier 195 will swing from 1.75 volts to 3.25 volts, half of the typical 1.0-4.0 volts swing on the input. This architecture improves the CMRR because the mismatch in the amplifier 195 output stage is divided by the differential loop gain. However, one source of error is resistor mismatch. The plus and minus channel in the differential paths each go through a resistor division path. If the gain of the two paths are slightly different, the CMRR will degrade since the circuit cannot distinguish the common mode and the differential mode signals The common mode signal is subtracted after the two paths have gone through the resistor division. Since the resistor matching in many integrated circuit fabrication processes is poor, (e.g., the resistor matching is as bad as 2%, which corresponds to 46 dB CMRR even if only one pair of resistors has such mismatching) a resistor trim feature is typically included. An example of such a trim feature is shown in U.S. patent application Ser. No. 09/506,180, entitled “An Amplifier Having an Adjustable Resistor Network,” by Hans W. Klein, Jian Li, and Yaohua Yang, filed on Feb. 17, 2000, which is incorporated herein by reference in its entirety.

One difficulty can arise due to the parasitic pole caused by the feedback resistors and the parasitic capacitance at the input of amplifier 195 , which can include C GS of the amplifier 195 input devices and the capacitance associated with transistors in a CMRR trim feature. Since it is desirable for the bandwidth of the loop to be large (in order to get good closed loop AC response), the parasitic pole is preferably pushed far out by using small resistors (typically 3.5 k-ohms) in the feedback.

Filter parameters are determined by the various combinations of resistances and capacitances in filter 130 . As discussed above, and illustrated in FIG. 2 , the resistances of the filter remain at fixed values, and filter programmability is achieved through the use of variable capacitors, specifically programmable capacitor arrays C 1 , C 2 , C 3 , C 4 , C 5 , C 213 , C 231 , C 435 , and C 453 . By varying the values of these capacitors, different types of low-pass filters (e g., Chebyshev, Bessel, Gaussian, Legendre, elliptic (Cauer), and linear phase equi-ripple) with different ranges of operating parameters can be implemented using filter 130 . Table 1 shows an example of the types of filters that can be implemented using filter 130 and varying the values of programmable capacitor arrays C 1 , C 2 , C 3 , C 4 , C 5 , C 213 , C 231 , C 435 , and C 453 . Again, it should be noted that filter 130 is typically implemented differentially, and thus the actual circuit includes two instances of each of the programmable capacitor arrays C 1 , C 2 , C 3 , C 4 , C 5 , C 213 , C 231 , C 435 , and C 453 .

TABLE 1
		Corner	Pole
Filter Category	Filter Type	Frequency	Spacing
Elliptic CC0510	θ = 22°-55°	50 kHz-500 kHz	±3.7%
(0.10 dB ripple)
Elliptic CC0510	θ = 22°-55°	50 kHz-500 kHz	±3.7%
(0.15 dB ripple)
Elliptic CC0510	θ = 22°-55°	50 kHz-500 kHz	±3.7%
(0.20 dB ripple)
Bessel		50 kHz-500 kHz	±3.7%
Linear Phase Equi-	Phase error = 0.05°,	50 kHz-500 kHz	±3.7%
ripple Error	0.5°
Gaussian	Gaussian to 6 dB,	50 kHz-500 kHz	±3.7%
	12 dB
Legendre		50 kHz-500 kHz	±3.7%

In order to implement these different types of filters in the parameter ranges specified, the capacitor sizes in filter 130 must be variable across wide ranges of values, and must vary, in sufficiently small steps. Table 2 illustrates an example of the programmable capacitor array requirements to implement the various filters of Table 1. Note that Table 1 is merely an example of one possible implementation, and those having ordinary skill in the art will readily recognize possible variations to implement different types of filters, different parameter ranges, and different resolutions.

TABLE 2
Capacitor	No. of	Minimum	Minimum	Maximum
Array	Capacitors	Capacitor (pF)	Capacitor (pF)	Spacing
C1	12	42.732	0.417	±3.72%
C2	9	48.369	1.241	±3.27%
C3	9	83.090	2.066	±3.63%
C4	8	59.041	1.856	±3.47%
C5	9	39.607	0.984	±2.95%
C213	11	9.359	0.122	±3.34%
C231	11	9.359	0.122	±3.34%
C435	12	32.359	0.326	±3.47%
C453	12	16.180	0.163	±3.47%

In the example of filter 130 , it should be noted that programmable capacitor arrays C 213 and C 231 have a fixed relationship, namely that they must have the same value. Similarly programmable capacitor arrays C 435 and C 453 have a fixed relationship, namely that programmable capacitor array C 453 is half the value of programmable capacitor array C 435 . Consequently, although a total of 93 capacitors are used in the programmable capacitor arrays, only 70 bits are needed to program these programmable capacitor arrays (11 bits for C 213 and C 231 combined, and 12 bits for C 435 and C 453 combined).

As illustrated by the examples in Table 2, each of the programmable capacitor arrays do not necessarily need to have he same number of capacitors, except where design restrictions require such similarity. This is due, in part, to the fact that higher resolution of individual programmable capacitor arrays is needed in some parts of the circuit, as compared to other parts of the circuit. Also, the values of each of the capacitors in any one programmable capacitor array are numerically weighted to provide the desired range and resolution of capacitance values. For example, a programmable capacitor array can be binarily weighted, wherein the values or each of the capacitors in the array are related as 1, 2, 4, 8, 16, etc. Other schemes for numerically weighting the capacitors in a programmable capacitor array, for example uneven weighting schemes (e.g., 1, 1.3, 1.6, 2.1, etc.), linear schemes, or other non-linear schemes, can be implemented as appropriate.

FIG. 3 illustrates portions of one of the programmable capacitor arrays, C 1 . Programmable capacitor array C 1 includes, for example, twelve user-selectable capacitors in parallel, four of which ( 300 A, 310 A, 320 A, and 330 A) are shown in FIG. 3 . Each user-selectable capacitor is coupled in series with a user-controlled switch such as switches 300 B, 310 B, 320 B, and 330 B. The switches allow a user to set the capacitance of programmable capacitor array C 1 by selecting which of the associated capacitors is electrically coupled into the circuit. User-selectable capacitors 300 A, 310 A, 320 A, and 330 A are shown as variable capacitors, because their respective values can be adjusted or trimmed by the manufacturer.

User-selectable capacitor 300 A is shown in greater detail, and it includes fixed capacitor 301 , and trim capacitors 303 A, 304 A, 305 A, 306 A, 307 A, 308 A, and 309 A, all of which are coupled in parallel. Each of the trim capacitors is coupled in series with a manufacturer-controlled switch such as switches 303 B, 304 B, 305 B, 306 B, 307 B, 308 B, and 309 B. The values of the trim capacitors are numerically weighted (e.g., binarily weighted), and typically scaled to each have a capacitance that is some set percentage of fixed capacitor 301 . Since all of the capacitors of integrated circuit 100 are typically made using the same process, the process variations that cause the deviation of the actual capacitance of a capacitor from its nominal capacitance will effect all capacitors proportionally. Thus, trim capacitors designed to have capacitances that are set percentages of the fixed capacitor they are to trim, can all be set by the same programming bits. That is, every capacitor will deviate from its nominal value by the same percentage, so only one set of trim bits (defining one trim percentage) need be used for all user-selectable capacitors in each of the programmable capacitor arrays, so long as the values of individual trim capacitors are set as a percentage of their associated fixed capacitor. Although not shown, user-selectable capacitors can include “fixed” trim capacitors in addition to the trim capacitors coupled in series with manufacturer-controlled switches. Alternately, the trim capacitors can be designed to be automatically electronically coupled into the circuit so that the trimming process involves opening switches to remove associated trim capacitors from the circuit. Access to the manufacturer-controlled switches is usually limited by an integrated circuit programming or I/O interface command or secret key value.

Although the trimming process focuses on user-selectable capacitor values, one goal of the trimming process is to achieve desired RC products, particularly in the context of a filter circuit such as filter 130 . To that end, a typical manufacturer's trim procedure can include programming filter 130 to behave (nominally) as a specific type of filter having specific parameters. Then, one can test the filter to determine just how close to the filter is to its nominal filtering characteristics, and adjust trim capacitors accordingly. Several of such tests can be performed, and the filter characteristics can be different for each test so that a range of filter behavior can be examined.

In FIG. 4 , part of programmable capacitor array C 1 is illustrated, including user-selectable capacitors 300 A, 310 A, 320 A, an example implementation of user-controlled switches 300 B, 310 B, and 320 B (in this example, CMOS transmission gates) and corresponding signal inverters 401 , 411 , and 421 . The input of programmable capacitor array C 1 is coupled in parallel to all of the user-controlled switches 300 B, 310 B, and 320 B. The other side of each user-selectable capacitor 300 A, 310 A, 320 A is coupled to the output of programmable capacitor array C 1 . Accordingly, by turning on or off individual switches, the capacitance of programmable capacitor array C 1 can be programmably controlled. The information which controls programmable capacitor array C 1 is stored within a memory, such as memory 450 . Capacitors that are not selected within the programmable capacitor array are, in one example, connected to a reference voltage via NMOS transistors 400 , 410 , and 420 , thereby eliminating the capacitors that are not selected from any circuit interaction. A similar scheme can be used to implement the trim capacitors and their associated manufacturer-controlled switches. However, in a preferred embodiment, the manufacturer-controlled switches used for trim capacitors are chosen to be PMOS transistors because the “on” voltage for the switches is 0V, instead of some non-zero value (e.g., 5V), which eliminates one potential distortion source associated with the power supply rejection ratio (PSRR). Nevertheless, a variety of different switch implementations for both the user-controlled switches and the manufacturer-controlled switches (e g, NMOS, PMOS, CMOS, or others) will be known to those having ordinary skill in the art.

Multiplexers 402 , 412 , and 422 , along with memory 450 are used to implement the multiple filter configuration capability discussed above in conjunction with FIG. 1 . Memory 450 stores pairs of configuration values A 1 and B 1 , A 2 and B 2 , etc., in associated memory storage elements 403 and 404 , 413 and 414 , etc. The configuration values represent bits used to program the various programmable capacitor arrays. The configuration of filter 130 can be quickly changed via A/B select line 440 . Depending upon which configuration is selected, multiplexers 402 , 412 , and 422 provide values stored in memory 450 for one of the two configurations. The dual-configuration architecture need not be implemented, and thus, a single memory storage element can be associated with each user-selectable capacitor. Alternately, more than two configurations can be stored using additional memory storage elements and appropriate multiplexers. Additionally, the trim capacitors and manufacturer-controlled switches can be programmed in a manner similar to that illustrated with respect to the user-selectable capacitors. Memory 450 is preferably a static memory, such as a electronically erasable (E 2 ) memory.

Switch control values are typically loaded into memory 450 via a shift register 460 , which in turn receives input values through an I/O interface 470 . Examples of suitable I/O interfaces include the serial peripheral interface (SPI) from Motorola, Inc., and the Joint Test Access Group (JTAG) interface associated with the IEEE Standard Test Access Port and Boundary Scan Architecture (IEEE 1149.1). Use of these, and other, interfaces for programming integrated circuits is well known to those having ordinary skill in the art. However, implementing the above described manufacturer-controlled switches requires some additional mechanism, such as a secret command or key to limit or prevent user access to the trim capacitors. Additionally, the memory for storing bit values for the manufacturer-controlled switches can be a memory separate from the memory storing values for the user-selectable capacitors.

The multiple feedback paths of filter 130 and the complexity of the programmable capacitor arrays presents difficulties in testing and debugging such an integrated circuit. It is very difficult to locate a problem in the filter without breaking the feedback network. Additionally, even if the circuit performs perfectly, verifying that the programmable capacitor arrays function properly can be extremely difficult. Consequently, it is advantageous to implement a special capacitor array test mode in the circuit for filter 130 . There are two basic elements to implementing the capacitor array test mode: (1) the ability to disable one or more of the ladder filter amplifiers A 1 , A 2 , A 3 , A 4 and A 5 (i.e., to “tristate” the amplifiers) thereby deactivating certain portions of the circuit; (2) an analog interconnect array (AIC) built around the filter to reconfigure the filter into an integrator to carry-out capacitor array tests. For example, one can configure an integrator with amplifier A 5 as the operational amplifier, the capacitor array under test as the feedback capacitor, and the resistors in the filter and the AIC serving as feedback and feed forward resistors. The remaining amplifiers are tristated during the test. A test signal, for example a step stimulus, is brought to the input of amplifier A 5 from an input pin. By measuring the step response of the integrator (e.g., the rise time) associated with each of the capacitors in a capacitor array, one can determine if the capacitor array is behaving as designed.

FIG. 5 illustrates a simplified circuit topology (based on filter 130 ) used for testing programmable capacitor array C 1 . Note that in this circuit, the differential nature of filter 130 is illustrated. Programmable capacitor array C 1 is the feedback capacitor for the integrator of FIG. 5 . Feedback resistance R x and capacitance C x represent the cumulative effect of portions of the filter circuit, and are therefore fixed. Turning on and off each of the capacitors in programmable capacitor array C 1 , one after another (starting from smallest to largest) and measuring the step response, one can determine if the capacitors are functioning, and if their values are in the expected range. Another method, particularly useful for very small capacitors, is to measure the total capacitance of programmable capacitor array C 1 by turning off all of the capacitors in the array, and adding them into the circuit one at a time. Input resistors R in can be selected, for example switched in as part of the AIC, to insure stability of the circuit and proper signal response. Other topologies, similar to that illustrated in FIG. 5 , for testing other programmable capacitor arrays in filter circuit 130 are readily apparent to those having ordinary skill in the art.

Like many other analog circuits, filter 130 is susceptible to output offset voltages due to, for example, component mismatch or component variations associated with integrated circuit fabrication processes. To overcome such offset voltages, an offset calibration feature can be included in the circuit. FIG. 6 illustrates an example of such a calibration feature associated with amplifier A 5 of filter 130 , that is the last amplifier of the filter. Calibration circuit 600 senses the output offset voltage when the entire circuit is placed in a calibration mode, develops a current to correct for the offset voltage, and injects the correction current into amplifier A 5 . One example of a calibration circuit 600 is described in U.S. patent application Ser. No. 09/053,251, entitled “Offset Voltage Calibration DAC with Reduced Sensitivity to Mismatch Errors,” by James L. Gorecki and Yaohua Yang, which is incorporated herein by reference in its entirety. Calibration mode can be initiated by, for example, a circuit power on signal asserted on input line 610 , or a calibrate signal asserted on input line 620 .

The description of the invention set forth herein is illustrative and is not intended to limit the scope of the invention as set forth in the following claims. Variations and modifications of the embodiments disclosed herein may be made based on the description set forth herein, without departing from the scope and spirit of the invention as set forth in the following claims.

Claims

1. A programmable capacitor array comprising:a plurality of user-selectable capacitors, each of the plurality of user-selectable capacitors being coupled in series with an associated user-controlled switch, and each of the plurality of user-selectable capacitors and associated user-controlled switches being coupled in parallel with each other of the plurality of user-selectable capacitors and associated user-controlled switches, wherein at least one of the user-selectable capacitors includes: at least one trim capacitor coupled in series with a manufacturer-controlled switch; and at least one fixed capacitor coupled in parallel with the at least one trim capacitor and the manufacturer-controlled switch.

2. The programmable capacitor array of claim 1 wherein the plurality of user-selectable capacitors are numerically weighted.

3. The programmable capacitor array of claim 2 wherein the plurality of user-selectable capacitors are binarily weighted.

4. The programmable capacitor array of claim 1 coupled into a fully differential circuit including at least one pair of differential input terminals and at least one pair of differential output terminals.

5. A programmable analog circuit comprising:an amplifier having at least one input terminal and at least one output terminal; and a programmable capacitor array coupled between the at least one input terminal and the at least one output terminal, the programmable capacitor array including: a plurality of user-selectable capacitors, each of the plurality of user-selectable capacitors being coupled in series with an associated user-controlled switch, and each of the plurality of user-selectable capacitors and associated user-controlled switches being coupled in parallel with each other of the plurality of user-selectable capacitors and associated user-controlled switches, wherein at least one of the user-selectable capacitors includes: at least one trim capacitor coupled in series with a manufacturer-controlled switch; and at least one fixed capacitor coupled in parallel with the at least one trim capacitor and the manufacturer-controlled switch.

6. The programmable analog circuit of claim 5 further comprising at least one resistor, the at least one resistor coupled to the amplifier and to the programmable capacitor array to form an active filter.

7. The programmable analog circuit of claim 6 wherein the active filter is one of a low pass filter, a high pass filter, and a band pass filter.

8. The programmable analog circuit of claim 6 wherein the active filter is one of a Chebyshev, Bessel, Gaussian, Legendre, elliptic, and linear phase equi-ripple filter depending upon a programmed capacitance of the programmable capacitor array.

9. The programmable analog circuit of claim 5 wherein the plurality of user-selectable capacitors are numerically weighted.

10. The programmable analog circuit of claim 9 wherein the plurality of user-selectable capacitors are binarily weighted.

11. The programmable analog circuit of claim 5 wherein the amplifier is a differential amplifier, the at least one input terminal includes a pair of differential input terminals, and the at least one output terminal includes a pair of differential output terminals.

12. The programmable analog circuit of claim 11 further comprising:a feedback amplifier including a pair of differential input terminals and a pair of differential output terminals, the output terminals of the feedback amplifier being coupled the input terminals of the differential amplifier.

13. The programmable analog circuit of claim 11 further comprising:a first input resistor coupled to a first one of the pair of differential input terminals of the feedback amplifier; and a second input resistor coupled to a second one of the pair of differential input terminals of the feedback amplifier.

14. The programmable analog circuit of claim 11 further comprising:a first feedback resistor coupled between a first one of the pair of differential input terminals of the feedback amplifier and a first one of the pair of differential output terminals of the feedback amplifier; and a second feedback resistor coupled between a second one of the pair of differential input terminals of the feedback amplifier and a second one of the pair of differential output terminals of the feedback amplifier.

15. A programmable analog circuit comprising:a differential input amplifier; a differential feedback amplifier coupled to the differential input amplifier; a differential filter amplifier coupled to the differential feedback amplifier and having a pair of input terminals and a pair of output terminals; and a programmable capacitor array coupled between at least one of the pair of input terminals of the differential filter amplifier and at least one of the pair of output terminals of the differential filter amplifier.

16. The programmable analog circuit of claim 15 wherein the programmable capacitor array includes:a plurality of user-selectable capacitors, each of the plurality of user-selectable capacitors being coupled in series with an associated user-controlled switch, and each of the plurality of user-selectable capacitors and associated user-controlled switches being coupled in parallel with each other of the plurality of user-selectable capacitors and associated user-controlled switches, wherein at least one of the user-selectable capacitors includes: at least one trim capacitor coupled in series with a manufacturer-controlled switch; and at least one fixed capacitor coupled in parallel with the at least one trim capacitor and the manufacturer-controlled switch.

17. The programmable analog circuit of claim 15 wherein the differential feedback amplifier includes a pair of differential input terminals and a pair of differential output terminals, the output terminals of the differential feedback amplifier being coupled the input terminals of the differential filter amplifier.

18. The programmable analog circuit of claim 17 further comprising:a first input resistor coupled between a first one of the pair of differential input terminals of the feedback amplifier and a first output terminal of the differential input amplifier; and a second input resistor coupled between a second one of the pair of differential input terminals of the feedback amplifier and a second output terminal of the differential input amplifier.

19. The programmable analog circuit of claim 17 further comprising:a first feedback resistor coupled between a first one of the pair of differential input terminals of the feedback amplifier and a first one of the pair of differential output terminals of the feedback amplifier; and a second feedback resistor coupled between a second one of the pair of differential input terminals of the feedback amplifier and a second one of the pair of differential output terminals of the feedback amplifier.