A/D Self-Testing Circuit

This invention relates generally to analog to digital converters, and more specifically, to successive approximation analog to digital converters which may be readily tested for accuracy.

Typically, analog to digital (A/D) converters are periodically tested in a system environment to determine if the A/D converter is functioning properly. An A/D converter can be tested by sampling a known voltage and determining if an accurate converted output is provided. The problem previously associated with tests of A/D converters have been related to the accuracy of the known voltage which is used to perform the test. A sampling voltage of the required accuracy is generally difficult to generate. Typical resistor voltage divider circuits do not provide a sampling voltage which is accurate enough. When conventional precision voltage reference circuits are used, substantial circuit size is dedicated for the single purpose of a test voltage function.

Accordingly, an object of the present invention is to provide an improved A/D self-testing converter.

Another object of the present invention is to provide an improved A/D converter having an accurate test voltage used to test the accuracy of the converter.

Yet another object of the present invention is to provide a self-testing A/D converter which minimizes circuit area and power consumption.

In carrying out the above and other objects of the present invention, there is provided, in one form, a circuit for testing the functionality of an analog to digital (A/D) converter system. The test circuit is used in conjunction with the system which operates in a conventional manner when the test circuit is not utilized. The system has an accurate reference voltage associated therewith and a capacitive digital to analog converter (DAC). The digital to analog converter has a plurality of rank ordered capacitors, all of the capacitors having first and second electrodes. The first electrodes of the capacitors are coupled together for providing a DAC output which is selectively coupled to a D.C. bias voltage. A first plurality of rank ordered switching means is provided with each switching means having a first terminal coupled to a predetermined second electrode of the capacitors of corresponding rank. The first terminal of each switch is selectively coupled between the accurate system reference voltage, a second reference voltage, a terminal coupled to an unknown voltage, and a successive approximation register (SAR) terminal. A first plurality of control signals controls the switching of the first plurality of switches. A second plurality of rank ordered switches is coupled to the successive approximation register and selectively couples the first and second reference voltages to a respective SAR terminal in response to a control signal provided by a successive approximation register. A digital result is provided by the successive approximation register which may be readily used to determine the functionality of the system.

The single FIGURE is a schematic diagram of a self-testing A/D converter constructed in accordance with the present invention.

Shown in the Figure is a binarily weighted capacitive D/A converter (DAC) 10 which is used within a successive approximation A/D converter. DAC 10 generally comprises a capacitive portion 11 and a switching portion 12. CApacitive portion 11 has a plurality of low ordered capacitors 14, 15, 16, 17 and 18 which in a preferred form are weighted 1, 1, 2, 4 and 8 units, respectively. Capacitors 14, 15, 16, 17 and 18 each have an electrode connected together at a node 19. A scaling capacitor 21 has a first electrode connected to node 19. In a preferred form, capacitor 21 is weighted 1.1 units. Capacitive portion 11 also has a plurality of high ordered capacitors 23, 24, 25 and 26 which in a preferred form are weighted 1, 2, 4 and 8 units, respectively. Capacitors 23, 24, 25 and 26 each have a first electrode connected together and to a second electrode of scaling capacitor 21 at a node 28 for providing the DAC output. A switch 29 has a first terminal connected to node 28 and a second terminal connected to a D.C. bias voltage source, V.sub.DC.

Switching portion 12 comprises a plurality of switches 31, 32, 33, 34, 35, 36, 37, 38 and 39 which have a first terminal connected to a second electrode of capacitors 14, 15, 16, 17, 18, 23, 24, 25 and 26, respectively. Switches 31-39 are each selectively switched between four terminals labeled "SAR Bit", "V.sub.X ", "V.sub.AG " and "V.sub.REF " in response to a first plurality of control signals provided by conventional external logic (not shown) which controls DAC 10 in the manner described below. The label "V.sub.REF " represents a known accurate system reference voltage which is associated with every A/D converter system. The label "V.sub.AG " represents an analog ground voltage and the label "V.sub.X " is an unknown sample voltage. The "SAR BIT" terminal of switches 31-39 is connected to a first terminal of a plurality of switches 50, 51, 52, 53, 54, 55, 56, 57 and 58, respectively. The first terminals of switches 50-58 are each coupled between second and third terminals labeled "V.sub.AG " (analog ground voltage) and "V.sub.REF " (system reference voltage), respectively, in response to a second plurality of control signals provided by a successive approximation register (SAR) 60 via conductors 61, 62, 63, 64, 65, 66, 67, 68 and 69, respectively. Although switching portion 12 is shown for the purpose of illustration as having individual switches 31-39 and 50-58 with each switch being a multiple position switch, it should be obvious that the functionality of each of switches 31-39 may be obtained by using multiple switches having other numbers of switching terminals. Furthermore, any type of switch such as a conventional transistor transmission gate may be used to implement switches 31-39 and 50-58.

In operation, DAC 10 operates in a conversion mode and a self-test mode. DAC 10 functions in a conventional manner when in the conversion mode. In particular, in the conversion mode an unknown analog voltage V.sub.X which is to be converted to a digital output by the successive approximation A/D converter is sampled onto all the capacitors 14-18 and 23-26 while switch 29 is conductive. Switch 29 establishes a known D.C. bias voltage, V.sub.DC, at output node 28 of DAC 10 while the second electrodes of capacitors 14-18 and 23-26 are charged to an unknown potential (V.sub.X). Since binarily weighted DACs are well known in the art, a discussion of the effects of the associated capacitor weights will not be given. It should be noted that the values of the capacitor weights are irrelevant for purposes of the present invention and the values shown in the Figure are for purposes of illustration only. Furthermore, although conventional half-bit correction is not illustrated in the Figure, it should be readily apparent that the present invention may be used with DACs which have half-bit correction. After a predetermined settling time, switches 31-39 are each switched from V.sub.X to the "SAR Bit" to be coupled to either first or second reference voltages, V.sub.AG or V.sub.REF, respectively, in response to the second plurality of rank ordered control signals provided by SAR 60 and associated successive approximation circuitry of the A/D converter. Simultaneously, switch 29 is made nonconductive. At this point in time, the voltage output at node 28 is proportional to the difference in the unknown analog voltage V.sub.X and the binarily weighted equivalent voltage represented by the successive approximation output referenced to V.sub.DC. The successive approximation circuitry determines whether the closest digital value of the unknown analog input voltage has been provided and SAR 60 provides a control signal which is either a logic "1" or "0" to each of switches 50-58. The control signals cause switches 50-58 to couple either V.sub.AG or V.sub.REF to the respective second electrodes of capacitors 14-18 and 23-26 thru switches 31-39, respectively. It should be well understood that the logic level of each of the control signals for switches 50-58 functions as an A/D output bit and correlates to which reference voltage switches 50-58 are individually coupled to. After the successive approximation circuitry has determined that a digital output which most closely approximates the unknown analog input voltage has been provided, SAR 60 provides a binarily weighted digital output corresponding to the unknown analog input voltage at the terminals labeled "A/D Output".

In the self-test mode, DAC 10 is initially charged to the known accurate system reference voltage by switching the reference voltage V.sub.REF to predetermined ones of capacitors 14-18 and 23-26 while all other capacitors are switched to V.sub.AG. The actual switching which occurs results from an arbitrary test code which is chosen by a user of the test mode. The resulting switching establishes a sampled voltage on capacitive portion 11 which is equal to the product of a ratio and V.sub.REF. The ratio is the ratio of the weighted capacitive sum of the capacitors switched to V.sub.REF and the effective capacitive sum of all the capacitors. During the time the self-test voltage is sampled, switch 29 is conductive so that the known D.C. bias V.sub.DC is coupled to the DAC output node 28. After a sufficient settling time, switch 29 is opened and switches 31-39 are connected to the SAR Bit terminals. The normal successive approximation conversion associated with the A/D converter proceeds from that point. If the resulting converted digital output code correlates to the arbitrary test code which was used to selectively sample V.sub.REF onto the capacitors of DAC 10, a user may quickly determine that the circuitry associated with the successive approximation A/D converter is fully functional. If the successive approximation routine is functioning properly and the A/D converter functions properly, the A/D converter should replicate the exact same sum of capacitances that are charged during the sample phase. Every combination of switching of capacitors 14-18 and 23-26 gives a unique and particular sum because the DAC is binarily weighted.

Upon close scrutiny, it should be apparent that the present invention necessitates an accurate reference voltage to be used as a test voltage because accuracy in a ratiometric data acquisition system is directly proportional to the accuracy of an associated reference voltage. However, in the present invention a test voltage is not generated or derived from a system reference voltage. Rather, the present invention only "effects" an accurate test voltage by utilizing the full value of an available accurate reference voltage which is always present in an A/D system. Instead of generating a reference test voltage or dividing a given system reference voltage by means such as a resistive divider, the present invention uses the entire system reference voltage but only selectively samples the system reference voltage onto a portion of the capacitive DAC. The entire value of the reference voltage cannot be sampled onto all of the capacitors without providing a digital output of all ones. Such an output would never indicate whether or not multiple bits may be electrically shorted together since all bits are supposed to have a digital value of one when a test value equal to the entire system reference voltage is used. Therefore, others have generated a test voltage from an accurate system reference voltage having a fractional value of the reference voltage and sampled that smaller test voltage onto all capacitors of a capacitive DAC so that a digital test result other than all ones results. The problem with such prior test circuits has been the inaccuracy of the test voltage which was derived from the reference voltage with respect to process, temperature and other variations as previously mentioned.

The present invention however utilizes switching portion 12 to effectively switch the full value of the accurate system reference voltage to only a portion of the capacitors of the DAC. A plurality of combinations of switching the capacitors exists and each combination provides a unique digital output sum. In other words, because it is a binarily weighted D/A converter, only a single unique combination of switching will give a particular sum. The plurality of effective test voltages which may be chosen provides multiple opportunities to test the digital circuitry associated with an A/D converter in a very stable manner with the same or multiple test values.

By now it should be apparent that the present invention provides a self-test circuit and method for testing the circuitry associated with an A/D system. The present invention allows the A/D system to perform a self-test so that a system host such as a microprocessor may interrogate the A/D system and determine if the system is functioning properly from a majority functionality standpoint. The present invention consumes very little circuit area since the only additional circuitry added to a conventional DAC is a portion of the switches of switching portion 12. Power consumption is also minimized as no D.C. current path is generated in association with the reference voltage utilized herein.