Self-Zeroing Amplifier

This invention relates generally to amplifiers and more particularly to offset adjustment circuitry in an amplifier.

Many applications require precision amplifiers. For example, data acquisition systems demand very high precision amplifiers.

One type of error in commercially available amplifiers is offset voltage. In a perfect amplifier, the output is zero when the input is zero. In a real amplifier, the output may not be precisely zero even when the input is zero. This deviation from a perfect amplifier is quantified by measuring the "offset voltage" of a real amplifier. With the output of the amplifier at zero volts, the voltage at the input is the offset voltage.

To reduce the problem of offset voltage, many commercially available integrated circuit (IC) amplifiers have pins to which a potentiometer can be connected. Adjusting the resistance of the potentiometer changes the offset voltage. If the potentiometer is properly adjusted, the offset voltage can be set very near zero or "nulled".

Several drawbacks of nulling the offset voltage with a potentiometer exist. A person usually performs the operation, which is often not convenient or cost-effective. Additionally, the offset voltage may change as the operating temperature of the amplifier changes or the amplifier ages. If the offset voltage changes, the amplifier is no longer nulled and an error is introduced.

A method of dynamic offset nulling has been used to avoid errors when the offset voltage changes. This method is called "chopper stabilization". The input to the amplifier is switched rapidly between an input signal and ground. A capacitor is connected to the output. When the input is switched to ground, the capacitor is switched between the output and ground. In this state, the output is at the offset voltage and the capacitor changes to the offset voltage. When the input to the amplifier is switched to the input signal, the capacitor is switched to be in series with the output. The capacitor is switched such that the voltage across the capacitor is subtracted from the output of the amplifier, thereby guaranteeing that the output is at zero volts when the input is at zero volts.

With chopper stabilization, the voltage across the capacitor adjusts if the offset voltage changes. Thus, the errors of adjusting the offset with a potentiometer do not occur. However, chopper stabilization has other drawbacks. Both the input and output constantly switch between the desired output signal and an undefined state. This rapid switching causes discontinuities in the output signal. The discontinuities can be filtered out, but only if the desired signal is of relatively low frequency. Additionally, the rapid switching creates high frequency noise in a system, which can be undesirable.

It is an object of this invention to provide a precise amplifier.

It is a further object of this invention to provide an amplifier with offset adjustment that can compensate for changes in the offset voltage.

It is yet another object of this invention to provide an integrated circuit chip which can compensate for offset voltage in a system.

The foregoing and other objects are accomplished in an integrated circuit containing an amplifier, control logic, a digital to analog converter (DAC), a counter, and switches. In response to a control input, the control logic switches the inputs of the amplifier to ground and resets the counter. The control logic then increments the digital value applied to the digital to analog converter. The output of the digital to analog converter is connected to the amplifier such that an increase in the digital input to the DAC changes the offset voltage of the amplifier. The digital input to the DAC is incremented by the counter until the control logic determines that the output of the amplifier changes from positive to negative. The control logic then latches the value of the digital input and disconnects the inputs to the amplifier from ground and reconnects them to the input signal.

In one embodiment of the invention, the switches do not directly connect the inputs of the amplifier to ground. Rather, integrated circuit package pins are provided for both the switches and the amplifier inputs. This arrangement allows circuitry to be included between the switch and the amplifier inputs. Any offset voltage generated by that circuitry can also be nulled.

In another embodiment of the invention, the amplifier is constructed on an integrated circuit chip. The logic section contains a comparator which indicates the offset voltage has been nulled when its output changes from positive to negative. The inputs of the comparator are inputs to the integrated circuit chip, allowing any signals to control when the offset voltage has been nulled. This arrangement allows offset induced by components outside the IC chip connected to the output of the amplifier to be nulled.

This invention will be better understood by reference to the following detailed description and accompanying figures, in which

FIG. 1 is a simplified block diagram of an amplifier constructed in accordance with the present invention;

FIG. 2 is a simplified schematic of the offset adjustment features of the amplifier in FIG. 1;

FIG. 3 is a simplified block diagram of an alternative embodiment of the invention;

FIG. 4A is a block diagram indicating how the amplifier of FIG. 3 can be connected to null offset voltage introduced by components connected to the input of the amplifier; and

FIG. 4B is a block diagram indicating how the amplifier of FIG. 3 can be connected to null offset voltage introduced by components connected to the input of the amplifier.

FIG. 1 shows a block diagram of circuitry which forms a precision amplifier 28. Precision amplifier 28 is constructed from known technology. For example, bipolar junction transistors might be connected to form an integrated circuit (IC) chip. The lines P.sub.IN+, P.sub.IN-, C.sub.EXT, CALIBRATE, and V.sub.OUT represent signal lines connected to pins outside the package containing the IC chip.

One of skill in the art will appreciate that an IC package contains more pins than are explicitly shown. Pins to provide standard functions in IC circuits are not explicitly shown. For example, no pins are shown for providing power to the IC, but one of skill in the art will recognize that such pins are required. Other standard elements of an IC circuit are not shown, but it will be appreciated that such elements are also required.

Precision amplifier 28 contains an amplifier 10 of known construction. The inputs of amplifier 10 are connected,

via switches 22 and 24, to input pins P.sub.IN+, and P.sub.IN-. The output of amplifier 10 is connected to output pin V.sub.OUT.

Amplifier 10 is also connected to digital to analog converter (DAC) 12 in a manner described in greater detail in conjunction with FIG. 2. Suffice it to say here that the output of DAC 12 is a current which changes the offset voltage of amplifier 10. Digital to analog converters are well known. Here, DAC 12 is a six-bit converter of known construction.

The digital input of DAC 12 is connected to the output of counter 14. Here, counter 14 is a six-bit counter of known construction. As is known, each time a pulse occurs on count input 14.sub.c, the digital value stored in counter 14 increases by 1. Counter 14 has latched outputs so the values stored in the counter appears on its output lines until it is changed. Counter 14 is also connected to logic 18 by control input 14.sub.R. A pulse on control input 14R causes the value stored in counter 14 to be reset to zero.

Counter 14 contains an overflow output 14.sub.O. Counter 14, having a fixed number of bits, can only count a fixed number of pulses. When counter 14 reaches its maximum i/15 count, overflow output 14.sub.O has a value of logic HI. This indicates to logic 18 that counter 14 cannot be incremented further.

The count input 14.sub.C to counter 14 is provided by clock 16. Clock 16 is of known construction and produces a pulse train. The period of the pulse train is dictated by the size of capacitor C.sub.1 which is outside of the IC package containing precision amplifier 28. Capacitor C.sub.1 is connected to clock 16 via pin C.sub.EXT. Here, C.sub.1 is selected to produce a pulse train with a frequency of roughly 200 KHz. Clock 16 also has a control input 16.sub.C. Clock 16 only produces a pulse train when control input 16.sub.C is in a logic HI state.

Control signals to clock 16 and counter 14 are provided by logic 18. Logic 18 contains known circuits configured to perform the logic functions described below.

Another input to logic 18 is provided by comparator 20. Comparator 20 is of known construction. When the value at input 20.sub.+ exceeds the value at input 20.sub.-, the output 20.sub.O has a logic HI value. When the value at input 20.sub.+ is less than the value at input 20.sub.-, the output 200 has a logic LO value. As shown in FIG. 1, input 20.sub.- is connected to ground and input 20.sub.+ is connected to the output of amplifier 10. Thus, the output 200 is a logic HI when the output of amplifier 10 is positive and a logic LO when the output of amplifier 10 is negative. The output 200 switches from logic LO to logic HI (a "LO to HI transition") when the output of amplifier 10 increases until it equals zero.

Actually, comparator 20 is designed to have some hysteresis. This means that input 20.sub.+ must slightly exceed input 20.sub.- for output 200 to switch from LO to HI. Likewise, input 20.sub.- must slightly exceed input 20.sub.+ for output 200 to switch from HI to LO. The hysteresis prevents comparator 20 from generating false signals when there is noise in the system.

In operation, switches 22 and 24 normally couple the inputs of amplifier 10 to the pins P.sub.IN+ and P.sub.IN-. Logic 18 is inactive and precision amplifier 28 operates, in this normal state, like an amplifier of the prior art.

Precision amplifier 28 may be placed in a "calibrate mode" of operation by a logic HI signal on input pin CALIBRATE. In the calibrate mode, the input offset voltage is nulled. When a logic HI signal should be placed on the CALIBRATE pin is determined by the use of precision amplifier 28. For example, if precision amplifier 28 is used in a microprocessor controlled data acquisition system, the microprocessor (not shown) will apply a logic HI to the CALIBRATE pin at appropriate times. Appropriate times might be when power is first applied to the system or before the system acquires data, or after every five minutes of operation.

In the calibrate mode, logic 18 controls the various components of precision amplifier 28 to null the offset. First, logic 18 operates switches 22 and 24 to connect them to their grounded terminals. This connects the inputs of amplifier 10 to ground. Logic 18 also generates a signal on line 14.sub.R to reset counter 14 to zero.

Logic 18 then places a signal on control input 16.sub.C to start clock 16 producing a pulse train. At each pulse, the value in counter 14 increments and the current out of DAC 12 changes. The change in current out of DAC 12 in turn changes the output voltage of amplifier 10. When the current from DAC 12 reaches a value which nulls the offset, output 20.sub.O of comparator 20 has a LO to HI transition.

Logic 18 detects this transition and places a signal on line 16.sub.C which stops clock 16 from generating a pulse train. Logic 18 then generates signals which connect switches 22 and 24 as they were in the normal mode. Precision amplifier 28 is then operating as a normal amplifier except that the current from DAC 12 nulls the offset voltage.

It is possible that the maximum value in counter 14 will not completely null the offset voltage. In that case, the counter will reach its maximum value and the overflow line 14.sub.O will go HI. If logic 18 detects a logic HI on overflow line 14.sub.O, it will return precision amplifier to the normal mode in the same fashion as when comparator 20 indicated the offset voltage had been nulled. However, when the calibration mode ends because of the value of overflow line 14.sub.O, the offset is not completely nulled. Rather, it is as close to being nulled as the circuitry of precision amplifier 28 can achieve.

Turning now to FIG. 2, additional details can be seen of how DAC 12 controls the offset voltage. FIG. 2 shows in simplified schematic form the input stage 202 of the amplifier 10. Input stage 202 is coupled to a second stage 204 which in turn is coupled to the output pin V.sub.OUT. Second stage 204 is any known second stage of an amplifier. Input stage 202 contains a pair of bipolar junction transistors Q.sub.1 and Q.sub.2. The collectors of transistors Q.sub.1 and Q.sub.2 are connected to the positive supply V.sub.CC through resistors R.sub.1L, R.sub.1M, R.sub.2E and R.sub.2G. The emitters of transistors Q.sub.1 and Q.sub.2 are connected together and connected to a lead of current source 200.

Input stage 202 is a traditional differential input stage. Current source 200 produces a current I.sub.l. The current I.sub.l splits such that part flows through transistor Q.sub.1 and part flows through transistor Q.sub.2 . Ideally, when more current flows through transistor Q.sub.2 , the voltage at line V.sub.C+ exceeds the voltage on line V.sub.C-. Conversely, when the current through transistor Q.sub.1 exceeds the current through transistor Q.sub.2 , the voltage at line V.sub.C- exceeds the voltage at line V.sub.C-. In this way, the output voltage changes in response to the currents through transistors Q.sub.1 and Q.sub.2 .

When the currents through transistors Q.sub.1 and Q.sub.2 are equal, the output voltage on pin V.sub.OUT would be zero if the resistors R.sub.1M and R.sub.2E, R.sub.1L and R.sub.2G and transistors Q.sub.1 and Q.sub.2 were identical (sometimes called "perfectly matched"). If the components of the amplifier were perfectly matched, a zero differential voltage between pins P.sub.IN+ and P.sub.IN- would produce the same current through transistors Q.sub.1 and Q.sub.2 and hence a zero output voltage. Of course, the components of precision amplifier 28 cannot be perfectly matched and some voltage will result on pin V.sub.OUT even if the input voltage is zero. This resulting voltage represents offset voltage.

As described above, the voltage at pin V.sub.OUT can be adjusted by adjusting the relative currents between transistors Q.sub.1 and Q.sub.2 . DAC 12 is connected to input stage 202 to adjust those currents. As seen, DAC 12 has two current outputs I.sub.OUT and I.sub.BOUT and one current input I.sub.IN. DAC 12 works in a conventional fashion. Namely, the input current I.sub.IN is divided between the output lines I.sub.OUT and I.sub.BOUT in proportion to the value of the digital input from counter 14. If the digital input is zero, the current I.sub.IN will flow through output I.sub.BOUT and no current will flow through line I.sub.OUT When the digital input has its maximum value, the current I.sub.IN will flow through line I.sub.OUT and no current will flow through line I.sub.BOUT. When the digital input has its middle value, equal amounts of current flow through lines I.sub.OUT and I.sub.BOUT. For other values of the digital input, some current will flow through each line in proportion to the value.

The manner in which DAC 12 can adjust the offset voltage can thus be seen. The value of current I.sub.IN is selected so that when applied to input stage 202 on line I.sub.BOUT, it lowers the output voltage by at least the amount of the expected offset voltage to be nulled. Thus, when counter 14 has a value of zero after precision amplifier 28 enters calibrate mode, the output V.sub.OUT should be less than the desired value. When counter 14 counts up, the output V.sub.OUT should increase to the desired value.

FIG. 2 also shows other details of the construction of precision amplifier 28. Current source 200 is seen to be implemented with a circuit known as a current mirror. The base of transistor Q.sub.93 is connected to a precise voltage reference V.sub.REF implemented in any known manner. The collector of transistor Q.sub.93 is connected to the positive supply V.sub.CC. The emitter of transistor Q.sub.93 is connected to ground via resistor R.sub.93. The current through transistor Q.sub.93 is relatively precisely controlled. Since the bases of transistors Q.sub.18 and Q.sub.48 are tied to the base of transistor Q.sub.93, the currents through transistors Q.sub.18 and Q.sub.48 "mirror" the current in transistor Q.sub.93. The transistors Q.sub.18, Q.sub.48 and Q.sub.93 and resistors R.sub.3, R.sub.56 and R.sub.93 can be made different sizes such that the currents in each transistor are different. However, the currents in transistors Q.sub.18, Q.sub.48 and Q.sub.93 will be relatively stable and, if any of the currents change, the rest of the currents exhibit proportionate change.

This proportionate change is significant because currents I.sub.l and I.sub.IN are both produced by current source 200. If current I.sub.l were to change, say because precision amplifier 28 heated up during use, the offset voltage might increase. Here, however, an increase in I.sub.l is mirrored by an increase in I.sub.IN, which increases the offset compensation. Thus, the offset voltage tends to stay nulled as precision amplifier 28 operates.

FIG. 2 also shows details of the construction of switches 22 and 24. Here, the switches are implemented using junction field effect transistors J.sub.51, J.sub.52, J.sub.53, J.sub.54.

Turning now to FIG. 3, an additional feature of the present invention may be seen. When precision amplifier 28 is fabricated on an IC chip, the inputs to comparator 20 and switches 22 and 24 can be connected to pins on the package containing the IC chip. These connections, as will be describe in conjunction with FIGS. 4A and 4B, allow precision amplifier 28 to null offsets created by sources external to precision amplifier 28.

The inputs of amplifier 10 are connected to pins IN.sub.+ and IN.sub.- via lines 317 and 318. Switches 22 and 24 are connected to pins S.sub.1 and S.sub.2. In FIG. 1, lines 317 and 318 are not present. Rather, the same inputs of amplifier 10 are connected to switches 22 and 24. To connect the inputs to switch 22 and 24, the gaps 302 and 306 could be filled and lines 317, 318, 304 and 308 could be omitted.

Likewise in FIG. 3, the inputs of comparator 20 are connected to pins COMP.sub.` and COMP.sub.+. In FIG. 1, the inputs of comparator 20 are connected to V.sub..sub.OUT and ground. In FIG. 3, the connection to the pins is made by lines 312 and 316. To connect the inputs of comparator 20 to ground and V.sub..sub.OUT, the lines 312 and 316 can be omitted and gaps 310 and 314 can be filled.

In integrated circuit fabrication, selective connection of certain portions of a circuit is well known. Integrated circuits are customized by changing the final metallization layer in which many components are connected. To fill a gap, a line is added to the metallization. To omit a line, the corresponding portion of the metallization is removed. In addition, when an integrated circuit chip is connected to pins of a package (a process called "bonding"), different portions of the circuit can be connected to the pins. In comparison to the expense and effort needed to fabricate an integrated circuit, small changes in the last metallization layer or the bonding process are simple.

The advantages of configuring precision amplifier 28 as shown in FIG. 3 can be seen by reference to FIGS. 4A and 4B. FIG. 4A shows precision amplifier 28A used to amplify the output of some circuits. For example, precision amplifier 28A could be used to amplify the output of a strain gauge sensor 400.

The offset voltage at pin V.sub.OUT could be due to imprecise matching of components inside the precision amplifier 28. In addition, sensor 400 could add a component to the offset voltage. To fully null the offset, precision amplifier 28A should compensate for both sources of error.

Since precision amplifier 28 is only placed in calibrate mode by an appropriate signal on the CALIBRATE line, it is not always necessary to connect the inputs of amplifier 10 to ground when nulling an offset voltage. In particular, if precision amplifier 28 is only placed in calibrate mode when the input to the amplifier is supposed to be zero, there is no need to ground the inputs. Consequently, strain gauge sensor 400 is connected directly to inputs IN+ and IN.sub.- rather than being connected via switches 22 and 24 (FIG. 3). During the nulling, then, strain gauge sensor is connected to the inputs of amplifier 10 (FIG. 3) and any offset introduced by that component is nulled.

The extent to which precision amplifier 28A can be used to null offset voltage attributed to circuits connected to the inputs of precision amplifier 28A depends on the extent to which a time can be identified when the expected output of precision amplifier 28A is known. In the circuit of FIG. 4A, that known output value is zero and the microprocessor or the other control means connected to the CALIBRATE line does not initiate calibration mode unless the system incorporating precision amplifier 28A is in a state which should produce a zero output.

It is also possible to null the offset of precision amplifier 28A when the expected output voltage of precision amplifier 28A is other than zero. From the foregoing description, it can be seen that comparator 20 (FIG. 3) indicates precision amplifier 28A is nulled when the voltage at pin COMP.sub.- equals the voltage at pin COMP.sub.+. FIG. 4B shows a resistor voltage divider comprising resistors R.sub.4A and R.sub.4B used to place a voltage of V.sub.NULL at pin COMP.sub.-. For the circuit of FIG. 4B, precision amplifier 28A is placed in calibrate mode only when the voltage connected to pin COMP.sub.+ is expected to equal V.sub.NULL.

When amplifier 28A is used in the configuration of FIG. 4B, the voltage applied to pin COMP.sub.+ is from the output of circuitry 402 connected to pin V.sub.OUT. In this fashion, precision amplifier 28A can null offset introduced by circuitry 402 in addition to offset introduced by precision amplifier 28A.

It should be noted in FIG. 4B that inputs to precision amplifier 28A are connected to pins IN.sub.+ and IN.sub.- via switches 22 and 24. The connections shown in FIG. 4B achieve the same connections between pins P.sub.IN+ and P.sub.IN- and amplifier 10 as shown in FIG. 1.

Having described several embodiments of this invention, it will be apparent to one of skill in the art that various modifications can be made to the disclosed embodiment. For example, alternative layouts of input stage or switches could be employed. It is felt, therefore, that this invention should be limited only by the spirit and scope of the appended claims.