Image
image
image
image


Design Articles

Auto-Zero Operational Amplifiers: Inherent Benefits in Portable Signal-Conditioning Applications

This self-correcting architecture provides excellent power-supply rejection and common-mode rejection, plus low bias current and noise to portable designs.

By Kevin Tretter, Microchip

At first glance, the term “auto-zero” operational amplifier (op amp) may appear to be something new, but in reality this architectural concept has been around for decades. This article will explore the history behind auto-zero op amps and provide a high-level overview of the architecture. Additionally, the article will explore the inherent benefits of this architecture for signal-conditioning applications. Finally, an example application will be analyzed to further compare the auto-zero architecture to that of traditional op amps.

A Brief History
Chopper amplifiers have been around for decades, dating back close to 60 years. The chopper amplifier was invented to address the need for an ultra-low-offset, low-drift op amp—something that was superior to the bipolar op amps available at the time. In the original chopper amplifier, the amplifier’s input and output are switched (or chopped), causing the input signal to be modulated, corrected for offset error and then unmodulated at the output. This technique allowed for low offset voltage and low drift, but also had limitations. Since the input to the amplifier is being sampled, the input-signal frequency had to be limited to less than half of the chopping frequency in order to prevent aliasing. In addition to the bandwidth limitation, the act of chopping causes significant glitches to appear, requiring filtering on the output to smooth out the resulting ripples.

figure 1
Figure 1: Simplified Chopper-Stabilized Functional Diagram

The next generation of self-correcting amplifiers improved on the chopper amplifier by creating a chopper-stabilized op amp. This architecture uses two amplifiers—a “main” amplifier and a “null” amplifier, as shown in Figure 1. The null amplifier corrects its own offset error by shorting the inputs and applying a correction factor to its own null pin, after which it monitors and corrects the offset of the main amplifier. This architecture has a big advantage over the older chopper amplifiers, as the main amplifier is always connected to the input and output of the IC. Thus, the bandwidth of the main amplifier determines the input-signal bandwidth. Therefore, the input bandwidth is no longer dependent upon the chopping frequency. Charge injection from the switching action is still an issue, which can cause transients and can couple with the input signal, causing intermodulation distortion.

The auto-zero architecture is similar in concept to that of a chopper-stabilized amplifier in that there is a nulling amplifier and a main amplifier. However, significant improvements have been made over the years to minimize noise, charge injection and other performance issues associated with chopper-stabilized op amps. Various manufacturers use different terms to define this architecture, such as “auto-zero,” “autocorrelating zeroing,” and “zero-drift.” Regardless of the terminology, the basic underlying architecture is the same.

Advantages of the Auto-Zero Architecture
As described above, the auto-zero architecture continually self corrects for the offset-voltage error of the amplifier. This results in several distinct advantages over traditional op amps.

Low Offset Voltage
The nulling amplifier continually cancels its own offset voltage, and then applies a correction factor to the main amplifier. The frequency of this correction varies depending upon the actual design, but typically occurs thousands of times per second. For example, the MCP6V01 auto-zero amplifier from Microchip Technology corrects the main amplifier every 100 μs, or 10,000 times each second. This continual correction allows for ultra-low offset voltages that are much lower than traditional op amps. Additionally, the process of correcting the offset voltage also corrects other DC specifications, such as power-supply rejection and common-mode rejection. Therefore, auto-zero amplifiers are able to achieve superior rejection to that of traditional amplifiers.

Low Drift Over Temperature and Time
All amplifiers, regardless of process technology and architecture, have an offset voltage that changes over temperature and time. Most op amps specify this offset drift over temperature in terms of volts per degree Celsius. This drift can vary substantially from amplifier to amplifier, but for a traditional amplifier is typically on the order of several micro-volts to tens of micro-volts per degree Celsius. This offset drift can be very problematic in high-precision applications; unlike initial offset errors, this drift cannot be accounted for with a one-time system calibration.

In addition to drifting over temperature, an amplifier’s offset voltage tends to change over time, as well. For traditional op amps, this drift over time (sometimes called aging) typically isn’t specified in the datasheet, but it can create significant errors over the life of the device.

The auto-zero architecture inherently minimizes both the drift over temperature and time by continually self correcting the offset voltage. In this way, an auto-zero amplifier can achieve significantly better drift performance over traditional op amps. For example, the MCP6V01 op amp mentioned previously has a maximum temperature drift of only 50 nV/°C.

Eliminates 1/f Noise
1/f noise, or flicker noise, is a low-frequency phenomenon caused by irregularities in the conduction path and noise due to the bias currents within the transistors. At higher frequencies, 1/f noise is negligible as the white noise from other sources begins to dominate. This low-frequency noise can be very problematic if the input signal is near DC, such as the outputs from strain gauges, pressure sensors, thermocouples, etc.

In an auto-zero based amplifier, the 1/f noise is removed as part of the offset-correction process. This noise source appears at the input and is relatively slow moving, hence it appears to be a part of the amplifier’s offset and gets compensated accordingly.

Low Bias Current
Bias current is the amount of current flow into the inputs of the amplifier to bias the input transistors. The magnitude of this current can vary from µA down to pA, and is strongly dependent upon the architecture of the amplifier-input circuitry. This parameter becomes extremely important when connecting a high-impedance sensor to the input of an amplifier. As the bias current flows through this high impedance, a voltage drop occurs across the impedance, resulting in a voltage error. For these applications, a low bias current is required.

Virtually all auto-zero amplifiers on the market today implement a CMOS input stage, which results in very low bias currents. However, the charge injection from the internal switching can result in slightly higher bias currents then that of a more traditional, CMOS-input op amp.

Quiescent Current
For battery-powered applications, quiescent current is a critical parameter. Because of the nulling amplifiers and other circuitry required to support the self-correcting auto-zero architecture, auto-zero amplifiers typically consume more quiescent current for a given bandwidth and slew rate, relative to traditional amplifiers. However, significant improvements have been made to increase the efficiency of this architecture. Some op amps, such as Microchip Technology’s MCP6V03, offer a Chip Select or shutdown pin in order to minimize quiescent current when the device is not active.

Application Example: Portable Pocket Weigh Scale
The previous section identified several parameters in which the auto-zero architecture helps to increase amplifier performance. This section will explore an example application using a strain gauge, which highlights some of the advantages of an auto-zero amplifier.

Portable weigh scales are popular devices for weighing small items such as precious metals, jewelry and medications. These devices are battery powered and typically require accuracy down to a tenth of a gram, if not better. Therefore, this application requires high-precision, low-power signal conditioning for the strain gauges used to measure the weight.

A strain gauge uses electrical resistance in order to quantify the amount of strain caused by an external force. There are several different types of strain gauges, the most common of which is a metallic strain gauge. This type of strain gauge is composed of a wire or small piece of metal foil. When a force is applied, the strain on the gauge is altered (either positively or negatively), resulting in a change in the strain gauge’s electrical resistance. This change in resistance can then be measured and the magnitude of the applied force quantified. Typically, one or more strain gauges are arranged in a Wheatstone-bridge configuration, due to the excellent sensitivity that this circuit offers. The change in the resistance value is small, so the overall voltage output of such a Wheatstone-bridge circuit is small. For this example, we will assume a 10 mV full-scale output.

figure 2
Figure 2: Simplified Application Circuit Diagram

Figure 2 is a simplified circuit that will be analyzed for this application. Please note that this circuitry is not intended to be a complete representation, but is simplified to show the benefits of the auto-zero architecture. For example, the outputs of the Wheatstone-bridge circuit should be buffered to provide a high-impedance input, which is not shown in the following circuit diagram. In this circuit, the amplifier is configured for a differential gain of 500, so a full-scale output from the Wheatstone bridge will ideally produce a 5V output from the amplifier.

Due to the high amount of gain required in this application, the offset voltage of the amplifier becomes critical. Any voltage offset due to the amplifier will be multiplied by the gain. For example, the MCP606 is a CMOS op amp that implements non-volatile memory to trim the input offset voltage, resulting in a maximum offset of 250 μV (at room temperature). In this application, the maximum offset error of the MCP606 can result in 125 mV of error at the output of the amplifier, or 2.5% of the full-scale range. Let’s compare this to the MCP6V01 auto-zero amplifier, which has a maximum offset of only 2 μV (at room temperature). This offset will result in a maximum error of 1 mV at the output of the amplifier, which is only 0.02% of the full-scale output.

Another advantage of the auto-zero architecture is its low drift over time and temperature. For this example, let’s assume that the portable weigh scale is specified from 0°C to 50°C. The temperature drift of the MCP606 is specified to be 1.8 μV/°C. The error due to drift across this temperature range could be as much as 90 μV, which again would be multiplied by the gain of the circuit, causing an additional 45 mV of error at the output of the amplifier. The MCP6V01, on the other hand, specifies a maximum drift of only 50 nV/°C. Hence, the drift error for this application is only 1.25 mV at the output of the amplifier circuitry, which is over 30 times better than the performance with the MCP606 amplifier.

As stated earlier, 1/f noise can be a limiting factor for low-frequency applications, such as the weigh-scale example used here. The MCP606 op amp exhibits the typical 1/f noise spectrum, with a corner frequency around 200 Hz. The 1/f noise begins to dominate at this point, resulting in a voltage-noise density well above 200 nV/ÖHz, below 1 Hz. The MCP6V01 op amp, due to its self-correcting auto-zeroing architecture, does not exhibit this 1/f noise, which remains constant at low frequency. For weigh-scale applications, the output of the load cell is a very slow-moving signal, so 1/f noise can be a critical factor.

Conclusion
Today’s auto-zero architecture can date its roots back to the early days of chopper amplifiers, but has improved significantly since that time. The old chopper amplifiers have many shortcomings that made system-level design quite troublesome. The new auto-zero architecture is much more user friendly and provides substantially better performance. As shown in the application example, the auto-zero architecture can offer much better performance over a traditional op amp in high-precision applications.

About the Author

Kevin Tretter is Product Marketing Manager, Analog & Interface Products Division, at Microchip Technology Inc. He can be reached at kevin.tretter@microchip.com.

This article originally appeared in the September, 2008 issue of Portable Design. Reprinted with permission.

 Share
Digg Reddit Del.icio.us Stumble Upon Facebook Twitter Google BlinkList Technorati Mixx Windows Live Bookmark MySpace Yahoo Bookmarks Diigo


Insert your comment

Author Name(required):

Author Web Site:

Author email address(required):

Comment(required):

Please Introduce Secure Code:


image
image