1
Edwin Goldberg and Jules Lehmann, U. S. Patent 2,684,999: Stabilized dc amplifier.
Demystifying Auto-Zero Amplifiers—Part 1
They esntially eliminate offt, drift,and 1/f noi. How do they work?Is there a downside?
by Eric Nolan
筷子消毒器INTRODUCTION
Whenever the subject of auto-zero or chopper-stabilized amplifiers comes up, the inevitable first question is “How do they really work?”Beyond curiosity about the devices’ inner workings, the real question in most engineers’ minds is, perhaps, “The dc precision looks incredible, but what kind of weird behavior am I going to have to live with if I u one of the in my circuit; and how can I design around the problems?” Part 1 of this article will attempt to answer both questions. In Part 2, to appear in the next issue, some very popular and timely applications will be mentioned to illustrate the significant advantages, as well as some of the drawbacks, of the parts.
CHOPPER AMPLIFIERS—HOW THEY WORK
The first chopper amplifiers were invented more than 50 years ago to combat the drift of dc amplifiers by converting the dc voltage to an ac signal. Initial implementations ud switched ac coupling of the input signal and synchronous demodulation of the ac signal to re-establish the dc signal at the output. The amplifiers had limited bandwidth and required post-filtering to remove the large ripple voltages generated by the chopping action.
Chopper-stabilized amplifiers solved the bandwidth limitations by using the chopper amplifier to stabilize a conventional wide-band amplifier that remained in the signal path 1. Early chopper-stabilized designs were only capable of inverting operation, since the stabilizing amplifier’s output was connected directly to the non-inverting input of the wide-band differential amplifier. Modern IC “chopper” amplifiers actually employ an auto-zero approach using a two-or-more-stage composite amplifier structure similar to the chopper-stabilized scheme. The difference is that the stabilizing amplifier signals are connected to the wide-band or main amplifier through an additional “nulling” input terminal, rather than one of the differential inputs. Higher-frequency signals bypass the nulling stage by direct connection to the main amplifier or through the u of feed-forward techniques, maintaining a stable zero in wide-bandwidth operation.
This technique thus combines dc stability and good frequency respon with the accessibility of both
inverting and noninverting configurations. However, it may produce interfering signals consisting of high levels of digital switching “noi” that limit the ufulness of the wider available bandwidth. It also caus intermodulation distortion (IMD), which looks like aliasing between the clock signal and the input signal, producing error signals at the sum and difference frequencies. More about that later.
Auto-Zero Amplifier Principle
Auto-zero amplifiers typically operate in two phas per clock cycle,illustrated in Figures 1a and 1b. The simplified circuit shows a nulling amplifier (A A ), a main (wide-band) amplifier (A B ), storage capacitors (C M1 and C M2), and switches for the inputs and storage capacitors. The combined amplifier is shown in a typical op-amp gain configuration.
In Pha A , the auto-zero pha (Figure 1a), the input signal is applied to the main amplifier (A B ) alone; the main amplifier’s nulling input is supplied by the voltage stored on capacitor C M2;and the nulling amplifier (A A ) auto-zeros itlf, applying its nulling voltage to C M1. In Pha B , with its nulling voltage furnished by C M1, the nulling amplifier amplifies the input difference voltage applied to the main amplifier and applies the amplified voltage to the nulling input of the main amplifier and C M2.
V V a.Auto-Zero Pha A: null amplifier nulls its own offt.
V V I b.Output Pha B: null amplifier nulls the main amplifier offt.
女兵报名
Figure 1.Switch ttings in the auto-zero amplifier.
Both amplifiers u the trimmable op-amp model (Figure 2), with differential inputs and an offt-trim input.
青社V I+V I –
N
O
= A(V I+ –V I –) +BV N TRIM GAIN
Figure 2.Trimmable op amp model.
In the nulling pha (Pha A—Figure 1a), the inputs of the nulling amp are shorted together and to the inverting input terminal (common-mode input voltage). T he nulling amplifier nulls its own inherent offt voltage by feeding back to its nulling terminal whatever opposing voltage is required to make the product of that
voltage and the incremental gain of the nulling input approximately equal to A A’s input offt (V OS). The nulling voltage is also impresd on C M1. Meanwhile, the main amplifier is behaving like a normal op amp. Its nulling voltage is being furnished by the voltage stored on C M2.
During the output pha (Pha B—Figure 1b) the inputs of the nulling amplifier are connected to the input terminals of the main amplifier. C M1 is now continuing to furnish the nulling amplifier’s required offt correction voltage. The difference input signal is amplified by the nulling amplifier and is further amplified by the incremental gain of the main amplifier’s nulling input circuitry. It is also directly amplified by the gain of the main amplifier itlf (A B). The op amp feedback will cau the output voltage of the nulling amplifier to be whatever voltage is necessary at the main amplifier’s nulling input to bring the main amplifier’s input difference voltage to near-null. Amplifier A A’s output is also impresd on storage capacitor C M2, which will hold that required voltage during the next Pha A.
The total open-loop amplifier dc gain is approximately equal to the product of the nulling amplifier gain and the wide-band amplifier nulling terminal gain. The total effective offt voltage is approximately equal to the sum of the main-amplifier and nulling-amplifier offt voltages, divided by the gain at the main amplifier nulling terminal. Very high gain at this terminal results in very low effect
ive offt voltage for the whole amplifier.
As the cycle returns to the nulling pha, the stored voltage on C M2 continues to effectively correct the dc offt of the main amplifier. The cycle from nulling to output pha is repeated continuously at a rate t by the internal clock and logic circuits. (For detailed information on the auto-zero amplifier theory of operation e the data sheets for the AD8551/AD8552/AD8554 or AD857x amplifiers).
Auto-Zero Amplifier Characteristics
Now that we’ve en how the amplifier works, let’s examine its behavior in relation to that of a “normal” amplifier. First, plea note that a commonly heard myth about auto-zero amplifiers is untrue: the gain-bandwidth product of the overall amplifier is not related to the chopping clock frequency. While chopping clock frequencies are typically between a few hundred Hz and veral kHz, the gain bandwidth product and unity-gain bandwidth of many recent auto-zero amplifiers is 1 MHz–3 MHz—and can be even higher.
A number of highly desirable characteristics can be easily inferred from the operating description: dc open-loop voltage gain, the product of the gains of two amplifiers, is very large, typically more than 10 million, or 140 dB. The offt voltage is very low due to the effect of the large nulling-terminal gain
on the raw amplifier offts. Typical offt voltages for auto-zero amplifiers are in the range of one microvolt. T he low effective offt voltage also impacts parameters related to dc changes in offt voltage—dc CMR and PSR, which typically exceed 140 dB. Since the offt voltage is continuously “corrected,” the shift in offt over time is vanishingly small, only 40 nV–50 nV per month. The same is true of temperature effects. The offt temperature coefficient of a well-designed amplifier of this type is only a few nanovolts per °C!A less obvious conquence for the amplifier’s operation is the low-frequency “1/f noi” characteristic. In “normal” amplifiers, the input voltage noi spectral density increas exponentially inverly with frequency below a “corner” frequency, which may be anywhere from a few Hz to veral hundred Hz. This low-frequency noi looks like an offt error to the auto-correction circuitry of the chopper-stabilized or auto-zero amplifier. The auto-correction action becomes more efficient as the frequency approaches dc. As a result of the high-speed chopper action in an auto-zero amplifier, the low-frequency noi is relatively flat down to dc (no 1/f noi!). This lack of 1/f noi can be a big advantage in low-frequency applications where long sampling intervals are common.
中风的原因
Becau the devices have MOS inputs, bias currents, as well as current noi, are very low. However, for the same reason, wide-band voltage noi performance is usually modest. T he MOS i
nputs tend to be noisy, especially when compared to precision bipolar-procesd amplifiers, which u large input devices to improve matching and often have generous input-stage tail currents. Analog Devices AD855x amplifiers have about one-half the noi of most competitive parts. There is room for improvement, however, and veral manufacturers (including ADI) have announced plans for lower-noi auto-zero amplifiers in the future.
拂晓是什么意思Charge injection [capacitive coupling of switch-drive voltage into the capacitors] occurs as the chopping switches open and clo. This, and other switching effects, generates both voltage and current “noi” transients at the chopping clock frequency and its harmonics. The noi artifacts are large compared to the wide-band noi floor of the amplifier; they can be a significant error source if they fall within the frequency band of interest for the signal path. Even wor, this switching caus intermodulation distortion of the output signal, generating additional error signals at sum and difference frequencies. If you are familiar with sampled-data systems, this will look much like aliasing between the input signal and the clock signal with its harmonics. In reality, small differences between the gain-bandwidth of the amplifier in the nulling pha and that in the output pha cau the clod-loop gain to alternate between slightly different values at the clock frequency. T he magnitude of the IMD is dependent on the internal matching and does not relate to the magnitude of the clock “
noi.”The IMD and harmonic distortion products typically add up to about –100 dB to –130dB plus the clod-loop gain (in dB), in relation to the input signal. Y ou will e below that simple circuit techniques can limit the effects of both IMD and clock noi when they are out of band.
中间的英文Some recent auto-zero amplifier designs with novel clocking schemes, including the AD857x family from Analog Devices, have managed to tame this behavior to a large degree. The devices in this family avoid the problems caud by a single clocking frequency by employing a (patented) spread-spectrum clocking technique, resulting in esntially pudorandom chopper-related noi. Since there is no longer a peak at a single frequency in either the intrinsic switching noi or “aliad” signals, the devices can be ud at signal bandwidths beyond the nominal chopping frequency without a large error signal showing up in-band. Such amplifiers are much more uful for signal bandwidths above a few kHz.
Some recent devices have ud somewhat higher chopping frequency, which can also extend the uful bandwidth. However, this approach can degrade V OS performance and increa the input bias current (e below regarding charge injection effects); the design trade-offs must be carefully weighed. Extreme care in both design and layout can help minimize the switching transients. As mentioned above, virtually all monolithic auto-zero amplifiers have MOS input stages, tending to res
ult in quite low input bias currents. T his is a very desirable feature if large source impedances are prent. However, charge injection produces some unexpected effects on the input bias-current behavior.
舒畅近义词At low temperatures, gate leakage and input-protection-diode leakage are very low, so the dominant input bias-current source is charge injection on the input MOSFETs and switch transistors. The charge injection is in opposing directions on the inverting and noninverting inputs, so the input bias currents have opposing polarities. As a result, the input offt current is larger than the input bias current. Fortunately, the bias current due to charge injection is quite small, in the range of 10 pA–20 pA, and it is relatively innsitive to common-mode voltage.
As device temperature ris above 40°C to 50°C, the rever leakage current of the input protection diodes becomes dominant; and input bias current ris rapidly with temperature (leakage currents approximately double per 10°C increa). The leakage currents have the same polarity at each input, so at the elevated temperatures the input offt current is smaller than the input bias current. Input bias current in this temperature range is strongly dependent on input common-mode voltage, becau the rever bias voltage on the protection diodes changes with common-mode voltage. In circuits with protection diodes connected to both supply rails the bias current polarity cha
nges as the common-mode voltage swings over the supply-voltage range.
Due to the prence of storage capacitors, many auto-zero amplifiers require a long time to recover from output saturation (commonly referred to as overload recovery). This is especially true for circuits using external capacitors. Newer designs using internal capacitors recover faster, but still take milliconds to recover. The AD855x and AD857x families recover even faster—at about the same rate as “normal” amplifiers—taking less than 100 µs. This comparison also holds true for turn-on ttling time. Finally, as a conquence of the complex additional circuitry required for the auto-correction function, auto-zero amplifiers require more quiescent current for the same level of ac performance (bandwidth, slew rate, voltage noi and ttling time) than do comparable nonchopped amplifiers. Even the lowest power auto-zero amplifiers require hundreds of microamperes of quiescent current; and they have a very modest 200-kHz bandwidth with broadband noi nearly 150 nV/√Hz at 1kHz. In contrast, some standard CMOS and bipolar amplifiers offer about the same bandwidth, with lower noi, on less than 10µA of quiescent current.
APPLICATIONS
Notwithstanding all of the differences noted above, applying auto-zero amplifiers really isn’t much diff
长寿之道erent from applying any operational amplifier. In the next issue, Part 2 of this article will discuss application considerations and provide examples of applications in current shunts, pressure nsors and other strain bridges, infrared (thermopile) nsors, and precision voltage references.
