Active-feedback IC serves as current-sensing instrumentation amplifier
High-speed current sensing presents a designer with some significant challenges. Most techniques for sensing current involve measuring the differential voltage the current produces as it flows through a sense element, such as a resistor or a Hall-effect device. The differential voltage across the sense element is generally small and is often riding on a common-mode voltage that is considerably larger than the differential voltage itself. Accurate amplification of the differential voltage requires a differential amplifier with high input impedance, high CMR (common-mode rejection); wide input-common-mode voltage range; and high, well-defined gain. Traditional instrumentation amplifiers have these features and often serve for low-frequency current sensing, but they perform poorly at high speeds. High-speed current sensing requires the kind of performance that instrumentation amps provide, but their abilities must extend to high frequencies. Figure 1 shows how high-speed active feedback amplifiers, such as the AD8129 and AD8130 differential receivers, are ideal for these high-speed instrumentation-amp applications. The AD8129 requires a minimum closed-loop voltage gain of 10 for stability, whereas the AD8130 is unity-gain-stable.
Active-feedback amplifier operation differs from that of traditional op amps; it provides a beneficial separation between the signal input and the feedback network. Figure 1 shows a high-level block diagram of an active-feedback amplifier in a typical closed-loop configuration. High-speed current sensing uses a resistor as the sense element. The input stages are high-impedance, high-CMR, wideband, high-gain transconductance amplifiers with closely matched transconductance parameters. The output currents of the transconductance amplifiers undergo summing, and the voltage at the summing node is buffered to provide a low-impedance output. Applying negative feedback around amplifier B drives VOUT to a level that forces the input voltage of amplifier B to equal the negative value of the input voltage at amplifier A, because the current from amplifier A equals the negative value of the current from amplifier B, and the gm values are closely matched. From the foregoing discussion, you can express the closed-loop voltage gain for the ideal case as: VOUT/VIN=1+RF/RG≡AV.
Measurement sensitivity in volts per amp is expressed as: VOUT/ISENSE=AVRSENSE. Minimizing the values of RF and RG also minimizes resistor and output-voltage noise arising from input-referred current noise. Because of the small sense resistance and high measurement frequencies, you must minimize parasitic effects in the input circuitry to avoid measurement errors. Parasitic trace inductance in series with the sense element is of particular concern, because it causes the impedance across the amplifier's input to increase with increasing frequency, producing a spurious increase in output voltage at high frequencies. Figure 2 illustrates a test circuit with RSENSE=1Ω and AV=20, which equates to a measurement sensitivity of 20V/A. The three-pole lowpass filter produces a defined bandwidth and attenuates spurious responses at the amplifier's output arising from input signals at frequencies outside the desired measurement bandwidth. The test circuit's frequency response in Figure 3 shows that the expected differential-to-single-ended gain of 20/101, or –14 dB, is flat to approximately 10 MHz and is down by 3 dB at 62 MHz. Figure 3 demonstrates the effectiveness of the high CMR of active-feedback amplifiers. The common-mode signal at the amplifier's input is approximately 50 times greater than the differential signal across the sense resistor.