Error processing SSI file
Analog Dialogue Current Issue

Two-Stage Current-Feedback Amplifier

by Will Drachler

The AD8011's excellent performance (table) and power economy are due to a unique two-gain-stage current-feedback architecture (Figure A1) and fabrication on Analog Devices's bonded-wafer extra-fast complementary bipolar (XFCB) process.

PAGE4FIG3.gif

AD8011 PERFORMANCE

Supply voltage(s) +5 V 5 V
Supply current, max 1 mA 1.2 mA
Quiescent power 5 mW 12 mW
Bandwidth, -3-dB, small-signal, typical
G = +1 325 MHz 400 MHz
G = +2 180 MHz 210 MHz
Bandwidth, -3 dB, 2.5 V p-p out, typical
G = +10 57 MHz 57 MHz
0.1-dB gain flatness 20 MHz 25 MHz
Slewing rate, typical, G = +2 2000 V/µs 3500 V/µs
Settling to 0.1%, 2-V step, G = +2 29 ns 25 ns
Distortion: 2nd, 3rd, R(L)=1 kOhms
5 MHz -84, -76 dB -75, -70 dB
20 MHz -59, -53 dB -62, -63 dB
Change in Gain error: R(L) = 1 kOhms, 150 Ohms 0.02%, 0.6% 0.02%, 0.02%
Change in Phase error: R(L) = 1 kOhms, 150 Ohms 0.06°, 0.8° 0.06°, 0.3°
Voltage noise @ 10 kHz, typical 2 nV/Sqrt(Hz) 2 nV/Sqrt(Hz)
Offset voltage, max 5 mV 5 mV
Output current available, typical 30 mA 30 mA
Input range 1.2 to 3.8 V 4.1 V
Output swing, R(L) = 1 kOhms 0.9 to 4.1 V 4.1 V

What is a current-feedback amplifier? A current feedback, or transimpedance, op amp (CFA) is used in circuits that look much like the circuits that employ voltage feedback op amps (VFAs). What is the difference? First, the negative input of a CFA responds to current; the output voltage is proportional to that current, hence transimpedance (V(out) = Z(t) I(in)). Instead of keeping the negative input current small by maintaining high input impedance, and using feed-back and voltage gain to keep the input voltage difference small, the CFA keeps the voltage difference small by virtue of its low input imped-ance (like looking back into a low-offset emitter follower); and it keeps its net input current small dynamically by feedback from the output.

When an ideal CFA is driven at the high-impedance positive input, the negative input, with its low impedance, follows closely in voltage; and the high gain for error current and the negative feedback through Rf require that the currents through Rf and Rin be equal; hence V(out) = V(in)[R(f)/R(in) + 1], just like for voltage-feedback amplifiers. A major difference is that the slew rate can be quite high, because large transient currents can flow in the input stage to handle rapid changes in voltage across the compensating capacitor(s). Also, the low impedance at the negative input means that stray input capacitance will not substantially affect the amplifier's bandwidth.

How does the AD8011 differ from conventional current amplifiers? Conventional IC current-feedback op amps, though built on complementary-bipolar processes, have been limited to a single gain stage, using current-mirror circuits for biasing and level-shifting (Figure A2a). Until now, fully complementary, two-gain-stage, current-feedback IC op amps have been impractical because of their high power consumption.

PAGE5FIG2.gif

The AD8011 employs a second gain stage consisting of a pair of complementary amplifiers, A2 and A2, exemplified by Q4 and Q5. The detailed design of current sources I(1) and I(2), and their bias circuit, are the (patent-applied-for) key to the success of the two-stage circuit; they keep the amplifier's quiescent power low, yet are capable of wide current excursions during slewing. Voltage developed at the output of the transconductance stage of a current-feedback amplifier (the high-impedance "C(p)" node) is unloaded by the unity-gain buffer to provide output drive.

A further advantage of the two-stage amplifier is the higher overall gain-bandwidth (for the same power), which means lower signal distortion and the ability to drive heavier external loads. Conversely, the second gain stage divides down the drive required by the buffer, as well as its nonlinearities, resulting in lower distortion and higher open-loop gain. The AD8011 retains the advantage of current-feedback amplifiers that higher closed-loop gains do not involve a proportional reduction in bandwidth. Finally, the amplifier's wide common-mode and output ranges permit it to operate on a single 5-volt supply, with half the power and little degradation of performance over 5-V operation.

Summing up, the novel bias circuit keeps the complementary currents equal and low--regardless of beta differences--albeit at the expense of slightly greater circuit complexity. However, the XFCB process permits fabrication of very small, fast transistors, so the die remains small (and the cost low). The AD8011's use of the second gain stage and the bias circuit provides it with all of the advantages of class-B operation, from input stage to output stage. It makes possible low distortion, high-speed, and high output current drive while running on low quiescent current.