previous

**VFB Op Amps Designed on Complementary Bipolar Processes**
We will now examine in more detail the current feedback (CFB) op amp topology which is very popular in high speed op amps. As mentioned previously, the circuit concepts were introduced decades ago, however modern high speed complementary bipolar processes are required to take full advantage of the architecture.

It has long been known that in bipolar transistor circuits, currents can be switched faster than voltages, other things being equal. This forms the basis of non-saturating emitter-coupled logic (ECL) and devices such as current-output DACs. Maintaining low impedances at the current switching nodes helps to minimize the effects of stray capacitance, one of the largest detriments to high speed operation. The current mirror is a good example of how currents can be switched with a minimum amount of delay.

Figure 1-106: Simplified current feedback (CFB) op amp

The current feedback op amp topology is simply an application of these fundamental principles of current steering. A simplified CFB op amp is shown in Figure 1-106 above. The non-inverting input is high impedance and is buffered directly to the inverting input through the complementary emitter follower buffers Q1 and Q2. Note that the inverting input impedance is very low (typically 10 to 100Ω), because of the low emitter resistance (ideally, would be zero). This is a fundamental difference between a CFB and a VFB op amp, and also a feature that gives the CFB op amp some unique advantages.

The collector outputs of Q1 and Q2 drive current mirrors, which mirror the inverting input current to the high impedance node, modeled by R

Now, consider a positive step voltage applied to the non-inverting input of the CFB op amp. Q1 immediately sources a proportional current into the external feedback resistors creating an error current, which is mirrored to the high impedance node by Q3. The voltage developed at the high impedance node is equal to this current multiplied by the equivalent impedance. This is where the term transimpedance op amp originated, since the transfer function is an impedance, rather than a unitless voltage ratio as in a traditional VFB op amp._{T}and C_{P}. The high impedance node is buffered by a complementary unity gain emitter follower. Feedback from the output to the inverting input acts to force the inverting input current to zero, hence the term Current Feedback. Note that in a ideal case, for zero inverting input impedance, no small signal voltage can exist at this node, only small-signal current.
Note also that the error current delivered to the high impedance node is not limited by the input stage tail current. In other words, unlike a conventional VFB op amp, there is no slew-rate limitation in an ideal CFB op amp. The current mirrors supply current-on-demand from the power supplies. The negative feedback loop then forces the output voltage to a value that reduces the inverting input error current to zero.

Figure 1-107: CFB op amp model and Bode plot

The model for a CFB op amp is shown in Figure 1-107 above, along with the corresponding Bode plot. The Bode plot is plotted on a log-log scale, and the open-loop gain is expressed as a transimpedance, T(s), with units of ohms.

The finite output impedance of the input buffer is modeled by R

_{O}. The input error current is i. By applying the principles of negative feedback, we can derive the expression for the op amp transfer function:
At the op amp 3dB closed-loop bandwidth frequency, f

_{cl}, the following is true:
Solving for f

_{cl}:
For the condition R

_{O}<< R2 and R1, the equation simply reduces to:
Examination of this equation quickly reveals that the closed-loop bandwidth of a CFB op amp is determined by the internal dominant pole capacitor, C

_{P}, and the external feedback resistor R2, and is independent of the gain-setting resistor, R1. This ability to maintain constant bandwidth independent of gain makes CFB op amps ideally suited for wideband programmable gain amplifiers.
Figure 1-108: AD8011 frequency response, G = +1, +2, +10

Because the closed-loop bandwidth is inversely proportional to the external feedback resistor, R2, a CFB op amp is usually optimized for a specific R2. Increasing R2 from its optimum value lowers the bandwidth, and decreasing it may lead to oscillation and instability because of high frequency parasitic poles.

The frequency response of the AD8011 CFB op amp is shown in Figure 1-108 above for various closed-loop values of gain (+1, +2, and +10). Note that even at a gain of +10, the closed-loop bandwidth is still greater than 100MHz. The peaking which occurs at a gain of +1 is typical of wideband CFB op amps used in the non-inverting mode, and is due primarily to stray capacitance at the inverting input. This peaking can be reduced by sacrificing bandwidth, by using a slightly larger feedback resistor. The AD8011 CFB op amp (introduced in 1995) still represents state-of-the-art performance, and key specifications are shown in Figure 1-109 below.

Figure 1-109: AD8011 key specifications

Traditional current feedback op amps have been limited to a single gain stage, using current-mirrors. The AD8011 (and also others in this family) unlike traditional CFB op amps, use a two-stage gain configuration, as shown in Figure 1-110 below.

Figure 1-110: Simplified two-stage current feedback op amp

Until the advent of the AD8011, fully complementary two-gain stage CFB op amps had been impractical because of their high power dissipation. The AD8011 employs a patented (Reference 2: Roy Gosser, DC-Coupled Transimpedance Amplifier, US Patent 4,970,470, filed October 10, 1989, issued November 13, 1990) second gain stage consisting of a pair of complementary amplifiers (Q3 and Q4). Note that they are not connected as current mirrors but as grounded-emitter gain stages. The detailed design of current sources (I1 and I2), and their respective bias circuits are the key to the success of the two-stage CFB circuit; they keep the amplifier's quiescent power low, yet are capable of supplying current-on-demand for wide current excursions required during fast slewing.

A further advantage of the two-stage amplifier is the higher overall bandwidth (for the same power), which means lower signal distortion and the ability to drive heavier external loads.

Thus far, we have learned several key features of CFB op amps. The most important is that for a given complementary bipolar IC process, CFB generally yields higher FPBW (hence lower distortion) than VFB for the same amount of quiescent supply current. This is because there is practically no slew-rate limiting in CFB. Because of this, the full power bandwidth and the small signal bandwidth are approximately the same.

Figure 1-111: Performance of selected CFB op amps

The second important feature is that the inverting input impedance of a CFB op amp is very low. This is advantageous when using the op amp in the inverting mode as an I/V converter, because there is less sensitivity to inverting input capacitance than with VFB.

The third feature is that the closed-loop bandwidth of a CFB op amp is determined by the value of the internal Cp capacitor and the external feedback resistor R2 and is relatively independent of the gain-setting resistor R1.
The performance for a selected group of current feedback op amps is shown in Figure 1-111 above. Note that the op amps are listed in order of decreasing power supply current.

Figure 1-112: Summary: CFB op amps

Figure 1-112 above summarizes the general characteristics of CFB op amps.

next **Effects of Feedback Capacitance in Op Amps**
## 0 comments:

## Post a Comment

Please wait for approval of your comment .......