Although current-feedback operational amplifiers (op amps) have been on the market since the 1990s, the current-feedback fully differential amplifier was not introduced until 2007.
There are several features of current-feedback fully differential amplifiers that set them apart from voltage feedback fully differential amplifiers. This article will detail several of these differences and how they impact circuit design.
Current feedback is the property where the feedback loop of an amplifier creates an output voltage proportional to the current fed back to the input. The transfer function is voltage out with respect to current, and the circuit is a transimpedance block.
When choosing between voltage- and current-feedback fully differential amplifiers there are several key differences. With a voltage-feedback amplifier, the input pins are high impedance; the feedback creates a condition where the difference in voltage between the two input pins is ideally zero. With a current-feedback amplifier, it is the reverse. The current-feedback fully differential amplifier’s input pins are low impedance; the feedback circuit drives the current flowing into the input pins to ideally 0A. This difference is largely transparent, since either way there is virtually no voltage between the input pins. However, it does introduce performance differences that are highlighted below.
With voltage-feedback amplifiers, you have significant flexibility on your feedback-resistor values because the voltage presented to the inverting input is the important parameter. In current-feedback amplifiers this is not the case; the current through the feedback resistor is the key parameter.
Figures 1 and 2 illustrate this behavior using the Texas Instruments TINA-TI software model for the LMH6554, a current-feedback, fully differential amplifier. The figures show that changing the feedback-resistor value will have a large impact on circuit performance. Because circuit performance will deteriorate when the feedback-resistor value is not close to the design value, the signal-gain range of the current-feedback amplifier is limited. This gain range may be specified in the data sheet, but if it is not, you can estimate the effective-gain range based on the resistor value.
Feedback-resistor impact on bandwidth
Feedback-resistor impact on bandwidth
For example, if the recommended feedback-resistor value is 250 Ω, you can see that a gain of 5 will result in an input impedance of 100Ω (differential). For many applications, an amplifier input impedance below 100Ω is not desirable, so the upper-gain range of this amplifier is limited to approximately 5V/V. The reason this limit is approximate is because there is no hard limit on how large you can set the feedback resistor. Increasing the feedback resistor results in an overcompensated amplifier, with the main drawback being reduced bandwidth. Since current-feedback amplifiers often have very high bandwidth, the bandwidth reduction may not be critical.
Changing the amplifierís compensation with the feedback resistor
While high gain can be problematic for a current-feedback amplifier, low gain is not. As gain decreases, you can fine-tune the amplifier compensation and therefore the stability with the feedback resistor. Figure 3 shows this fine-tuning, again using the LMH6554 model.
Constant bandwidth with higher gain
If a low-gain circuit exhibits overshoot and peaking in the frequency response, you can increase the feedback-resistor value to eliminate the instability. For this reason, current-feedback fully differential amplifiers are often used as buffers to analog-to-digital converters (ADCs). An ADC input buffer requires high speed, high linearity, excellent stability and a low settling time. Current-feedback fully differential amplifiers can meet these requirements.