Frequency compensation

In electrical engineering, frequency compensation is a design technique for amplifiers which use negative feedback or those, such as operational amplifiers, that are intended for use with negative feedback. The primary goal of frequency compensation is to avoid unintentionally creating positive feedback, although the technique enjoys secondary benefits such as providing nearly constant phase over a larger range of operating frequencies.

Motivation

Most amplifiers use negative feedback to trade gain for other desirable properties, such as improved bandwidth or noise immunity. Ideally, the phase characteristic of an amplifier's frequency response would be constant; however, device limitations make this goal physically unattainable. In general, it should be expected that even the simplest amplifiers have at least two poles. An unfortunate consequence is that at some critical frequency, the phase nears or passes through -180°. At this frequency, negative feedback becomes positive feedback; not only are the benefits of negative feedback lost, but the gain is destabilized and nonlinear effects become more pronounced. If a flat frequency response over the passband is of primary importance, then frequency compensation may not be necessary or beneficial. However, for wideband amplifiers, and in particular op-amps, which are expected to be used for a wide variety of purposes, frequency compensation is essential.

It can be shown that by using negative feedback to add a pole at low frequencies, existing high frequency poles are shifted even higher. Proper design can ensure there is only a single pole in the expected range of operating frequencies. This means that, at frequencies of interest, the phase never nears -180°.

Practice

A particularly illustrative example of frequency compensation in practice is the op-amp. Because negative feedback is so important in many op-amp circuits, the need for frequency compensation outweighs its drawbacks. In fact, most op-amps have so much frequency compensation that they might be said to be overcompensated for certain applications.

An op-amp will typically have a pole added at a very low frequency, on the order of 4 Hz. The result is a nearly flat phase response, at -90°, from about 40 Hz to very high frequencies, which is a secondary benefit. More important, however, is the fact that over most practical frequencies, negative feedback is preserved. Frequency compensation might be accomplished by adding a capacitor in a feedback loop in the second stage of the amplifier.

Drawbacks

As noted above, op-amps are overcompensated for certain applications. This is not the case for all amplifiers that use frequency compensation. However, an unavoidable consequence of this technique is a decreased bandwidth. This tradeoff is usually considered acceptable, as the bandwidth can be restored by the more traditional use of negative feedback.

Another, potentially disastrous, drawback of frequency compensation arises when active components are used in the feedback loop. For example, a hypothetical amplifier configuration which uses an op-amp for both the forward amplifier and the feedback loop would have -180° phase at even low frequencies. This type of configuration should be avoided.

In op-amps, slewing, a nonlinear phenomenon that limits performance, is partially due to frequency compensation. Although slewing would occur even without frequency compensation, its effect is more pronounced because of it.