Buffers (electronics)

Electronic circuits whose primary function is to connect a high-impedance source to a low-impedance load without significant attenuation or distortion of the signal. Thus, the output voltage of a buffer replicates the input voltage without loading the source. An ideal voltage buffer is an amplifier with the following properties: unity gain, A_B = 1; zero output impedance, Z_out = 0; and infinite input impedance, Z_in = ∞. For example, if the voltage from a high-impedance source, say a strain-gage sensor with 100 k&OHgr; output resistance, must be processed by further circuitry with an input impedance of, say, 500 &OHgr;, the signal will be attenuated to only 500/100,500 ≈ 0.5% of the sensor voltage if the two circuits are directly connected, whereas the full strain-gage voltage will be available if a buffer is used.

Buffers are generally applied in analog systems to minimize loss of signal strength due to excessive loading of output nodes (illus. a). Two kinds of circuits are frequently used: the operational-amplifier-based buffer and the transistor follower.

Buffer circuit

The operational-amplifier-based buffer circuit (illus. b) is based on an operational amplifier (op amp) with unity-gain feedback. The open-loop gain, A(s), of the operational amplifier should be very high. To form the buffer, the amplifier is placed in a feedback loop. The buffer gain, A_B(s), is then given by Eq. (1). (1) Here, s = j&ohgr; is the Laplace transform variable, j = -1; &ohgr; = 2&pgr;f is the radian frequency in radians per second (rad/s); and f is the frequency in hertz (Hz). The magnitude of A_B is approximately equal to unity, that is, |A_B| approaches 1, if |A| becomes very large. A common representation of the frequency dependence of the operational-amplifier gain is given by Eq. (2), (2)  where &ohgr;_t is the operational amplifier's unity-gain frequency. By using this notation, Eq. (1) becomes Eq. (3), (3)  which shows that the buffer's bandwidth is approximately equal to the unity-gain frequency of the operational amplifier, typically 1 MHz or higher.

Under the assumption that the frequency of interest is much less than &ohgr;_t, it follows from Eq. (2) that the magnitude of the operational-amplifier gain, A(s), is much greater than 1. In that case, it can be shown that, because of the feedback action, the buffer's input impedance is much larger than that of the operational amplifier itself [by a factor of A(s)]. Similarly, the buffer's output impedance is much smaller than that of the operational amplifier [again, by a factor of A(s)].

The very low output impedance of operational-amplifier-based buffers assures that a load impedance, Z_L(s), does not affect the buffer's gain, A_B. Also, operational-amplifier-based buffers have no systematic offset. The high-impedance input node of a buffer may in practice have to be shielded to prevent random noise from coupling into the circuit. This shielding can be accomplished with a coaxial cable. To eliminate the capacitive loading of the source by the effective input capacitance of the cable, the shield can be driven with the output voltage of the buffer so that no voltage difference exists between the signal line and the shield. The driven shield is referred to as the guard. See Amplifier, Operational amplifier

The bipolar junction transistor (BJT) emitter follower and the field-effect transistor (FET) source follower are very simple but effective buffer circuits. Both consist of a single transistor and a bias-current source; they are used in applications where power consumption and circuit area must be reduced to a minimum or where specifications are not too demanding.

The performance of a transistor follower circuit depends strongly on the source and load impedances, that is, on the surrounding circuitry. In fact, the transistors are so fast that the frequency response is usually determined by loading. In general, follower circuits exhibit a systematic direct-current (dc) offset equal to the base-to-emitter voltage, V_BE, in BJTs and equal to the gate-to-source voltage, V_GS, for FET. Only followers made with depletion-mode field-effect transistors can be biased with zero V_GS to avoid this offset. See Emitter follower, Transistor

Buffer circuits should have small dc offset voltages (dc outputs when no input is applied), small bias currents (to minimize the effect of high-impedance sources), large linear signal swing (to minimize distortion), and high slew rate (to handle fast transitions of the applied signals).

Buffers should have a low-frequency gain of unity and wide bandwidth (to reproduce the applied signals faithfully), low phase margins (to prevent peaking and overshoots), and low equivalent input-referred noise (to have wide dynamic range). Field-effect-transistor input buffers exhibit the lowest noise for high-impedance signal sources. See Gain