August 1, 1997

Variable-gain stage uses voltage-output DAC

Eugene Palatnik, BCI International, Waukesha, WI

A voltage-output DAC normally generates an output voltage that is a fraction of the input-reference voltage. In many designs, however, it is necessary to multiply the input voltage by a programmable gain, or, in other words, to create a variable-gain stage. You can easily implement a variable-gain stage using a low-cost voltage-output DAC (Figure 1). IC₁ includes three independent DACs, two of which have internal buffers. This design uses the third DAC (DAC C with output OUT_C) for the variable-gain stage. This no-buffer design minimizes noise and voltage offset, but you can use buffered DACs, too. Various linearity tests reveal a maximum nonlinearity error of 13.2 mV, which is less than the ±1-LSB (±19.5-mV) integral nonlinearity of the MAX512.

The output voltage of the DAC, V_DAC, is as follows, where N is the code clocked into the DAC:

IC_2A compares V_DAC with V_IN. To close the feedback loop, the output of IC_2A connects to DAC C's voltage-reference input, REF_C. In steady-state mode, V_IN=V_DAC, and, because V_REF=V_OUT,

When N equals its maximum value of 255, the gain of the circuit equals 1.004. The maximum practical gain is at N=1, which equates to a gain of 256. R₁ and C₁ reduce the loop gain as the frequency increases, which helps make the circuit stable and reduces the bandwidth and overall noise.

An 80C51 µC controls the DAC using a three-wire-plus-reset interface. The µC serially clocks the gain value in the DAC while the SELECT line is low. Pulling SELECT high sets the new gain value. Because the DAC and op amp are rail-to-rail parts, the circuit can operate in rail-to-rail mode. The DAC itself is not critical; you can use voltage-output DACs with external-reference inputs from different vendors.

(DI #2055)

| EDN Access | Feedback | Table of Contents |