Multiplying Digital-to-Analog Converter Applications

While conventional Digital-to-Analog Converters (DACs) take a digital code and produce a fixed analog voltage or current, multiplying DACs are the chameleons of the signal chain. They accept both a digital code and an external analog reference input, producing an output that is the product of these two variables. This unique capability allows them to function as digitally programmable resistors, gain stages, and attenuators, making them indispensable for creating flexible, software-controlled analog systems in applications ranging from test equipment to audio processing.

The Core Principle: From Static Output to Variable Gain

The fundamental difference lies in the architecture. A standard voltage-output DAC has an internal, fixed reference voltage. Its output is simply the digital code scaled against this fixed reference: $V_{out}=Code\times V_{re{f}_{fixed}}$.

A Multiplying DAC (MDAC), however, is designed to use an external reference voltage ($V_{REF}$) that can vary over a wide range, often including bipolar (positive and negative) signals. Its output is given by: $$V_{out}=(DigitalCode)\times V_{REF}$$

For a binary-weighted DAC with N bits, the code is a fraction. If $D$ is the decimal equivalent of the digital input (ranging from 0 to $2^N-1$), the output becomes: $$V_{out}=\left(\frac{D}{2^N}\right)\times V_{REF}$$

This equation reveals the core behavior: the digital code acts as a scaling factor, or multiplier, applied to the incoming analog reference signal. If $V_{REF}$ is a DC voltage, the DAC sets a precise output level. If $V_{REF}$ is an AC signal like a sine wave, the DAC directly scales its amplitude—it becomes a digitally controlled attenuator or amplifier.

Application 1: Digitally Programmable Gain and Attenuation

This is the most direct application. By using the signal you wish to scale as the $V_{REF}$ input to the MDAC, the digital code sets the gain. Consider an audio mixing console where you need to digitally control the volume of a channel. The audio signal itself is fed into the MDAC's $V_{REF}$ pin. The microcontroller, based on the user's fader position, sends a corresponding digital code to the MDAC. The output is the original audio signal, multiplied by the code's fractional value.

• Attenuation: When the code is less than full scale (e.g., D/1024 for a 10-bit DAC), the output is an attenuated version of $V_{REF}$. A code of 512/1024 applies -6 dB of attenuation.
• Amplification: Many MDACs can be configured in an op-amp circuit to provide actual gain. If the reference input is connected to the output of a fixed-gain stage, the MDAC in the feedback loop can digitally program the overall circuit gain to values greater than 1. This is crucial in automatic test equipment (ATE), where a single instrument must measure signals from microvolts to tens of volts by dynamically adjusting its input amplifier's gain under software control.

Application 2: Precision Offset and Bipolar Output Adjustment

MDACs excel in calibration circuits. By using a stable DC voltage as $V_{REF}$, the MDAC generates a precise, digitally adjustable offset voltage. This offset can be summed with another signal to null out system errors or to establish precise bipolar output ranges.

A common configuration is the 4-quadrant multiplying DAC. Here, both the reference and the output can be positive or negative. The digital code is often expressed in two's complement format. A code of mid-scale (e.g., 0x800 for a 12-bit DAC) produces a 0V output regardless of $V_{REF}$. Codes above mid-scale yield an output with the same polarity as $V_{REF}$, while codes below mid-scale invert the polarity. This is essential in programmable power supplies and motor control, where you need to digitally set both positive and negative output voltages or control the direction of a current.

Application 3: Advanced Signal Processing and Filtering

The ability to multiply a digital coefficient by an analog signal enables sophisticated analog computing. One advanced application is in digitally controlled analog filters.

For instance, in a state-variable or programmable filter design, the filter's critical parameters—like cutoff frequency ($f_c$) and Q (quality factor)—depend on the ratio of resistor values. By replacing these fixed resistors with MDACs, the filter's characteristics become software-programmable. The digital code sets the MDAC's effective "resistance," thereby tuning $f_c$ and Q on the fly. This creates a flexible filter block for communications systems or signal analyzers that must adapt to different bandwidths.

Furthermore, in complex modulation schemes or waveform generation, an MDAC can be used to apply a digital envelope or weighting function to an analog carrier wave, performing a fundamental mixing operation directly in the digital domain's control of an analog process.

Common Pitfalls

1. Ignoring Reference Input Impedance and Bandwidth: The $V_{REF}$ pin is not a perfect input. It has a dynamic impedance that can vary with the digital code, potentially distorting the source signal if not driven properly. Additionally, the MDAC has a specified multiplying bandwidth—the maximum frequency of $V_{REF}$ for which the DAC can maintain accurate multiplication. Using an AC reference signal beyond this bandwidth results in increased distortion and gain error. Always buffer the reference source and check the bandwidth specification for AC applications.

1. Code-Dependent Glitches During Updates: When the digital input code changes, the internal switches of the DAC transition, causing brief, high-energy glitches at the output. In a static-output DAC, these are troublesome. In an MDAC with a live AC signal on $V_{REF}$, these glitches are modulated by the reference signal, creating unwanted spectral noise. To mitigate this, use a deglitcher circuit (a sample-and-hold) at the output, or synchronize code updates during the zero-crossings of the reference signal when possible.

1. Assuming Perfect Linearity for Attenuation: While an MDAC is a precise digital attenuator, its performance is limited by its analog linearity specifications—namely Differential Non-Linearity (DNL) and Integral Non-Linearity (INL). At very low digital codes (deep attenuation), these non-linearities become a larger percentage of the small output signal, potentially distorting it. For high-fidelity attenuation applications, select MDACs with excellent low-code linearity.

1. Neglecting Temperature Coefficients: The gain accuracy of the multiplication depends on the stability of the internal resistor ladder network. The gain temperature coefficient specifies how much the scale factor drifts with temperature. If your system operates over a wide temperature range and requires stable attenuation, this coefficient can be a significant source of error, sometimes more critical than the initial gain error.

Summary

• A Multiplying DAC (MDAC) outputs the product of its digital input code and an external analog reference voltage ($V_{OUT}=Code\times V_{REF}$), enabling it to act as a digitally controlled variable resistor or gain element.
• Its primary applications include digitally programmable gain/attenuation for signal conditioning (as in automatic test equipment), generating precision offset voltages for system calibration, and creating bipolar output ranges for programmable power supplies.
• Advanced uses leverage the multiplication for real-time analog signal processing, such as building digitally tuned filters or applying modulation envelopes.
• Successful implementation requires careful attention to the reference signal integrity, management of switching glitches, understanding of low-code linearity limits, and consideration of temperature drift effects on multiplication accuracy.