18-Bit, Linear, Low Noise, Precision Bipolar ±10 V DC Voltage Source

Circuit Functionality and Benefits

The circuit shown in Figure 1 provides an 18-bit programmable voltage with an output range of −10 V to +10 V, while having an integral nonlinearity of ±0.5 LSB, a differential nonlinearity of ±0.5 LSB, and low noise.

The digital input of the circuit is serial and compatible with standard SPI, QSPI, MICROWIRE®, and DSP interface standards. For high precision applications, the circuit provides high precision and low noise performance by combining precision components such as the AD5781, ADR445, and AD8676.

The reference buffer is critical to the design because the input impedance of the DAC reference input is highly code dependent, which will cause linearity errors if the DAC reference source is not adequately buffered. The AD8676 has a high open-loop gain of 120 dB and has been proven and tested to meet the requirements of this circuit application for settling time, offset voltage, and low impedance drive capability. The AD5781 is characterized and factory calibrated to buffer its voltage reference input using the dual-channel AD8676 op amp, further enhancing the reliability of the companion components.

This combination of devices offers industry-leading 18-bit resolution, ±0.5 LSB integral nonlinearity (INL) and ±0.5 LSB differential nonlinearity (DNL), guaranteed monotonicity, low power consumption, small PCB size, and cost-effectiveness.

Figure 1. 18-Bit Precision, ±10 V Voltage Source (Simplified Schematic: Decoupling and All Connections Not Shown)

Circuit Description
The digital-to-analog converter (DAC) shown in Figure 1 is the AD5781, an SPI-interfaced, 18-bit, high-voltage converter that provides ±0.5 LSB INL, ±0.5 LSB DNL, ​​and 7.5 nV/√Hz noise spectral density. The AD5781 also features very low drift over temperature (0.05 ppm/°C). The precision architecture used by the AD5781 requires forced sensing to buffer its voltage reference inputs to ensure that the specified linearity is achieved. The amplifiers (B1 and B2) selected to buffer the reference inputs should have low noise, low temperature drift, and low input bias current. The AD8676 amplifier is recommended for this function, which is an ultra-precision, 36 V, 2.8 nV/√Hz dual op amp with a low offset drift of 0.6 μV/°C and 2 nA input bias current. In addition, the AD5781 is characterized and factory calibrated to use this dual op amp to buffer its voltage reference inputs, further enhancing the reliability of the companion device.

In Figure 1, the AD5781 is configured in a gain-of-2 mode, which allows a single reference to be used to generate a symmetrical bipolar output voltage range. This mode of operation uses an external op amp (A2) and on-chip resistors (see the AD5781 data sheet) to provide a gain of 2. These internal resistors are thermally matched to each other and to the DAC ladder resistors, allowing ratiometric thermal tracking. The output buffer is also an AD8676, which has low noise and low drift. This amplifier (A1) is also used to amplify the +5 V reference voltage from the low noise ADR445 to +10 V. R2 and R3 in this gain circuit are precision metal foil resistors with a tolerance and temperature coefficient resistance of 0.01% and 0.6 ppm/°C, respectively. For best performance over the entire temperature range, R1 and R2 should be in a single package, such as the Vishay 300144 or VSR144 series. R2 and R3 are both selected to be 1 kΩ to keep system noise low. R1 and C1 form a low-pass filter with a cutoff frequency of approximately 10 Hz. This filter is used to attenuate the reference noise.

Figure 2. Integral nonlinearity vs. DAC code.

Linearity Measurements
The precision performance of the circuit shown in Figure 1 is shown in the data in Figures 2 and 3, which show the integral nonlinearity and differential nonlinearity as a function of DAC code. It is clear from the figures that both characteristics are within the specifications of ±0.5 LSB and ±0.5 LSB, respectively.

The total unadjusted error of the circuit is a combination of dc errors, namely INL error, offset error, and gain error. Figure 4 shows a plot of the total unadjusted error versus DAC code. The error is largest at DAC codes 0 and 262,143. This is expected and is caused by the absolute error in the reference output, the mismatch of the external resistors R2 and R3 (see Figure 1), and the mismatch of the AD5781 internal resistors RFB and R1 (see Figure 5).

Figure 3. Differential nonlinearity vs. DAC code.

Figure 4. Total unadjusted error vs. DAC code

Figure 5. Circuit with internal gain of 2 (schematic diagram)

The absolute error of the reference voltage is specified at 0.04%; the mismatch between resistors R2 and R3 in this example is specified at 0.02%; and the mismatch between internal resistors R1 and RFB is specified at 0.01%. Therefore, the total gain error is 0.07% of the full-scale range, or 184 LSBs. Figure 4 shows the measured value of 20 LSBs (or 0.007% of the full-scale range), indicating that all components perform significantly better than their specified tolerances. Noise

Measurement
To achieve high accuracy, the peak-to-peak noise at the output of the circuit must be maintained below 1 LSB, which is 76.29 μV for 18-bit resolution and a 20 V peak-to-peak voltage range. Figure 6 shows the peak-to-peak noise measured over a 0.1 Hz to 10 Hz bandwidth for 10 seconds. The peak-to-peak values ​​for the three conditions are 1.34 μV (midscale output), 12.92 μV (full-scale output), and 15.02 μV (zero-scale output). The noise is lowest at the midscale output, where the noise comes only from the DAC core. When midscale code is selected, the DAC attenuates the noise contribution of each reference voltage path.

Figure 6. Voltage noise (0.1 Hz to 10 Hz bandwidth)

However, in real applications, there will not be a high-pass cutoff frequency at 0.1 Hz to attenuate 1/f noise, but it will include frequencies down to DC in its passband; therefore, the measured peak-to-peak noise is more realistic, as shown in Figure 7. In this case, the noise at the output of the circuit was measured over 100 seconds, which fully covers frequencies down to 0.01 Hz. The upper cutoff frequency is approximately 14 Hz and is limited by the measurement setup. For the three conditions shown in Figure 7, the corresponding peak-to-peak values ​​are 1.61 μV (midscale output), 43.33 μV (full-scale output), and 36.89 μV (zero-scale output). The worst-case peak-to-peak value (43.33 μV) is roughly equivalent to ½ LSB.

Figure 7. Voltage noise measured over 100 seconds.

As the measurement time gets longer, lower frequencies will be included and the peak-to-peak value will get larger. At lower frequencies, temperature drift and thermocouple effects become sources of error. These effects can be minimized by selecting components with low thermal coefficients. In this circuit, the dominant source of low frequency 1/f noise is the voltage reference. The voltage reference also has the largest temperature coefficient value in the circuit at 3 ppm/°C.

COMMON VARIATIONS
The AD5781 supports a variety of output ranges from 0 V to +5 V, up to ±10 V, and anywhere in between. If a symmetrical output range is desired, a gain of 2 configuration can be used, as shown in Figure 1. This mode is selected by setting the RBUF bit in the AD5781 internal control register to Logic 0. If an asymmetrical range is desired, separate references can be applied to VREFP and VREFN; the output buffer should be configured to provide unity gain as described in the AD5781 data sheet. This can be accomplished by setting the RBUF bit in the AD5781 internal control register to Logic 1.

Circuit Evaluation and Test
The circuit shown in Figure 1 was constructed on a modified AD5781 evaluation board. Details on the AD5781 evaluation board and test methods are provided in the Evaluation Board User Guide UG-184.