7 Test setup and comparison

We used CMRR and offset voltage errors as a measure of the relative performance of each circuit across temperature. For each device, we connected a precision source measure unit to both input pins of the DA and used a calibrated 8.5-digit multimeter to measure the change in offset voltage. We repeated all testing over five averaged sweeps to obtain an accurate representation of device performance, configuring the devices to sweep a common-mode voltage of -35V to 35V, with a split supply configuration of ±18V. We performed overtemperature testing with an oven, along with sufficient soak time to ensure uniform temperature across the test board.

Forcing a differential voltage across the device inputs while maintaining a common-mode voltage at mid-supply tests both the gain error and gain error drift. Sweeping each device with a corresponding input range forces the output to range from –10V to +10V, enabling you to compare the corresponding slope to the ideal slope and thus evaluate the percentage of gain error.

Table 1 compares the performance of CMRR and the offset between a discrete DA and two TI integrated DAs across different operating temperatures.

Table 2 compares the gain error and drift performance between the same discrete DA and integrated DAs across different operating temperatures.

As expected, the integrated DAs performed exceptionally well in achieving high CMRR, low gain error and low gain error drift compared to the discrete DA. While the offset voltage of the discrete DA outperformed one of the integrated DAs, it is possible to compensate for this through software calibration.

Figure 5 shows a simplified layout of each of the three DA variants and compares the sizes of each solution. For comparison purposes, we used the smallest device packages, along with resistors and capacitors in the 0402 package.

Figure 5 Size comparison