As an alternative to the iterative approach presented in Section 3.8, the Performance Analysis tool can be used to better select the optimal value of R_s. By applying the Performance Analysis tool to the initial sweep results presented in Figure 3-12, it is possible to graph settling error as a function of the value of R_s. To launch the Performance Analysis tool, go to Trace ➔ Performance Analysis... using the menu in the upper left of the PSpice for TI simulation window. The Performance Analysis tool is shown in Figure 3-15.

Use the Wizard to create a Performance Analysis Trace. When prompted to choose a measurement, select YatX as shown in Figure 3-16. When prompted to complete the measurement expression, enter V(Vch, Voa_ss) as the name of the trace to search and 3.12 µs as the X value at which to measure the Y value as shown in Figure 3-17. Note that this measurement expression measures the Y value of the V_error trace at 3.12 µs, which is just past the end of an acquisition window. Thus, this measurement expression measures the settling error.

When finished using the Wizard, a Performance Analysis Trace of settling error as a function of the value of R_s should appear as shown in Figure 3-18. Note that the settling error scale has been adjusted to show a range of -2 mV to +2 mV. The settling error scale can be further adjusted as desired by double-clicking on the vertical axis and modifying the Data Range.

Examining the Performance Analysis Trace shows that the settling error is approximately zero when the value of R_s is roughly 26 Ω and again when the value of R_s is roughly 44 Ω. However, Figure 3-12 shows that smaller values of R_s tend to result in greater oscillations of V_error than larger values of R_s. Greater oscillations of V_error are not ideal because small changes to the duration of the acquisition window can lead to large variations in settling error. Additionally, the slope of the Performance Analysis Trace at values of R_s near 26 Ω is steeper than the slope of the Performance Analysis Trace at values of R_s near 44 Ω. A steeper slope of the Performance Analysis Trace is not preferable because small variations in the value of R_s can lead to large changes in settling error. For this reason, values of R_s around 26 Ω are ignored in favor of values of R_s around 44 Ω.

Based on the Performance Analysis Trace, any value of R_s between 24 Ω and 56 Ω inclusive is expected to provide adequate settling. As a result, a final value of R_s can be selected from the set of standard resistor values in the range from 24 Ω to 56 Ω inclusive with a preference for standard resistor values near 44 Ω. The optimal value of R_s is the standard resistor value in this range that provides the least settling error. According to the Performance Analysis Trace, the 5% standard resistor value in this range that provides the least settling error is is 43 Ω. Thus, the final value of R_s is selected as the optimal value of 43 Ω.