Setting the Error Tolerance

For Interpolating sweeps.

The Error tolerance value is the maximum relative difference allowed between two successive interpolation solutions. Because interpolating sweeps converge non-uniformly, that is, the error will jump around, false convergence is possible. To prevent this, a secondary check based on the Loewner matrix must also succeed: the condition of the Loewner matrix must also be less than the error tolerance.

The default 0.5 percent for interpolating sweeps is usually satisfactory. However, if the Interpolated sweep Advanced Options include Enforce Causality and Enforce Passivity, you may need to specify a lower tolerance (for example, 0.1). This is because of the sensitivity of Enforce Causality and Passivity.

To set the error tolerance for an Interpolating sweep:

1. Open the Edit Sweep dialog box (by either viewing the properties of an existing Sweep or by Adding a Frequency sweep to an existing Setup).
2. Type a value in the Error Tolerance box.

If you expect to use the same settings often, you can click Set Defaults. To use previously saved settings, click Use Defaults.