Configuration

The key parameters when setting up envelope modeling are the fundamental tone, harmonic order, and step size. To speed up the simulation, you can swap the simulation step size and the total number of simulation frequencies.

For example, if there are two large input signals with a bandwidth of 100 MHz each, with a center frequency of 10 GHz and 10.1 GHz, respectively, then you can simulate these two signals using two separate fundamental tones of [10 10.1] GHz. Each tone has a harmonic order of `3 (total of 25 simulation frequencies), and the simulation step size is 1/200 MHz = 5 ns.

It is also possible to configure the radio frequency subsystem so that both signals are in the same simulated bandwidth with a central frequency of 10.05 GHz. In this case, the harmonic order is set to 3 (total of 4 simulation frequencies), and the simulation step size is 1/400 MHz = 2'.5 ns. The latter configuration is faster, since the number of simulation frequencies is 3 times smaller, and the simulation step size is only '2` times smaller.

When setting up an envelope simulation, avoid overlapping envelopes. The thermal noise generated by the passive components is accounted for separately in each subband, which makes it possible to overlap individual envelopes.

The simulation step size should be small enough to take into account the bandwidth of the signal and the in-band growth of the spectrum.

For example, the complex input signal has a sampling frequency of 10 MHz. The minimum time step required to simulate this signal is '1/20` MHz = 50 ns. You can use a resampling factor from 4 to 8, which corresponds to a simulation time step from 25 ns to 12.5 ns. This allows us to detect the growth of the spectrum caused by nonlinear effects.

By default, the block Configuration allows you to automatically interpolate a base signal with a lower frequency into a radio frequency signal with a higher frequency. If you disable this property, it is recommended to use the same step size as the input signals. The input port resamples the input signal with the step size specified in the block Configuration. Using the same step size avoids the undesirable effects of spectrum overlap. It is best to resample the input signals before importing them into the library. RF Blockset using analog (continuous time) or digital (discrete time) interpolation filters.

The circuit envelope solver in the library RF Blockset performs the solution of a set of nonlinear equations based on a set of system variables. These system variables are determined from the circuit topology and simulation frequencies. Relative tolerance and absolute tolerance are used to minimize the convergence error of system variables. The number of iterations used in each time step significantly affects the speed of solutions and the trade-off between accuracy and speed. This compromise is governed by the criterion of stopping iterations. This stopping criterion is based on three sub-criteria:

Stop calculations if the first two sub-criteria are met or the last sub-criterion is met. If only one of the sub-criteria is met, an error is returned stating that the nonlinear solver did not work.