Introduction

Interaction with SPICE-like simulators.

The library support the generation of netlists from so-called meta-netlists, running SPICE-like simulators and reading back the results. Currently the following simulators are supported:

• ngspice

• Xyce

These capabilities are implemented in the objects NGSpice and Xyce respectively.

A meta-netlist is a netlist with meta-netlist variables denoted by alphanumerical variables surrounded by double dollers $$var$$. This allows, for example, to

• Check performance across corners.

• Sweep parameters in netlist places where SPICE expressions/variables are not permitted.

• Run multiple simulation types without manually editing the netlist.

• …

Given a meta-netlist with variables s1_dc, s1_mag, sim_cmd and sim_output, a basic usage example is as follows:

>>> import matplotlib.pyplot as plt
>>> from fbespice import *
>>> sim_vars = { 's1_dc' : 0.65,
                 's1_mag' : 0.4,
                 's1_freq' : 100e6}
>>> tran_vars = dict(sim_vars,
      **{'sim_cmd' : " ".join([".tran",
                       str(0.1/sim_vars['s1_freq']),
                       str(100/sim_vars['s1_freq'])]),
         'sim_output' : ".print tran v(out)"})
>>> xyce = Xyce("path/to/testbench")
>>> xyce.netlist_from_meta(tran_vars)
>>> df_tr = xyce.run().read_results(XyceResID.TRAN)
>>> ac_vars = dict(sim_vars,
      **{'sim_cmd' : ".ac dec 5 1k 1e9",
         'sim_output' : ".print ac vdb(out) vp(out)"})
>>> xyce.netlist_from_meta(ac_vars)
>>> df_ac = xyce.run().read_results(XyceResID.AC)
>>> plt.plot(df_tr['time'], df_tr['v(out)'])
>>> plt.semilogx(df_ac['freq'], df_ac['vdb(out)'])

ngspice simulation results are expected in raw format and can be read, on top of the NGSpice.read_results(), also by read_raw(). The library also supports reaing Xyre results in raw format. However, given that Xyce can only output results in raw format for a subset of its simulation capabilities, Xyce results are expected in it’s native STD format by Xyce.read_results() and can also be read with read_xyce().