Running AC Analysis

Transient analysis solves for the evolution of a circuit's currents and voltages over time. We will showcase here the steps to run a transient analysis, using the circuit from the Butterworth filter example. If you have not already, we recommend reviewing Running Transient Analysis, and Running Parametric Sweeps as we will be building on top of the concepts covered within.

As a first step, we load the SPICE file and prepare a parameterization. We will here sweep parameters setting the input voltage to this op-amp, looking to see the output voltage follow the input voltage, but saturating at some point based on the properties of the amplifier:

using CedarEDA
# Import some useful suffixes for SI units
using CedarEDA.SIFactors: p, n, u, m, k, M, G

sm = SimManager(joinpath(@__DIR__, "butterworth.spice"))
sp = SimParameterization(sm;
    # Sweep over these SPICE .parameters
    params = ProductSweep(
        l1 = 0.5n:0.25u:2.3u, # 10 values (step is 0.25u)
        c2 = 100p:50p:800p, # 15 values (step is 50p)
        l3 = 100n:50n:500n, # 9 values (step is 50n)
    ),
)

SimParameterization with parameterization:
 - c2 (15 values: 100p .. 800p)
 - l1 (10 values: 500p .. 2.2505u)
 - l3 (9 values: 100n .. 500n)
Total: 1350 simulations to sweep over.

We will use explore() to visualize the frequency response, and so we will mark the signals we are interested in visualizing:

# Add signals to be automatically plotted
set_saved_signals!(sp, [
    sp.probes.node_vin,
    sp.probes.node_vout,
]);

# Run the AC simulations with 40 points per decade from 10kHz to 1GHz
ac_sol = ac!(sp, acdec(40, 10k, 1G))

explore(sp, ac_sol)

Because this is a third-order butterworth filter, we expect the filter to drop roughly three orders of magnitude for every decade of frequency increase. Inspecting the figure above, we see this is indeed the case.

AC solution objects

Just as with a transient analysis solution, an AC analysis solution contains within it the parameters and op properties, however here we will focus on the ac property:

julia> ac_sol.acFrequency Response with 10 properties:
  cc2
  gnd
  ll1
  ll3
  node_0
  node_n1
  node_vin
  node_vout
  rr4
  v1

When we index the ac property, pulling out a solution for a particular parameterization point, the return value is a complex-valued Signal, representing the frequency domain. A common task is to visualize the magnitude or phase of the signal, which can be done via the abs() and angle() methods in Julia. As showcased in Working with Plots, the inspect() function can take in arrays of signals to plot. Here, we demonstrate how to visualize the effect that varying the first parameter has on our output:

inspect([abs(sig) for sig in ac_sol.ac.node_vout[1, 1, :]])