1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
|
import os
import sys
import tempfile
import unittest
from pathlib import Path
import numpy as np
import pytest
import skrf
class VectorFittingTestCase(unittest.TestCase):
@pytest.mark.filterwarnings('ignore::UserWarning')
def test_ringslot_with_proportional(self):
# perform the fit
nw = skrf.data.ring_slot
vf = skrf.vectorFitting.VectorFitting(nw)
vf.vector_fit(n_poles_real=2, n_poles_cmplx=0, fit_proportional=True, fit_constant=True)
self.assertLess(vf.get_rms_error(), 0.02)
@pytest.mark.filterwarnings('ignore::UserWarning')
def test_ringslot_default_log(self):
# perform the fit without proportional term
nw = skrf.data.ring_slot
vf = skrf.vectorFitting.VectorFitting(nw)
vf.vector_fit(n_poles_real=4, n_poles_cmplx=0, init_pole_spacing='log')
self.assertLess(vf.get_rms_error(), 0.01)
@pytest.mark.filterwarnings('ignore::UserWarning')
def test_ringslot_without_prop_const(self):
# perform the fit without proportional term
nw = skrf.data.ring_slot
vf = skrf.vectorFitting.VectorFitting(nw)
vf.vector_fit(n_poles_real=4, n_poles_cmplx=0, fit_proportional=False, fit_constant=False)
self.assertLess(vf.get_rms_error(), 0.01)
@pytest.mark.filterwarnings('ignore::UserWarning')
def test_ringslot_custompoles(self):
# perform the fit with custom initial poles
nw = skrf.data.ring_slot
vf = skrf.vectorFitting.VectorFitting(nw)
vf.poles = 2 * np.pi * np.array([-100e9, -10e9 + 100e9j])
vf.vector_fit(init_pole_spacing='custom')
self.assertLess(vf.get_rms_error(), 0.01)
def test_190ghz_measured(self):
# perform the fit without proportional term
s2p_file = Path(__file__).parent.parent.parent / 'doc/source/examples/vectorfitting/190ghz_tx_measured.S2P'
nw = skrf.network.Network(s2p_file)
vf = skrf.vectorFitting.VectorFitting(nw)
vf.vector_fit(n_poles_real=4, n_poles_cmplx=4, fit_proportional=False, fit_constant=True)
self.assertLess(vf.get_rms_error(), 0.02)
def test_no_convergence(self):
s2p_file = Path(__file__).parent.parent.parent / 'doc/source/examples/vectorfitting/190ghz_tx_measured.S2P'
# perform a bad fit that does not converge and check if a RuntimeWarning is given
nw = skrf.network.Network(s2p_file)
vf = skrf.vectorFitting.VectorFitting(nw)
with pytest.warns(RuntimeWarning) as record:
vf.vector_fit(n_poles_real=0, n_poles_cmplx=5, fit_proportional=False, fit_constant=True)
assert len(record) == 1
def test_dc_enforcement(self):
# perform the fit on data including a dc sample (0 Hz)
s4p_file = Path(__file__).parent / 'cst_example_4ports.s4p'
nw = skrf.Network(s4p_file)
vf = skrf.VectorFitting(nw)
vf.auto_fit()
# rough check on fit quality
self.assertLess(vf.get_rms_error(), 0.2)
# evaluate model error at the dc point (real part)
# the dc point should always be real (it still often has a tiny imaginary part)
vf_fit = np.empty(nw.nports ** 2, dtype=complex)
abs_real_errors = np.empty(nw.nports ** 2)
for i in range(nw.nports):
for j in range(nw.nports):
vf_ij = vf.get_model_response(i, j, nw.f[0])
vf_fit[i * nw.nports + j] = vf_ij
abs_real_errors[i * nw.nports + j] = np.abs(np.real(vf_ij - nw.s[0, i, j]))
self.assertTrue(np.all(abs_real_errors < 1e-10))
def test_read_write_npz(self):
# fit ring slot example network
nw = skrf.data.ring_slot
vf = skrf.vectorFitting.VectorFitting(nw)
with pytest.warns(UserWarning) as record:
vf.vector_fit(n_poles_real=3, n_poles_cmplx=0)
self.assertTrue(len(record) == 1)
# export (write) fitted parameters to .npz file in tmp directory
with tempfile.TemporaryDirectory() as name:
vf.write_npz(name)
# create a new vector fitting instance and import (read) those fitted parameters
vf2 = skrf.vectorFitting.VectorFitting(nw)
vf2.read_npz(os.path.join(name, f'coefficients_{nw.name}.npz'))
# compare both sets of parameters
self.assertTrue(np.allclose(vf.poles, vf2.poles))
self.assertTrue(np.allclose(vf.residues, vf2.residues))
self.assertTrue(np.allclose(vf.proportional_coeff, vf2.proportional_coeff))
self.assertTrue(np.allclose(vf.constant_coeff, vf2.constant_coeff))
@pytest.mark.skipif("matplotlib" in sys.modules, reason="Raise Error only if matplotlib is not installed.")
def test_matplotlib_missing(self):
vf = skrf.vectorFitting.VectorFitting(skrf.data.ring_slot)
with self.assertRaises(RuntimeError):
vf.plot_convergence()
def test_passivity_enforcement(self):
vf = skrf.VectorFitting(None)
# non-passive example parameters from Gustavsen's passivity assessment paper:
vf.poles = np.array([-1, -5 + 6j])
vf.residues = np.array([[0.3, 4 + 5j], [0.1, 2 + 3j], [0.1, 2 + 3j], [0.4, 3 + 4j]])
vf.constant_coeff = np.array([0.2, 0.1, 0.1, 0.3])
vf.proportional_coeff = np.array([0.0, 0.0, 0.0, 0.0])
# test if model is not passive
violation_bands = vf.passivity_test()
self.assertTrue(np.allclose(violation_bands, np.array([4.2472, 16.434]) / 2 / np.pi, rtol=1e-3, atol=1e-3))
self.assertFalse(vf.is_passive())
# enforce passivity with default settings
vf.passivity_enforce(f_max=2)
# check if model is now passive
self.assertTrue(vf.is_passive())
def test_autofit(self):
vf = skrf.VectorFitting(skrf.data.ring_slot)
vf.auto_fit()
self.assertTrue(vf.get_model_order(vf.poles) == 6)
self.assertTrue(np.sum(vf.poles.imag == 0.0) == 0)
self.assertTrue(np.sum(vf.poles.imag > 0.0) == 3)
self.assertLess(vf.get_rms_error(), 1e-05)
suite = unittest.TestLoader().loadTestsFromTestCase(VectorFittingTestCase)
unittest.TextTestRunner(verbosity=2).run(suite)
|
