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
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
|
import json, sys, os.path, configparser
import numpy as np
import matplotlib.pyplot as plt
from scipy.signal import butter, sosfilt, freqz
CONFDIR = os.path.join(os.path.dirname(__file__), "../conf")
# This information is not in the blackbox file
DEFAULT_AMP_GAIN = 18.50
AMP_GAIN = {
"apple,j180": 16.0,
"apple,j313": 16.0,
"apple,j274": 21.0,
"apple,j375": 21.0,
"apple,j473": 21.0,
"apple,j474": 21.0,
"apple,j475": 21.0,
}
def db(x):
return 10 ** (x / 20)
def smooth(a, n=3):
l = len(a)
ret = np.cumsum(a, dtype=float)
ret[n:] = ret[n:] - ret[:-n]
ret = ret[n - 1:] / n
pad = l - len(ret)
return np.pad(ret, (pad//2, (pad + 1)//2), "edge")
def butter_lowpass(cutoff, fs, order=5):
return butter(order, cutoff, fs=fs, btype='low', output="sos", analog=False)
def butter_highpass(cutoff, fs, order=5):
return butter(order, cutoff, fs=fs, btype='high', output="sos", analog=False)
def butter_lowpass_filter(data, cutoff, fs, order=5):
sos = butter_lowpass(cutoff, fs, order=order)
y = sosfilt(sos, data)
return y
def butter_highpass_filter(data, cutoff, fs, order=5):
sos = butter_highpass(cutoff, fs, order=order)
y = sosfilt(sos, data)
return y
def pilot_filter(data, fs):
data = butter_lowpass_filter(data, 100, fs, 6)
return butter_highpass_filter(data, 10, fs, 3)
class Model:
def __init__(self, idx, an, name, conf):
self.idx = idx
self.an = an
self.name = name
self.conf = conf
self.tr_coil = float(conf["tr_coil"])
self.tr_magnet = float(conf["tr_magnet"])
self.tau_coil = float(conf["tau_coil"])
self.tau_magnet = float(conf["tau_magnet"])
self.t_limit = float(conf["t_limit"])
self.t_headroom = float(conf["t_headroom"])
self.z_nominal = float(conf["z_nominal"])
self.z_shunt = float(conf.get("z_shunt", 0))
self.is_scale = float(conf["is_scale"])
self.vs_scale = float(conf["vs_scale"])
self.a_t_20c = float(conf["a_t_20c"])
self.a_t_35c = float(conf["a_t_35c"])
self.is_chan = int(conf["is_chan"])
self.vs_chan = int(conf["vs_chan"])
self.t_ambient = an.fdr["t_ambient"]
self.t_coil = an.fdr["blocks"][0]["speakers"][self.idx]["t_coil"]
self.t_magnet = an.fdr["blocks"][0]["speakers"][self.idx]["t_magnet"]
self.m_x = []
self.m_t_coil_tg = []
self.m_t_coil = []
self.m_t_magnet_tg = []
self.m_t_magnet = []
self.l_x = []
self.l_t_coil = []
self.l_t_magnet = []
def run_model(self):
off = 0
t = 0
for blk in self.an.fdr["blocks"]:
sr = blk["sample_rate"]
cnt = blk["sample_count"]
data = blk["speakers"][self.idx]
isense = self.an.cvr[off:off+cnt, self.is_chan] * self.is_scale
vsense = self.an.cvr[off:off+cnt, self.vs_chan] * self.vs_scale
dt = 1 / self.an.sr
alpha_coil = dt / (dt + self.tau_coil)
alpha_magnet = dt / (dt + self.tau_magnet)
self.l_x.append(t)
self.l_t_coil.append(data["t_coil"])
self.l_t_magnet.append(data["t_magnet"])
for x, (i, v) in enumerate(zip(isense, vsense)):
self.m_x.append(t + x / sr)
p = i * v
tvc_tgt = self.t_magnet + p * self.tr_coil
self.t_coil = tvc_tgt * alpha_coil + self.t_coil * (1 - alpha_coil)
tmag_tgt = self.t_ambient + p * self.tr_magnet
self.t_magnet = tmag_tgt * alpha_magnet + self.t_magnet * (1 - alpha_magnet)
self.m_t_coil_tg.append(tvc_tgt)
self.m_t_coil.append(self.t_coil)
self.m_t_magnet_tg.append(tmag_tgt)
self.m_t_magnet.append(self.t_magnet)
t += cnt / sr
off += cnt
def analyze(self, outfile):
plt.clf()
fig, ax1 = plt.subplots(figsize=(30,15))
ax1.set_title(self.name)
ax1.set_xlabel('time (s)')
ax1.set_ylabel('temperature')
ax1.tick_params(axis='y')
ax2 = ax1.twinx() # instantiate a second axes that shares the same x-axis
color = 'tab:red'
ax2.set_ylabel('power', color=color)
ax2.tick_params(axis='y', labelcolor=color)
ax1.plot(self.m_x, self.m_t_coil, "r")
# ax1.plot(self.m_x, smooth(self.m_t_coil_tg, 1000), "y")
ax1.plot(self.m_x, self.m_t_magnet, "b")
# ax1.plot(self.m_x, smooth(self.m_t_magnet_tg, 1000), "c")
ax1.plot(self.l_x, self.l_t_coil, "om")
ax1.plot(self.l_x, self.l_t_magnet, "og")
i = self.an.cvr[:, self.is_chan] * self.is_scale
v = self.an.cvr[:, self.vs_chan] * self.vs_scale
sr = self.an.fdr["sample_rate"]
ilp = pilot_filter(i, sr)
vlp = pilot_filter(v, sr)
p = butter_lowpass_filter(i * v, 10, sr, 1)
p = smooth(p, 4000)
plp = butter_lowpass_filter(ilp * vlp, 10, sr, 1)
plp = smooth(plp, 4000)
vlprms_sq = butter_lowpass_filter(vlp * vlp, 10, sr, 1)
vlprms_sq = smooth(vlprms_sq, 4000)
r = vlprms_sq / plp
# ax2.plot(self.m_x, p, "b")
rref = np.average(r[1 * sr:2 * sr])
print(f"Initial resistance: {rref} ohms")
# Clear out the first second, since it tends to contain garbage
r[:1*sr] = rref
#r = butter_lowpass_filter(r - rref, 2, sr, 2) + rref
a = self.a_t_35c # XXX why are there two values at different temperatures?
tref = self.l_t_coil[0]
t = ((r - self.z_shunt) / (rref - self.z_shunt) - 1) / a + tref
ax1.plot(self.m_x, t, "k")
ax2.plot(self.m_x, p, "r")
# ax2.plot(self.m_x, plp, "g")
# ax2.plot(self.m_x, vlprms_sq, "b")
for level in (-1000, -6, -10, -15):
gain = AMP_GAIN.get(self.an.fdr["machine"], DEFAULT_AMP_GAIN)
pbase = (db(gain - 30) ** 2) / (self.z_nominal + self.z_shunt)
ptest = (db(gain + level) ** 2) / (self.z_nominal + self.z_shunt)
p = pbase + ptest
ax2.axhline(y=p, color='r', linestyle='--')
fig.tight_layout() # otherwise the right y-label is slightly clipped
plt.savefig(outfile)
class Analyzer:
def __init__(self, base):
self.fdr = json.load(open(base + ".fdr"))
data = open(base + ".cvr", "rb").read()
cvr = np.frombuffer(data, dtype="int16").astype("float") / 32768
maker, model = self.fdr["machine"].split(",")
cf = os.path.join(CONFDIR, maker, model + ".conf")
print(f"Using config file: {cf}")
self.conf = configparser.ConfigParser()
self.conf.read(cf)
ch = int(self.conf["Globals"]["channels"])
samples = len(cvr) // ch
self.cvr = cvr.reshape((samples, ch))
print(f"Got {samples} samples ({ch} channels)")
assert ch == self.fdr["channels"]
self.sr = self.fdr["sample_rate"]
speaker_configs = []
for key in self.conf.sections():
if not key.startswith("Speaker/"):
continue
print(key)
name = key.split("/")[1]
speaker_configs.append((name, self.conf[key]))
# Match the order that speakersafetyd uses (by group)
speaker_configs.sort(key=lambda x: int(x[1]["group"]))
self.speakers = []
for i, cfg in enumerate(speaker_configs):
self.speakers.append(Model(i, self, cfg[0], cfg[1]))
def analyze(self):
for i, spk in enumerate(self.speakers):
print(f"Processing speaker {i}")
spk.run_model()
spk.analyze(f"speaker_{i}.png")
# break
if __name__ == "__main__":
a = Analyzer(sys.argv[1])
a.analyze()
|
