-
Notifications
You must be signed in to change notification settings - Fork 0
/
noisedown8and6
413 lines (255 loc) · 12.7 KB
/
noisedown8and6
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
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
#TODO: add support for ADC histogram plotting.
#TODO: add support for determining ADC input level
import corr,time,numpy,struct,sys,logging,pylab,matplotlib
import matplotlib.pyplot as plt
import cmath
import math
import numpy as np
import time
from scipy.fftpack import fft, ifft
bitstream = 'ramtestreg_2017-04-12_1129.bof'
katcp_port=7147
from SNAPsynth import LMX2581
fpga=LMX2581('rpi3-2')
time.sleep(0.1)
print fpga.listdev()
print fpga.listbof()
fpga.write_int('fft_shift',2**11-1)
fpga.write_int('adc16_use_synth',0)
fpga.write_int('sync_gen2_sync',1)
fpga.write_int('sync_gen1_sync_period_var',2**16-2)
fpga.write_int('sync_gen1_sync_period_sel',1)
print fpga.est_brd_clk()
fpga.write_int('cnt_rst',1)
print fpga.read_int('acc_len')
acc_cnt=fpga.read_int('acc_cnt')
acc_cnt_loop=fpga.read_int('acc_cnt')
print fpga.read_int('acc_cnt')
fpga.write_int('cnt_rst',0)
time.sleep(5)
print time.time()
while (acc_cnt==acc_cnt_loop):
acc_cnt_loop=fpga.read_int('acc_cnt')
print fpga.read_int('acc_cnt')
time.sleep(0.001)
print time.time()
print fpga.read_int('acc_cnt')
print time.time()
### Parameter Setting ###
j=cmath.sqrt(-1)
fpga.write_int('acc_len',2*10**5)
Fs=250*10**6
length_acc=10
a=np.arange(1024)
a=a[28:233]
slice_length=len(a)
### Define Arrays ###
autocor_fft3out2_mean=np.arange(length_acc,dtype=np.float64)
autocor_fft3out2_sliced=np.arange(slice_length,dtype=np.float64)
auto_div_cross=np.zeros(1024,dtype=np.float64)
t_auto=np.arange(length_acc,dtype=np.float64)
division_auto=np.zeros(length_acc,dtype=np.float64)
autocor_fft3out2_8=np.zeros(1024,dtype=np.float64)
autocor_fft3out2_sum_8=np.zeros(1024,dtype=np.float64)
autocor_fft3out28_sliced=np.arange(slice_length,dtype=np.float64)
autocor_fft3out28_sum=np.arange(slice_length,dtype=np.float64)
cor_fft3out1_fft3out2_sum_re=np.zeros(1024,dtype=np.float64)
cor_fft3out1_fft3out2_mean_re=np.arange(length_acc,dtype=np.float64)
cor_fft3out1_fft3out2_mean_im=np.arange(length_acc,dtype=np.float64)
t=np.arange(length_acc,dtype=np.float64)
cor_fft3out1_fft3out2_mean=np.arange(length_acc,dtype=np.float64)
cor_fft3out1_fft3out2_re=np.zeros(1024,dtype=np.float64)
cor_fft3out1_fft3out2_im=np.zeros(1024,dtype=np.float64)
cor_fft3out1_fft3out2_sum=np.arange(slice_length,dtype=np.float64)
cor_fft3out1_fft3out2=np.arange(slice_length,dtype=np.float64)
division=np.zeros(length_acc,dtype=np.float64)
cor_fft3out1_fft3out2_sliced_re=np.arange(slice_length,dtype=np.float64)
cor_fft3out1_fft3out2_sliced_im=np.arange(slice_length,dtype=np.float64)
cor_fft3out1_fft3out2_re=np.zeros(1024,dtype=np.float64)
cor_fft3out1_fft3out2_im=np.zeros(1024,dtype=np.float64)
cor_fft3out1_fft3out2_mag=np.arange(slice_length,dtype=np.float64)
cor_fft3out1_fft3out2_mag_sliced=np.arange(slice_length,dtype=np.float64)
cor_fft3out1_fft3out2_sum=np.arange(slice_length,dtype=np.float64)
cor_fft3out1_fft3out2_sum_1=np.zeros(1024,dtype=np.float64)
cor_fft3out1_fft3out2_mag_1=np.zeros(1024,dtype=np.float64)
cor_fft3out1_fft3out2_re=np.zeros(1024,dtype=np.float64)
cor_fft3out1_fft3out2_im=np.zeros(1024,dtype=np.float64)
cor_fft3out1_fft3out2_mag_momentary=np.zeros(1024,dtype=np.float64)
cor_fft3out1_fft3out2_sum_re_sq=np.zeros(1024,dtype=np.float64)
cor_fft3out1_fft3out2_sum_im_sq=np.zeros(1024,dtype=np.float64)
cor_fft3out1_fft3out2_mag_momentary_1=np.zeros(1024,dtype=np.float64)
cor_fft3out1_fft3out2_mag_1=np.zeros(1024,dtype=np.float64)
cor_fft3out1_fft3out2_mean_mag=np.arange(length_acc,dtype=np.float64)
autocor_fft3out2_sum_1_sq=np.zeros(1024,dtype=np.float64)
autocor_fft3out2_sum_8_sq=np.zeros(1024,dtype=np.float64)
multi_mean_auto=np.zeros(1024,dtype=np.float64)
corr_phase=np.zeros(1024,dtype=np.float64)
### First value of Sum Autocorrelation ###
autocor_fft3out2=struct.unpack('>1024q',fpga.read('Correlator_fft2_out1_Memory2_SharedBRAM3_wideband',8192,0))
autocor_fft3out2_sliced=autocor_fft3out2[28:233]
autocor_fft3out2_sum=autocor_fft3out2_sliced
### First value of Sum Crosscorrelation ###
cor_fft3out1_fft3out2_re=np.array(struct.unpack('>1024q',fpga.read('Correlator_fft2_out2_Memory1_SharedBRAM1_wideband',8192,0)),dtype=np.float64)
cor_fft3out1_fft3out2_im=np.array(struct.unpack('>1024q',fpga.read('Correlator_fft2_out2_Memory1_SharedBRAM2_wideband',8192,0)),dtype=np.float64)
autocor_fft3out2_1=struct.unpack('>1024q',fpga.read('Correlator_fft2_out2_Memory2_SharedBRAM1_wideband',8192,0))
autocor_fft3out2_8=struct.unpack('>1024q',fpga.read('Correlator_fft3_out1_Memory1_SharedBRAM3_wideband',8192,0))
cor_fft3out1_fft3out2_sum_re=cor_fft3out1_fft3out2_re
cor_fft3out1_fft3out2_sum_im=cor_fft3out1_fft3out2_im
autocor_fft3out2_sum_1=autocor_fft3out2_1
autocor_fft3out2_sum_8=autocor_fft3out2_8
### Open Files ###
cross_sum_re=open('cross_sum_re.txt','w')
cross_sum_im=open('cross_sum_im.txt','w')
auto_sum_1=open('auto_sum_1.txt','w')
auto_sum_8=open('auto_sum_8.txt','w')
mean_div_acclen_cross=open('mean_div_acclen_cross.txt','w')
mean_div_acclen_auto=open('mean_div_acclen_auto.txt','w')
auto_cross_ratio=open('auto_cross_ratio.txt','w')
phase_cross=open('phase_cross.txt','w')
cross_sum_mag=open('cross_sum_mag.txt','w')
auto_mean_6=open('auto_mean_6.txt','w')
cross_mean=open('cross_mean.txt','w')
### For Loop Start ###
for m in range (1,length_acc):
print 'Loop value:' ,m
acc_cnt=fpga.read_int('acc_cnt')
acc_cnt_loop=fpga.read_int('acc_cnt')
print fpga.read_int('acc_cnt')
while (acc_cnt==acc_cnt_loop):
acc_cnt_loop=fpga.read_int('acc_cnt')
print fpga.read_int('acc_cnt')
time.sleep(0.001)
print time.time()
cor_fft3out1_fft3out2_re=np.array(struct.unpack('>1024q',fpga.read('Correlator_fft2_out2_Memory1_SharedBRAM1_wideband',8192,0)),dtype=np.float64)
cor_fft3out1_fft3out2_im=np.array(struct.unpack('>1024q',fpga.read('Correlator_fft2_out2_Memory1_SharedBRAM2_wideband',8192,0)),dtype=np.float64)
print fpga.read_int('acc_cnt')
autocor_fft3out2_1=struct.unpack('>1024q',fpga.read('Correlator_fft2_out2_Memory2_SharedBRAM1_wideband',8192,0))
autocor_fft3out2_8=struct.unpack('>1024q',fpga.read('Correlator_fft3_out1_Memory1_SharedBRAM3_wideband',8192,0))
### Concatenate crosscorrelation ###
cor_fft3out1_fft3out2_sum_re_pre=cor_fft3out1_fft3out2_sum_re # [m-1]th value of the summation of sum of the real part of the cross
cor_fft3out1_fft3out2_sum_im_pre=cor_fft3out1_fft3out2_sum_im # [m-1]th value of the summation of sum of the imaginary part of the cross
cor_fft3out1_fft3out2_sum_re_con=np.concatenate([cor_fft3out1_fft3out2_sum_re_pre,cor_fft3out1_fft3out2_sum_re])
cor_fft3out1_fft3out2_sum_im_con=np.concatenate([cor_fft3out1_fft3out2_sum_im_pre,cor_fft3out1_fft3out2_sum_im])
print cor_fft3out1_fft3out2_sum_re_con
np.savetxt('cross_sum_re' , cor_fft3out1_fft3out2_sum_re_con)
np.savetxt('cross_sum_im' , cor_fft3out1_fft3out2_sum_re_con)
### Sum the real and imginary part of Cross and take the magnitude ###
cor_fft3out1_fft3out2_sum_re=np.add(cor_fft3out1_fft3out2_sum_re,cor_fft3out1_fft3out2_re)
cor_fft3out1_fft3out2_sum_im=np.add(cor_fft3out1_fft3out2_sum_im,cor_fft3out1_fft3out2_im)
cor_fft3out1_fft3out2_sum_re_con=np.concatenate([cor_fft3out1_fft3out2_sum_re_pre,cor_fft3out1_fft3out2_sum_re])
cor_fft3out1_fft3out2_sum_im_con=np.concatenate([cor_fft3out1_fft3out2_sum_im_pre,cor_fft3out1_fft3out2_sum_im])
complex_val = cor_fft3out1_fft3out2_sum_re + 1j*cor_fft3out1_fft3out2_sum_im
cor_fft3out1_fft3out2_mag = np.abs(complex_val)
cor_fft3out1_fft3out2_mag_momentary=cor_fft3out1_fft3out2_mag[28:233]
cor_fft3out1_fft3out2_mean_mag[m]=np.mean(cor_fft3out1_fft3out2_mag_momentary)
t[m]=m*1*10**5
division[m]=float(cor_fft3out1_fft3out2_mean_mag[m])/float(t[m])
### Concatenate the Autocorrelation ###
autocor_fft3out2_sum_1_pre=autocor_fft3out2_sum_1 # [m-1]th value of the summation for autocorrelation
autocor_fft3out2_sum_8_pre=autocor_fft3out2_sum_8 # [m-1]th value of the summation for autocorrelation
autocor_fft3out2_sum_1=np.add(autocor_fft3out2_1,autocor_fft3out2_sum_1)
autocor_fft3out2_sum_8=np.add(autocor_fft3out2_8,autocor_fft3out2_sum_8)
autocor_fft3out2_sum_1_con=np.concatenate([autocor_fft3out2_sum_1_pre,autocor_fft3out2_sum_1])
autocor_fft3out2_sum_8_con=np.concatenate([autocor_fft3out2_sum_8_pre,autocor_fft3out2_sum_8])
np.savetxt('auto_sum_1' , autocor_fft3out2_sum_1_con)
np.savetxt('auto_sum_8' , autocor_fft3out2_sum_8_con)
### Slice the autocorrelation,add and take the mean ###
autocor_fft3out2_1_sliced=autocor_fft3out2_1[28:233]
autocor_fft3out2_sum=np.add(autocor_fft3out2_sliced,autocor_fft3out2_sum)
autocor_fft3out2_mean[m]=np.mean(autocor_fft3out2_sum)
t_auto[m]=m*1.*10**5
division_auto[m]=autocor_fft3out2_mean[m]/float(t_auto[m])
cor_fft3out1_fft3out2_sliced_re=cor_fft3out1_fft3out2_sum_re[28:233]
cor_fft3out1_fft3out2_sliced_im=cor_fft3out1_fft3out2_sum_im[28:233]
### Magnitude of Summation Crosscorrelation ###
complex_val = cor_fft3out1_fft3out2_sum_re + 1j*cor_fft3out1_fft3out2_sum_im
cor_fft3out1_fft3out2_mag = np.abs(complex_val)
### Phase of the Crosscorrelation ###
corr_phase=numpy.arctan2(cor_fft3out1_fft3out2_sum_re,cor_fft3out1_fft3out2_sum_im)
### Multiplicative Mean of Autocorretlation ###
autocor_fft3out2_sum_1_sq=np.sqrt(autocor_fft3out2_sum_1)
autocor_fft3out2_sum_8_sq=np.sqrt(autocor_fft3out2_sum_8)
multi_mean_auto=np.multiply(autocor_fft3out2_sum_1_sq,autocor_fft3out2_sum_8_sq)
### Autocorrelation / Crosscorrelation ###
for m in range (1,1024):
auto_div_cross[m]=float(multi_mean_auto[m])/float(cor_fft3out1_fft3out2_mag[m])
### Save Text ###
np.savetxt('mean_div_acclen_cross' , division)
np.savetxt('mean_div_acclen_auto' , division_auto)
np.savetxt('auto_cross_ratio' , auto_div_cross)
np.savetxt('phase_cross' , corr_phase)
np.savetxt('cross_sum_mag' , cor_fft3out1_fft3out2_mag)
np.savetxt('auto_mean_6' , autocor_fft3out2_mean)
np.savetxt('cross_mean' , cor_fft3out1_fft3out2_mean_mag)
### Frequency Axis ###
cor_fft3out1_fft3out2_len=len(cor_fft3out1_fft3out2_mag)
print cor_fft3out1_fft3out2_len
cor_fft3out1_fft3out2_freq=np.linspace(0,Fs,cor_fft3out1_fft3out2_len)
cor_fft3out1_fft3out2_freq_sliced=cor_fft3out1_fft3out2_freq[28:233]
cor_fft3out1_fft3out2_freq_sliced_len=len(cor_fft3out1_fft3out2_freq_sliced)
print cor_fft3out1_fft3out2_freq_sliced_len
autocor_fft3out2_len=len(autocor_fft3out2)
autocor_fft3out2_freq=np.linspace(0,Fs,autocor_fft3out2_len)
### Seperation of first values ###
auto_div_cross_1=auto_div_cross[3:1024]
autocor_fft3out2_freq_1=autocor_fft3out2_freq[3:1024]
print auto_div_cross_1[152]
print auto_div_cross_1[151]
print auto_div_cross_1[150]
#division=division[2:]
#division_auto=division_auto[2:]
#t_1=t[2:]
#t_auto=t_auto[2:]
print autocor_fft3out2_sum_1[367]
print autocor_fft3out2_sum_1[368]
print autocor_fft3out2_sum_1[369]
print autocor_fft3out2_sum_1[370]
print autocor_fft3out2_sum_1[371]
plt.subplot(3,3,1)
plt.plot(t[1:],division[1:])
plt.ylabel('Mean of Crosscorrelation divided by Accumulation length')
plt.xlabel('Accumulation Length')
plt.title('Mean of Crosscorrelation/Accumulation length vs Accumulation length')
plt.subplot(3,3,2)
plt.plot(t_auto[1:],division_auto[1:])
plt.ylabel('Mean of Autocorrelation divided by Accumulation length')
plt.xlabel('Accumulation Length')
plt.title('Mean of Autocorrelation/Accumulation length vs Accumulation length')
plt.subplot(3,3,3)
plt.semilogy(cor_fft3out1_fft3out2_freq,cor_fft3out1_fft3out2_mag)
plt.ylabel('Crosscorrelation of Input8 & Input6')
plt.xlabel('Frequency(Hz)')
plt.title('Crosscorrelation of Input8 & Input6 vs Frequency')
plt.subplot(3,3,4)
plt.semilogy(autocor_fft3out2_freq,autocor_fft3out2_sum_1)
plt.ylabel('Autocorrelation of Input6')
plt.xlabel('Frequency(Hz)')
plt.title('Autocorrelation of Input6 vs Frequency')
plt.subplot(3,3,5)
plt.semilogy(autocor_fft3out2_freq,autocor_fft3out2_sum_8)
plt.ylabel('Autocorrelation of Input8')
plt.xlabel('Frequency(Hz)')
plt.title('Autocorrelation of Input8 vs Frequency')
plt.subplot(3,3,6)
plt.semilogy(autocor_fft3out2_freq_1,auto_div_cross_1)
plt.ylabel('Autocorrelation/Crosscorrelation')
plt.xlabel('Frequency(Hz)')
plt.title('Autocorrelation/Crosscorrelation vs Frequency')
plt.subplot(3,3,7)
plt.plot(t,cor_fft3out1_fft3out2_mean)
plt.ylabel('Mean of Crosscorrelation')
plt.xlabel('Frequency(Hz)')
plt.title('Mean of Crosscorrelation vs Frequency')
plt.subplot(3,3,8)
plt.plot(cor_fft3out1_fft3out2_freq,corr_phase)
plt.xlabel('Frequency(Hz)')
plt.ylabel('Phase of Crosscorrelation')
plt.title('Phase of Crosscorrelation vs Frequency')
plt.subplot(3,3,9)
plt.plot(autocor_fft3out2_freq[3:],cor_fft3out1_fft3out2_sum_re[3:])
plt.plot(autocor_fft3out2_freq[3:],cor_fft3out1_fft3out2_sum_im[3:])
plt.ylabel('cor_fft3out1_fft3out2_sum_im')
plt.ylabel('Crosscorrelation Real and Imaginary Parts')
plt.xlabel('Frequency(Hz)')
plt.title('Crosscorrelation Real and Imaginary Parts vs Frequency')
plt.show()