Recipes for recording from a single analog source

Recording a short segment from a digital line

Read 200 milliseconds worth of data from line “di0” with a sampling frequency of 50 kilohertz and display the results using matplotlib:

from picodaq import *
import matplotlib.pyplot as plt

with DigitalIn(line=0, rate=50*kHz) as di:
    data = di.read(200*ms)

plt.plot(data)
plt.xlabel('Sample #')
plt.ylabel('Binary value')

Retrieving timestamps

The read function can also return timestamps along with your data:

with DigitalIn(line=0, rate=50*kHz) as di:
    data, times = di.read(200*ms, times=True)

plt.plot(times, data)
plt.xlabel('Time (s)')
plt.ylabel('Binary value')

In this case, I had an approximately 100-Hz square TTL pulse train connected to the input:

Reading a short segment from two lines

Read 200 ms worth of data from channels “di0” and “di1” with a sampling frequency of 50 kilohertz and display the results using matplotlib:

with DigitalIn(lines=[0, 1], rate=50*kHz) as di:
    data = di.read(200*ms)

plt.plot(data)
plt.xlabel('Sample #')
plt.ylabel('Binary value')

(The second line sampled a rather boring signal.)

Retrieving raw binary data

On occasion, e.g., for debugging purposes, it may be useful to retrieve raw binary data.

import numpy as np

with DigitalIn(lines=[0, 1], rate=50*kHz) as di:
    data = di.read(200*ms, raw=True)

bits = np.unpackbits(data, bitorder='little').reshape(-1, 8).T
plt.imshow(bits[:,:200], aspect='auto', interpolation='none')
plt.xlabel('Byte #')
plt.ylabel('Bit #')

The result is a packed array of bytes (np.uint8) in which the bits represent the interleaved data from all of the recorded lines. In this case, bits 0, 2, 4, and 6 of each byte derive (in that order) from line “di0” whereas bits 1, 3, 5, and 7 derive from line “di1”. Plotting that with numpy and matplotlib involves a little np.unpackbits magic, as the above code illustrates.

Of course you could extract the two lines separately using np.unpackbits(data, bitorder='little').reshape(-1, 2).T, though at that point you might as well call di.read(...) without the raw parameter set to True.