Source code for awg_creation_demo

#!/usr/bin/env python

This script will generate a binary file that can be used to output a waveform on D-TACQ AO modules.

To have the UUT output a single sample the file should have 32 int16s ie (CH1 CH2 ... CH32) where the value for each
channel can be different (they will not be different for this example). The user can extend this to as many samples as
necessary and to any value necessary i.e the values do not need to be the same across channels (channel 1 could be a
sine wave, channel 2 could be a sinc wave etc).

In this script we are outputting the same identical sine wave across all 32 channels for sine_size number of samples.
There is also an argument called even_ch_to_zeros that, if set, will set all of the even channels to zero to illustrate
how the channels are ordered.

Example usage:

To create the default file:


To create binary file with even channels set to 0 and a wave size of 3000 samples.

python --sine_size=3000 --even_ch_to_zeros=1

To upload awg files once it has been generated use the following python script:

python --files="../acq400/waves/example_awg" --capture=1 --awglen=<awg_length> <UUT name>

import numpy as np
import matplotlib.pyplot as plt
import argparse

[docs]def create_array(sine_size, args): # function that creates an array with sine_size * nchan elements # Firstly, an array containing a sine wave is generated. This is entirely arbitrary and is meant to demonstrate a # waveform being generated. We scale the sine wave by 32767 because the DACs are 16 bit. # DAC codes valid in range -32768 to 32767. If the UUT gets a code of 32768 the output will just saturate. Also # be aware that the astype(np.int16) function used below will wrap back to -32768 if a number larger than 32767 is # used. The same goes for numbers < -32768: they wrap up to 32767. sine_wave = np.sin(np.linspace(0, 2*np.pi, sine_size)) # Creates sine wave sine_wave = 32767 * sine_wave # Scales the sine wave to 16 bit sine_wave = np.rint(sine_wave) # round the sine wave to integers # An empty list is created that will eventually contain our waveform waveform = [] # For every point in the sine wave: duplicate it nchan times (one value for each one of the channels) for num in sine_wave: waveform.extend(args.nchan*[num]) # extends the single value taken from the sine wave data across 32 channels. if args.even_ch_to_zeros == 1: # if this argument is chosen then all of the even channels will be set to 0. for index, element in enumerate(waveform): # loop over all of the channels in order if index % 2 != 0: # if the index of that channel is odd then set it to 0. # We check for odd channels here because in python we index from 0, whereas real # channels are indexed from 1. waveform[index] = 0 # Here we set the value of the "even" channel to 0. # change waveform into a numpy array so we can use numpy functions on it. waveform = np.array(waveform) # We then ensure that the waveforms points are of type int16. This is important as the UUT requires the data type # to be int16. waveform = waveform.astype(np.int16) return waveform
[docs]def write_array_to_disk(waveform, dir): # function that writes numpy array to disk. # This is just a binary file. waveform.tofile(dir, "") return None
[docs]def generate_awg(args): #function that generates an awg waveform waveform = create_array(args.sine_size, args) write_array_to_disk(waveform, args.dir)
[docs]def run_main(): parser = argparse.ArgumentParser(description='generate awg waveform') parser.add_argument('--sine_size', default=30000, type=int, help="Size of sine wave to generate.") parser.add_argument('--nchan', default=32, type=int, help="Number of channels in AO module.") parser.add_argument('--even_ch_to_zeros', default=0, type=int, help="Whether to set even channels to zero.") parser.add_argument('--dir', default="waves/example_awg", type=str, help="Location to save files") generate_awg(parser.parse_args())
if __name__ == '__main__': run_main()