Multifunction DAQ

cancel
Showing results for 
Search instead for 
Did you mean: 

USB-6361 Background AI record while using AO

Solved!
Go to solution

Hello everyone,

 

I'm new using nidaqmx python API with a USB-6361 DAQ and I face an issue while trying to use an analog input and an analog output simultaneously. My setup is a a simple control of an actuator and a sensor.

The idea is to start the reading task to record the output of my sensor while I emit a signal with the acutator. The timings look like this (input and output could be started simultaneously):

 

timing.png

 

Here is my current minimum working example :

 

import numpy as np
import nidaqmx
from nidaqmx import stream_readers
from nidaqmx import stream_writers

out_chan = 'Dev1/ao0'
in_chan = 'Dev1/ai0'

fs = 1 * 10 ** 6         # Sampling at 1MHz
f0 = 100 * 10 ** 3       # Signal at 100kHz
nb_pts = (fs // f0) // 2
signal = np.append(np.ones(nb_pts), 0)
result = np.zeros(fs)    # Recording during 1s

with nidaqmx.Task() as out_task, nidaqmx.Task() as in_task:
    out_task.ao_channels.add_ao_voltage_chan(out_chan)
    out_task.timing.cfg_samp_clk_timing(
        rate=fs,
        sample_mode=nidaqmx.constants.AcquisitionType.FINITE,
        samps_per_chan=len(signal))
    writer = stream_writers.AnalogSingleChannelWriter(out_task.out_stream, auto_start=True)

    in_task.ai_channels.add_ai_voltage_chan(in_chan)
    in_task.timing.cfg_samp_clk_timing(
        rate=fs,
        sample_mode=nidaqmx.constants.AcquisitionType.FINITE,
        samps_per_chan=len(result))
    reader = stream_readers.AnalogSingleChannelReader(in_task.in_stream)

    reader.read_many_sample(result)
    writer.write_many_sample(signal)
    out_task.wait_until_done()    
    in_task.wait_until_done()

 

 

My issue is that when I start the input task, it blocks my script during the full reading duration. Then, the output task is played after that.

 

Playing with the output part of the code, I noticed that the writer stream is non-blocking because I can switch it to CONTINUOUS and let it run while doing other things. Thus I tried to start this output before recording the at these high frequencies, the reader misses some data at the beginning.

 

I know that triggers exist, but I did not understand how to use them, as the python documentation is quite scarce about them.

0 Kudos
Message 1 of 3
(2,581 Views)
Solution
Accepted by topic author thomas_w

I'm a LabVIEW guy and can't really help with detailed syntax for the python API.  Here are some general tips:

 

1. I would tend to avoid "auto-start" behavior.  You'll have more control over timing & sequencing when you choose to start your tasks explicitly.  (Note: the input task will likely auto-start when you try to read from it before you've explicitly started it.)

 

2.  After you've done most of the config, I would change things at the tail end of your posted code something like this:

    writer.write_many_sample(signal)

    // insert code to start the input task (but don't read yet!)
    // insert code to start the output task

    reader.read_many_sample(result)
    // the read call will wait to collect all samples so the
    // next "wait_until_done()" is likely redundant
    in_task.wait_until_done()

    out_task.wait_until_done()

 

3. There are also ways to sync tasks in hardware if you need to.  For now, this makes sure you start acquisition *just before* starting your signal generation.  They aren't sync'ed, but you also won't be missing stuff.

 

 

-Kevin P

CAUTION! New LabVIEW adopters -- it's too late for me, but you *can* save yourself. The new subscription policy for LabVIEW puts NI's hand in your wallet for the rest of your working life. Are you sure you're *that* dedicated to LabVIEW? (Summary of my reasons in this post, part of a voluminous thread of mostly complaints starting here).
Message 2 of 3
(2,553 Views)

Thank you very much Kevin!

 

Separating the in_task "start" and the reading command did the trick. For those interested, here the corrected code:

 

import numpy as np
import nidaqmx
from nidaqmx import stream_readers
from nidaqmx import stream_writers

out_chan = 'Dev1/ao0'
in_chan = 'Dev1/ai0'

fs = 1 * 10 ** 6         # Sampling at 1MHz
f0 = 100 * 10 ** 3       # Signal at 100kHz
nb_pts = (fs // f0) // 2
signal = np.append(np.ones(nb_pts), 0)
result = np.zeros(fs)    # Recording during 1s

with nidaqmx.Task() as out_task, nidaqmx.Task() as in_task:
    out_task.ao_channels.add_ao_voltage_chan(out_chan)
    out_task.timing.cfg_samp_clk_timing(
        rate=fs,
        sample_mode=nidaqmx.constants.AcquisitionType.FINITE,
        samps_per_chan=len(signal))
    writer = stream_writers.AnalogSingleChannelWriter(
        task_out_stream=out_task.out_stream,
        auto_start=False)
    writer.write_many_sample(signal)

    in_task.ai_channels.add_ai_voltage_chan(in_chan)
    in_task.timing.cfg_samp_clk_timing(
        rate=fs,
        sample_mode=nidaqmx.constants.AcquisitionType.FINITE,
        samps_per_chan=len(result))
    reader = stream_readers.AnalogSingleChannelReader(in_task.in_stream)

    in_task.start()
    out_task.start()
    reader.read_many_sample(result)

# Do something with <result>

 

 

I think it will be sufficient for now. If I need more precise synchronisation, I will dig the hardware triggers way.

0 Kudos
Message 3 of 3
(2,496 Views)