LabVIEW

cancel
Showing results for 
Search instead for 
Did you mean: 

Thermocouples and Accelerometer VI

Dear All,

I want to record data using 3 thermocouples and 4 accelerometers.
Continuous samples = 2048, @ 2048 Hz.
The test should run for 5 minutes (300 s).

The data is logged in the .csv file after each (1/2048 s = 0.000488 s) 

Could you please confirm if my VI and the generated .csv file are correct?

Any suggestions to improve are most welcome. 

0 Kudos
Message 1 of 4
(785 Views)

You can Create Simulated NI-DAQmx Devices in NI MAX if you don't have actual hardware.

-------------------------------------------------------
Applications Engineer | TME Systems
0 Kudos
Message 2 of 4
(771 Views)

Hi,

 

A few suggestions :

- Arrange your Front panel and Bloc Diagram to fit in one screen

- Manage the errors

- Avoid duplicated code and make subVIs for reuse

- Minimise the use of Express VIs and learn how to use DAQmx API and Signals analysis palette

 

If your beginner at developping with LabVIEW I suggest you to following the NI training LabVIEW Core 1&2, find ressources on internet and youtube, reach Local LabVIEW User Group, etc. 

 

Loïc.

0 Kudos
Message 3 of 4
(768 Views)

Please read Learn 10 Functions in NI-DAQmx and handle 80 Percent of your Data Acquisition Applications.  Disregard the first point which talks about the Dreaded DAQ Assistant.  I urge you to rewrite your routine using (almost) no Express VIs, particularly no DAQ Express VIs, and no Dynamic Data Wires!  

 

I'm assuming (which could, I admit, be wrong) that you have two basically different devices, a Thermocouple (or 3), and an Accelerometer (or 4).  In my experience, these different types of data are acquired by different pieces of hardware, so you (ideally) have two DAQmx sets of functions.

 

The "Learn 10 Functions" White Paper talks about DAQmx "Tasks", which you can generate with MAX (make up a better name than "MyAnalogInTask" or, my preference, you can define them in your LabVIEW Project. 

 

Unless you are planning on doing "maximally efficient FFTs" on your signals, I see no point in specifying Samples to Read and Rate as a power of 2.  A sampling rate in "rounder numbers base 10" (such as 2000) makes more "engineering" sense to me.

 

You should be able to do most of your DAQmx stuff with 4 functions:  Create Channel (sets range), Timing, Read, and Stop.  Read goes in a While Loop.  Set your sampling Rate to 2000, Continuous, and Sample Size to 2000.  Note this gives you 2000 points once a second -- if you need to "see" them more often, change the Sample Size to 50 (which gives you 20 points at a time -- there's rarely a need to process every point in "real time").

 

You'll see you now have two "simple" streams of data that can be treated as an array of 3 Waveforms (Temp) and 4 Waveforms (Accel), which plot rather nicely (perhaps in two plots).  There are also simpler output schemes that will handle the Waveform data and avoid the Dynamic Data Wires.

 

Bob Schor

0 Kudos
Message 4 of 4
(762 Views)