I'm going to guess that you are largely "self-taught" about LabVIEW, and have not taken full advantage of the "LabVIEW Training Resources" listed on the first page of this Forum.
I'm also going to guess that you've not had any experience with "Signal Analysis", including acquiring signals by modern Data Acquisition techniques, including sampling (and how that works).
LabVIEW is a language designed by Engineers (and largely targetted for Engineers) to build "Virtual Instruments" for testing (and sometimes replacing) hardware designs. Indeed, the name comes from Laboratory Virtual Instrument Engineering Workbench (hence the funny style of capitalization).
LabVIEW includes a particularly robust (and fairly User-Friendly, once you "Learn 10 Functions in NI DAQmx" and handle 80% of your Data Acquisition Applications -- I encourage you to search that phrase on the Web) and can easily handle the tasks you are trying to do, but you need to learn a little more in order to do it properly.
LabVIEW is a Data Flow language, which makes it behave a little differently (as far as timing goes) from other computer languages you may have used. You've seen, already, that there are things such as Controls (or "inputs"), Indicators (or "outputs"), Functions, and Structures (Loops, Case Structures, etc.) and something called "Wires" that connect them and carry Data. There are three important "Rules" for Data Flow:
- No Structure or Function will start running until every Wire coming into it has Data.
- No Structure or Function will put Data on its "output" Wires until the Structure or Function exits.
- No Structure or Function can exit until everything "inside it" has run.
LabVIEW also differs from other computing languages in elevating "Time" to have an elevated importance. This is especially important when dealing with Data Acquisition and DAQmx.
Suppose you want to acquire data and save it for later analysis. Let's say you want to acquire the data at 1 kHz, and acquire it for 10 minutes. That means taking 1000 x 60 x 10 = 600 k points. Should you take them one-at-a-time? That's a lot of tiny operations (almost a million of them!). Better that you take them, say, 1000 at a time, so every second, you get 1000 points. No problem writing 1000 points to disk in a second, right?
Most Data Acquisition (or DAQ) hardware that works with LabVIEW can take single samples, or multiple samples based on an internal (and generally quite accurate) "Hardware Clock". So you can ask the Hardware to acquire 1000 points at 1 kHz and do it "over and over again until I say Stop" (or until you do it 600 times, if you want 10 minutes worth). Suppose you put such a DAQmx Read function inside a For Loop (with 600 wired to N). How fast will the Loop run?
The Third Law of Data Flow says that if a DAQ Read function inside the Loop is acquiring 1000 samples at 1 kHz, that operation will take exactly 1 second, so the Loop (without you having to do any other timing!) will run at exactly 1 Hz (or once per second). Now all you have to do is write the data to disk!
But that's also simple. Before entering the Loop, open a Disk File for writing, and bring the Wires (representing the Error Line and the File Reference) into the For Loop. Take the output from the DAQmx Read holding the 1000 points of data and wire in to a File Write function (which you assume takes less than 1 second). Bring the File Reference wire to the right edge of the For Loop, and put the "Close File" function outside the Loop to close the file.
Now run the program. You enter the For Loop at t = 0 (say). The DAQmx Read starts to run, but the File Write function just sits there because there's no "Data" from the DAQmx Read to the File Write present yet. At t = 1 (second), the DAQmx Read finishes and passes the data to the File Write. [Oops -- I forgot to say that you configure the DAQmx Read to "sample continuously", meaning as soon as it finishes the first 1000-point sample, it continues sampling without a break to get the next 1000 points]. Assume the File Write takes 0.7 seconds. So at 1.7 seconds, the For Loop finishes Loop 0, and starts Loop 1.
Now we're in Loop 1, with 70% of the 1000 points already collected. So 0.3 seconds later, the DAQmx finishes, passes the data to the File Write, and at 1.7 + 0.3 + 0.7 = 2.7 seconds later, Loop 1 finishes, and on to Loop 2. We continue the DAQmx Read/File Write sequence until we reach Loop 599, after which the For Loop exits, passes the File Reference to Close File, and you're done.
Give it a whirl.
Bob Schor
