06-17-2021 10:33 AM - edited 06-17-2021 10:37 AM
Hi All,
I have done a bunch of searching on this topic but can't seem to find an example of how to iterate through a data file to "play back" a orbital/time based plot.
Currently I have a working VI based off of the Sound on Vibration Toolkit example "Demonstration of Analysis VIs (Simulated)" In this example, a state machine is used to read the data file, pass it to a shift register which then passes it to the "processing" state where it is chunked up into user-defined blocks using "svx_Get Wav subset.vi" However, since I need even-angle sampled data to properly display the data on an orbital plot, I am using "OAT Even Angle Subset.vi" to chunk up the data for plot display (more explanation below).
Here's the problem: each loop is painfully slow. Practically speaking, I was hoping that I could scan through my data files in faster-than-real-time to identify points of interest for later scrutiny. I'm able to do this with a parallel vi I created for looking at the FFT of the same data. Maybe this is just an impossibility with resampling to even-angle signals.... At 27,000 rpm it should take ~0.22 seconds to amass 100 rpms but the loop iteration time is around 5s. Seems like a lot of overhead. I tried chunking the data using "svx_Get Wav subset.vi" but for some reason the phase between the tach, X & Y data gets skewed.
I realize that not a lot of people on the board have the S&V toolkit. Here's the abbreviated code just for the data processing state. Maybe someone could take a look and suggest a better way to maintain the phase relationship but chunk more efficiently? Also, I realize that branching the even angle signals for "Shaft Speed" and "Time per subset" is inefficient, but removing these features only speeds things up by 1 second.
Thanks in advance!
Josh
Solved! Go to Solution.
07-02-2021 09:28 AM
For anyone out there battling with the same obscure issues:
The initial problem seemed to stem from using the 'svx_Get Wav Subset.vi' which is an example sub.vi included in the S&V toolkit. I had been using it based on another S&V example mentioned above along with its added functionality. I'm still a bit unclear exactly what about this .vi seems to botch the even-angle resampling but the solution was to swap this .vi out with 'WDT Get Waveform Subset DBL' building the needed functionality directly into my parent .vi. In doing so, I avoided using the very slow to process "OAT Even Angle Subset.vi"
The iteration time went down to 1/10 of what it was for a given block size.