DIAdem

cancel
Showing results for 
Search instead for 
Did you mean: 

Diadem spectrogram help

Hello,

 

I've been struggling to create a spectrogram or waterfall plot of RPM vs. Frequency vs. Amplitude.

The data is from a synchronized microphone and Tachometer. Both of these are waveform channels.

I can generated many time interval FFTs, but can't get them onto a graph.

 

This old post has a broken link to download a tool for it unfortunately:

https://knowledge.ni.com/KnowledgeArticleDetails?id=kA03q000000x14vCAA&l=en-US

 

What am I missing? All help is appreciated, thank you.

0 Kudos
Message 1 of 5
(1,774 Views)

Got it!

The only issue now is getting the FFT index to corelate to RPM accurately. 

Marco893_0-1634835625682.png

 

0 Kudos
Message 2 of 5
(1,751 Views)

Hi Marco,

 

If you perform a time interval FFT calculation, then you will get an XYZ graph with Time, Frequency and Amplitude axes.  In order to get an XYZ graph with RPM, Frequency and Amplitude, you need to instead perform an order analysis calculation.  The DIAdem order analysis calculation takes as inputs a Time channel, an RPM channel, and a Signal channel, so you will first need to convert your tachometer channel to an RPM channel to use as an input to the order analysis calculation.

 

It may expedite you steps to use a dialog-driven script application designed to perform these steps and automatically pop up resulting graphs, so you might want to give the "Joint Time Frequency and Order Analysis Tool for DIAdem" a try:

https://knowledge.ni.com/KnowledgeArticleDetails?id=kA03q000000x14vCAA&l=en-US

 

Brad Turpin

Principal Technical Support Engineer

NI

0 Kudos
Message 3 of 5
(1,731 Views)

Thank you Brad,

 

I think I'm starting to understand this all better.

The link to the tool/script from your link is unfortunately broken, be cause that does appear to be exactly the tool for the job.

 

When I use the "Order analysis, frequency domain" though, I get several repeat FFTs, and strange scaling.

The best I have found is using the raw signal file. RMS and or filtering seems to give it hiccups.

 

There is some better correlation (single speed electric motor revving) when plotted on an overhead 2D, but a lot of the dB values, and choppyness of the FFTs makes me question that It's processing correctly.

 

Marco893_1-1634921753256.png

 

0 Kudos
Message 4 of 5
(1,721 Views)

Hi Marco,

 

My word, the download link is broken!  I'll look into that, but in the meantime here is the latest version of the Joint Time Frequency Order Analysis script/dialog application.  Let me know if you have any trouble with it.

 

Brad Turpin

Principal Technical Support Engineer

NI

0 Kudos
Message 5 of 5
(1,705 Views)