One-Third-Octaves for Dummies 1

[271828]

Hello all. I do a fair amount of work with structural vibrations, but our usual frequency band of interest is fairly narrow, like 0-20 Hz. No background in acoustics or higher frequency mechanical vibrations. Therefore, I've never had to deal with plots using one-third-octaves. I've done some reading online and have an idea of what's going on, but not quite enough to allow me to really deal with these effectively. My current application is checking a structure's ability to support a CT scanner. The manufacturer's tolerance plot is a spectrum with log scale rms acceleration vs One-Third-Octave Center Frequency in Hz. I'm not sure how to turn this into a spectrum with linear x and y axes for comparison with the predicted acceleration spectrum. If someone would be kind enough to point me to a good discussion, I'd appreciate it.

BTW, I searched, but only found the following, which was helpful, but I need more.

http://www.eng-tips.com/viewthread.cfm?qid=150400

 

[GregLocock]

The y axis is the rms of the enrgy in each third octave band, each of which has a bandwidth 22% of its centre frequency.

So the psd is a horizontal line, height (rms_accel_in band)^2/(.22*centre frequency of band in hz).

However you'd be better off turning your linear Hz measurement into third octaves for comparison.





Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[271828]

Thanks Greg.

A really silly question: Why would one want to use a one-third octave spectrum instead of a regular FFT magnitude? I work with relatively low frequency vibrations in the 3 Hz to 20 Hz range 99% of the time. A few of the equipment manufacturers express their vibration limits in terms of one-third octave spectra, hence my original question.

 

[electricpete]

I don’t work with anything like Octaves or 1/3 octaves. I can’t see why anyone would want to other than for historical reasons..... maybe it’s a holdover from the old days before getting an FFT was trivial?

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[Strong]

Here is some background reading for floor vibration criteria for sensitive instruments in third octave bands:

http://colingordon.com/pdf/Gordon-SPIE99.pdf

As Pete says, it probably evolved from a historical use of available instruments. Third octave filters are really quite narrow in the frequency range below 20 Hz!

Walt

 

[Tmoose]

In the 80s I happened to see a few reports from acoustic consultants generated after they had been asked to perform surveys in search of sound and vibration problems in bog office buildings. It was almost sad to see how many potential sources they had to include in their "analysis" when any of the lower "machinery" octaves was being bad.

 

[GregLocock]

I agree 1/3 octave specs are probably for historical reasons, rather than any real advantage. I suppose you could argue that the bandwidth of the ear is roughly 1/3 octave, but that is a very problematical statement.

Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?