Split a random vibration spectrum by 2 - HOW ?

[julienper1]

Hi all,
Im new on this forum and I hope you will be able to help me
I have an equipment to shake from 5Hz to 150Hz and my shaker is too smooth for very low frequencies (from 5Hz to 10Hz)=> overtravel. I would like to use another test equipment (like jack) in order to test from 5Hz to 10Hz and my shaker from 10Hz to 150Hz.
Is it possible to split my vibration spectrum by 2 (from 5 to 10Hz on a specific test equipment and from 10 to 150Hz on my existing shaker) and keep the same energy or fatigue on the equipment tested ? Does a specific law exist ?
I can precise if it s not clear to you (Im French and my english is sometimes "poor" compared to you )
Hope you will have some ideas for me !!
Julien

 

[SomptingGuy]

If you are measuring some kind of transfer function and both tests are linear, you should be able to stitch the two spectra together.

 

[julienper1]

What I want to do is spliting a spectra not stitching
I m looking for the theory which is introducing such calculation... what is uneasy !

 

[SomptingGuy]

Stitching is what you do to the bits you've had to split up when you want to reconstruct the whole spectrum.

My point is that the amplitude and phase of a linear transfer function can be determined point by point (i.e. frequency by frequency) or all together. You are choosing to do it in two measurements.

 

[julienper1]

I m not sure to understand properly what you mean, in fact, what I have imagined to do is to use the beginning of the spectrum on a test equipment during t1 and the end of the spectrum on another equipment during t2.
For example, if the PSD is 0.1g²/Hz during 1hour from 5Hz to 150Hz, what would be the duration of excitation t1 (with I assume 0,1g²/Hz) and the duration of excitation t2 with I assume 0,1g²/Hz).
I dont think that it is linear and using transfer functions seems to be a little complicated because I m not sure that I can determinate it (What i have never done before)
Thanks for your support whatever !!

 

[GregLocock]

From a practical point of view you won't get the same interactions between different frequencies if you split the test like that, and you are using random excitation. You may find a rule allowing you to do so, but I wouldn't trust it.

If you are using swept sine then yes you can, but I get the impression you are using random.





Cheers

Greg Locock

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[julienper1]

Im afraid you are right, I dont find any laws or else to split random spectra.
If someone have any information please advice, perhaps we will find on this forum a vibration genious !! (hope so )
Whatever thanks to all, sure I ll be back to help or to ask for help in future !!
See U
Julien