Transmissibility of Wire rope isolator

[m2fm2]

hello,

New to this great forum! So on to my question.

I have accelerometer data (time-history) for a wire rope isolator I am trying to characterize.

The setup was 2 accels mounted on either side of the isolator (input and output sides). To get a ratio of input to output, is it as simple as dividing the two time histories? I'm still new to working in the frequency domain, but could I FFT the two datasets and divide the two to get a "frequency domain" transmissibility? Or divide in the time domain first and then FFT (which doesn't sound right to me)?

thanks,
Ken

 

[GregLocock]

To get a ratio of input to output, is it as simple as dividing the two time histories?


No


I'm still new to working in the frequency domain, but could I FFT the two datasets and divide the two to get a "frequency domain" transmissibility?

No, but nearly there.

Or divide in the time domain first and then FFT (which doesn't sound right to me)?

No, that's worse.

Presumably you want an output as a transfer function, of transmissibility. This is easy enough, it is merely the ratio of the fft amplitudes, and the difference between the phases, at each frequency.Usually this is estimated by using the self product of the output spectrum divided by the cross produst, in the complex plane.

Note that transmissibility is very dependent on the end conditions, it only really works as a in-situ test in my opinion. (eg consider the case where you earth the output end- the transmissibility will be zero).



Cheers

Greg Locock


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

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

 

[m2fm2]

Greg,

Thanks for the reply! You have clued me in that I should be looking at it from a 'transfer function' perspective. Looking back on the digital data analysis course I attended last year...duh...the phase information needs to be preserved and be part of the calculation! Thanks for that as well.

cheers!
Ken