Is it possible to compare Random response data with Frequency Response

[saintsaens]

i have two sets of data: 1) Random PSD g^2/hz data from a test we ran and 2) frequency response data from the manufacturer (i.e. acceleration versus frequeny). Is it possible to compare these two to see if over testing was done?

 

[BrianE22]

frequency response data...

would normally be output/input vs. frequency. Are you sure about what you have in 2) above?

 

[IRstuff]

You should just ask the manufacturer what levels he used and what sweep parameters he used. Otherwise, pure frequency response as alluded to by BrianE22 is normalized, and there is no way to deduce the stimulus level.

TTFN

FAQ731-376
Chinese prisoner wins Nobel Peace Prize

 

[saintsaens]

I mean accel versus frequency. I obtained this data by taking FFT of time domain in-situ manufacturer data from a test where he mounted accels to hard points on the chassis of vehicle.

So now i have two curves. 1)freq./g**2 from random test we ran. 2) freq/g curve from vendor.

Now to compare the two do I just look at what happens at resonance of the part mounted to the chassis (which i make) and use Miles equation? Then see which gives the higher loading?

I am sure there I cannot just make a head to head comparison but only look at resonance to see if there was an overtest situation.

 

[IRstuff]

Maybe, you should be comparing the total g-rms vs. the bandwidth-integrated stimulus in the sine sweep as a function of time.

TTFN

FAQ731-376
Chinese prisoner wins Nobel Peace Prize