continuation of "sine on random..." from 23 Apr 09

[rjkene]

vibrationnewbie and IRstuff both make valid points -- yes, there is a fundamental difference between sinusoidal and random vibration. However, you can still add the sinusoids to the random vibration in a time series and do a spectral analysis, getting both on the same graph. I've done this in Matlab -- the random component is easy to simulate using a method from R. Mitchell (IFFTs, overlap, etc) and I have no trouble getting the spectral levels for the random component. My question is for the sinusoids! MIL-STD-810-G specifies them in g^2/Hz. Figure 519.6C-1 places a 40 Hz sinusoid on the chart with a level of 22.5 dB re 1g^2/Hz. What amplitude sinusoid is this? I can fudge in a number and try it to see where it comes out on the PSD graph, but I'd like to know how MIL-STD-810-G wants us to calculate it!