I have a PXI system with:
Additionally I have a vibration shaker along with its power amplifier.
I want to perform a random vibration test using these devices. This test is demonstrated here.
So my questions are:
Solved! Go to Solution.
Not my field so pay close attention to other answers from actual experts. Just trying to offer a couple thoughts in the meantime, FWIW.
- Prior to your control-loop work, I'd think you'd want to first do some open-loop work to characterize a freq domain transfer function of your power amp and shaker. Then for a given desired accel, you'd know how to compensate the nominal control output so that the shaker produces the corresponding *actual* accel.
- I don't think a PSD converts unambiguously back into the time domain. Other forms of FFT can (such as complex with Re/Im parts). I'm not sure what the impact of this ambiguity will be on your efforts to construct a time domain signal from a PSD envelope.
- minimizing the difference between reference and measured PSD as a real-time control law sounds like a very advanced and specialized bit of work. It sure doesn't sound like something that could exist as a pre-packaged module. The fact you're hoping to find such an easy way out makes me suspect you may not yet have a sufficient understanding of the complexity of what you're trying to do.
Just to set the right expectations - a project like this is gonna take very significant time and effort.
I have done acoustic tests with Sound and Vibration toolkit, but your requirement does not seem to fit into any of the available toolkit features.
Typically SVT covers the measurement aspect of sound and vibration and some level of sound generation aspects but nothing on vibration generation aspects. I am not experienced enough in vibration domain to provide any valuable thoughts but looking at the closed feedback loop requirement, it does not look simple to me. I would agree with Kevin's analysis, PSD may not be converted into a time-domain signal, even if you do it, when generated using a shaker may not result in the same exact PSD.
Interesting requirement, will follow this thread for insights from other experts in the field.