Continue to Site

Eng-Tips is the largest engineering community on the Internet

Intelligent Work Forums for Engineering Professionals

  • Congratulations waross on being selected by the Eng-Tips community for having the most helpful posts in the forums last week. Way to Go!

Isolation of Random Vibration

Status
Not open for further replies.

spongebob007

Military
Sep 14, 2007
265
Let me start by saying I have zero practical experience with random vibration and only a very fuzzy knowledge of the theory. I have successfully designed isolated systems to attenuate shock and harmonic vibratory loads.

What I am dealing with is one of our current products that consists of commercial electronics that are mounted using vibration mounts. The assembly has successfully passed vibration testing so the isolated system has proven itself acceptable for harmonic excitation. Recently we have sold this equipment to a new customer and they require it to pass random vibration. The equipment failed. I examined the equipment and determined that the most likely cause of failure was the vibration mounts bottomed durning testing and generated enough force to cause damage to the isolated equipment.

For harmonic motion, design of an isolated system is as simple as selecting isolators such that the driving frequency is greater than 1.4 times the isolated system frequency (2X or more ideally) and that there is sufficient excursion space for the isolated system to move under load. The isolated structure also needs to be stiff to avoid unwanted amplifications. But how does one go about selecting an isolator stiffness when frequency content between 1Hz and 100Hz is going to be hitting the sturcutre essentially all at the same time?

One of the lessons learned from harmonic motion is that sometimes a poorly designed isolated system is worse than no isolation at all. Along those lines, I was also considering just removing the isolators altogether and trying to harden the equipment to withstand the full force of the vibration.
 
Replies continue below

Recommended for you

Don't give up so quickly. You can design a random time signal to meet the spectrum, turn the transfer function of your isolation system into a time based filter cascade, then run many different frames of the time signal through the filter set and study the resulting deflections.

That sounds complicated, each step is simple, and you will learn a lot about signal processing.

If anything random is easier than sine based stuff. The only gotcha, and it is enormous, is the crest factor of the time signal (ratio of peak to RMS), we often use 3 or 5. There are various ways of achieving that, some ugly, some not. I only know ugly.

If you have a persistent problem with overtravel fit snubbers, we do it all the time.

I'm kind of hoping you know matlab or mathcad, not just excel...

Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376
 
Status
Not open for further replies.

Part and Inventory Search

Sponsor