Most of our projects depend on liberal application of vibration isolation systems on mechanical equipment. Especially in nanotech labs and other high-tech settings, you simply can't throw enough concrete and steel at the problem. It's far better -- and far cheaper -- to just minimize the vibrational forces that get applied to the structure in the first place.
But it bears repeating: resilient-support isolation systems can't eliminate vibrations. At best, they can only only reduce vibrations. Critically: the effectiveness of vibration isolators depends on frequency. In fact, they actually make matters worse if mis-applied.
This means that the exact same "isolator" that works great in one application might be worse-than-useless in another. That's right: your isolator can become an amplifier if you're not careful.
The plot below shows the force transmissibility of a simple spring vibration isolator system, something like the free-standing springs that you often see base-building machinery (like pumps, fans, and chillers) installed upon.
Isolation transmissibility curve for a steel spring with 1" static deflection, which works out to be a 3Hz isolator. Note that the x-axis is shown both in Hz and the equivalent RPM. Of course, this is somewhat simplified; real springs don't perform quite so beautifully at high frequencies.
Here's how to read the above plot: "transmissibility" is the ratio of output force to input force. Above, we've expressed this in decibels, but you could use decimal notation too. The "input" is the force produced by the machine, like the imbalance force that shows up at the shaft speed. This gets applied to the top of the spring system. The "output" is the force that shows up at the bottom of the springs and gets applied to the building structure.
Positive numbers (in decibels) mean amplification and negative numbers mean attenuation. Zero dB means that there's no change: the output force is the same as the output force. At the very lowest frequency (0Hz), nothing is moving, and the transmissibility tends to zero decibels. This should be obvious, since all of the static weight of the machine gets transmitted to the floor. At high frequencies, the transmissibility is negative: the output forces applied to the floor are lower than the input forces generated by the machine. This is the attenuation we were looking for. But between the two, around the spring resonance, we actually get amplification. Just how much amplification depends on the damping in the system. As a practical matter, isolators are only useful for frequencies well above the spring frequency times the square root of 2.
Now, there's a ton of ways that machine vibration isolation can go wrong, but I think it would help if we at least had a better word for these systems. When you hear "isolator" it's easy to forget that these systems just don't isolate at all frequencies.