All this talk of impact testing has me thinking... I have a variety of hammers, but don't have a shaker. Seems to me that a shaker, attached to a structure, where you could "sweep" the frequency range while monitoring vibration at multiple points (fan wheel, shaft, bearing housings, pedestal, foundation) would be the quickest way to determine what is going to shake the most, and at what frequency.

I saw Westinghouse once use a motor with an eccentric disk to do this on a turbine casing, but it looked sort of crude, and the "mounting" options would be pretty limiting.

I've never used an actual "shaker" but wouldn't that be the best way to go since they are fairly small, are stud mounted to the machine, and -- I assume -- can be powered with a "swept frequency" power source?

How well do these work? And how expensive is the entire setup? Is there a "poor man's" equivalent for any or all of the necessary components?

Thanks in advance.

If you purchase a metal cone 500 W base loudspeaker that can make a car look dizzy, you have the material for a good shaker. Now replace the cone with a diaphragm dual spring so the coil remains floating but the whole speaker is now just a large magnet system. Add some handles or eyes for hanging. Add a centre pin from the coil so you can input the force with a rare earth magnet foot. Feed the coil with a 200W car HiFi amplifier with a mains to 12VDC (actually 13.6VDC) power supply or bring the car battery and a simpler charger. Tune this using an old but stable simple tone generator from Ebay. The bill will sum up below 1000 USD plus a bit hobby work like boxes, rugged connectors and cables.
Ling Dynamics sells a good shaker range and you can sure find pro amplifiers and sweep generators or even a generator that, using a reference accelerometer, can sweep constant velocity or constant acceleration.

The W gadget you saw is available as shakers for dust like product handling. You can make one using one or two balls blown around in an outer ring of a ball bearing. Drawback is the force is prop. to speed square. Adding ref sensor gives you a nice system to record swept recordings of 1x vectors in Bode or Nyquist plots.

In the period before small FFT units, the above was all we had on hand. Reports were sure thinner but results just as good.

I have never used one, but it seems to me that it takes longer to setup and sweep a shaker test than to just do a bump. So if bump could do the same thing, that would be preferred. The advantage of a shaker as I understand it is large structures where you can't get enough energy in with a bump to excite the modes of interest (or perhaps the background vib is high and you want to increase s/n ratio)... shaker can put in a lot more energy at each frequency.

For big structures, such as a ship, the place I used to work used a mechanical shaker. It was a bit more advanced then the eccentric on a motor. It used a variable speed motor to turn a pair of counter-rotating arms, with up to 20 pounds of weight at the end of each arm. Pretty scarey to be around....especially if it broke loose from the deck during a test!

The advantage of counter-rotating arms is that the resulting vibration is directional - the sideways components cancel out. The downside, as Arne points out, is that the force varies as the square of the speed. When sweeping a wide range, it was often necessary to stop the test and change weights. Large weights were needed at slow speed to get enough force, less weight at high speed to prevent the shaker from flying apart.

In yet a different past life, we used a small electrodynamic shaker from Wilcoxon. It was roughly 10" in diameter and 12" from top to bottom. It was easy to use, mounting to a single tapped hole with a stud. It worked against a seismic mass so the single mounting point was all that was needed. I can't picture what the amplifier looked like...I think it was only about 50 watts so it wasn't huge. Add to the amplifier a signal generator of some sort to make it all work.

This system was relatively simple to install and use but does require some setup to use. Like a mechanical shaker, the output level is far from flat. In this case, the shaker has a resonance in the few Khz range.

Jon
Spintelligent Labs

Upside is that when you have tuned to a resonance and having the reference from your shaker being excenter, hydraulic, pneumatic or voice coil you can do a basic modal map in your head or for that entering the data in a software as you can read the phase and amplitude at any point you can access while shaking it. Then you know what to do to fix it, possibly. Olov

OLI, I think you and I are both "visual" thinkers... one thing I'd like to do with a shaker (that you can't do with a hammer), is to sweep the frequency range with the meter set up in monitor mode with "peak hold" turned on. Another method is to use a multi-channel analyzer to monitor multiple points while you sweep the frequency, to see how the different points react to the same input force.

Many of you must be asking, "Why?" Some folks think "creatively" as opposed to "analytically" and solve problems in perhaps different ways. You don't always have to take the methodical, crunch-the-numbers approach to solving vibration problems.

Rusty,

A shaker at a single location and acting in a single direction does not excite all modes in a structure or create uniform response of one mode to another. The aerospace and automobile folks use multiple input from more than one shaker. The dynamic mass of the shaker can influence the natural frequencies being measured. I have done some shaker testing with components on top of the shaker in a lab, used a shaker to excite a starter-generator on a jet engine while attached to a plane, and used a shaker connected to robotic laser-welder. The limited experience I have had means that I stay away from field test with shakers unless a compelling reason comes up. Like hammer impact testing, the use of a shaker or multiple shakers involves both art and science.
I am not knocking your's or Arne's creativity or practicality, but I am saying that for me 99% of the time, I would not consider a shaker for field testing of typical machinery structures.

Walt

Thanks, Walt.... "been there, done that" input is the best kind.

So I normally don´t want to know all of them, 6 degrees of freedom + shaft + fan axially + selected modes of all of those, yikes. No wonder we have work to do. I just like to have the behaviour of the one I need to kill to make the machine to behave and those I invite when doing the things I have to do to kill it. For that a shaker test is what you would like to do and on rare cases you are lucky and can enter the results in a proper FEM model and do all the fiddling in software not in iron and concrete until it seems to work in the model. It does happen that it also works out in reality if the FEM guy is gifted (experienced). These days you normally have to get by with a knock test and a hopefully gifted suggestion based on previous battles with similar beasts. Concrete makes most of them silent if it´s not in the rotating parts but some try that there also. All tools are qualified in the battle against the resonance beast. Good luck to all of you in the battle. May we all struggle for a resonance free world, normally (sometimes) feels like Don Quijote and the windmills but keeps us busy. Olov

Walt, the limitation of getting all modes to show when using one point and one direction is the same as for a hammer input at one point and one direction.

But for both cases, I usually place the point and the direction so that the chance is great to see most of them. A single point hammer modal test usually follow symmetry axis and do not show twisting modes well. but just a 45 degree off and there you go, all pop up nicely.

Advertising warning: There is a soft at some 100 USD where you tap in or ship in via excel all the vectors from you what-ever-method-test and you can see the machine moving instantly. Plus send that soft (all in all less than 500 kB) and the data to a customer so he can enjoy it directly on his screen.

Arne, email me.... did you ever get a non-terrorist email address ?

Arne,
My comment was not directed to an experienced person like yourself, but to the neophytes who must learn how to do a structural vibration test in order to solve many machine vibration problems.

Walt

Walt, don´t worry, I knew exactly what you mean. We have to put on a certain show here so the neos will read better and not fall asleep so fast. :-) I gave seminars with my father for years and we often woke up the sleepers by giving each other sore, unpolite or provocative comments or stupid questions. All woke up and listened to each word, in which time we brought the difficult knowledge over. On one occasion, a mechanic came up to me afterwards and said a bit angrily - You should not tease your father like that! We both had to explain to him that it was just a trainers trick, a show that we put on. That´s what you have friends for, right?

What I always wanted to do was use explosive bolts and a wire rope to apply step loading using a force transducer. I wonder why no one ever let me do this. Location, location, location – something about gas plants and explosions that scared people.

I did have some large structures to test.

I once met a consultant that claimed to have done the ultimate knock test, 30 sec burntime rocketmotor bolted to a bridge... I have only used the empty reel of a papermchine either w. 2 guys and ropes or vertically only using the crane. Nobody let me fire explosives either. Sigh. Olov

Bill,

I hope you don't have to FLY to job site with that set of TOOLS!

Walt

One of my tools might scare people!