I am trying to build a vibration simulator. The end user is a consumer NOT NASA, so I can' use exotic actuators that cost between $500 and $4k a pop.

I am thinking of using simple solenoids.
http://www.mcmaster.com/#electric-actuators/=7r552n
I will start with two (two degrees of freedom, up/down and left/right). My initial thought (please suggest better) is to mount the "platform" on springs thus holding the dead weight (<50lbs) and allowing for motion up/down. The motions to be reproduced are not repeating sine waves, but more like a bumpy car ride with variation.

If the springs can flex horizontally an inch or so then I would be able to combine the two movements, if not then, the whole spring support would be slid side to side. Either way the problem is controlling the solenoids.

DISCLAIMER: I do not pretend to know what I am talking about from here on!

I have seen some interesting "linear solenoids"
http://www.ledex.com/linear-solenoids/selection-soft-shift.html
And am wondering how I could control them.

In reading this forum I came across several threads discussing the National Instruments products
http://zone.ni.com/devzone/cda/tut/p/id/2991
But I am unclear what current they can handle or how the signals could be used to control devices that can handle larger current.

Then there is the STAMP
http://www.parallax.com/tabid/295/Default.aspx

As a seasoned developer, I can write code in C/C++/BASIC for all the control requirements, but how to reproduce the signals at the platform is a little beyond my knowledge at present.

All suggestions welcome

 

I had to build a test platform for endurance years ago. I used a scaled up version of a cellphone or pager vibrator. (offset weight on a motor) And standard linear solenoids. Basically the same thing that pinball machines use to fire the ball (The newer machines not the pull-stick versions )

They are quite easy to use. You send a 24v pulse to the solenoid and thats it.
You can use darlington pairs or mosfets to get the juice. The amount of weight you need to move will determine the currents involved.

If you remember the dancing toaster in Ghostbusters 2, it used linear solenoids for the dancing feet.

 

If it were just one axis, I made a vibration rig of my own for testing car components at work (quality and development dept).

Just a big subwoofer speaker, with a machined plate attached to it, with a housing and connection box.

I made a transistor amplifier, which is driven off of a programmable function generator.

simple enough really... something to give you some ideas.

 

I used to work for a major air conditioner manufacturer, they used a "shaker table" to simulate a cross country truck ride to see if any panels would vibrate loose.
It was an actual flat bed truck body(20' from a wreck, cheaper than building one) mounted on 6 captive coil springs.
Fastened to the bottom of the bed was an electric motor with offset weights mounted to the shaft (like wrechted said). The bed was free to vibrate up and down as well as side to side. The motor was on a timer and the weights were variable.
You'll need a motor with good bearings, not just sleeve bushings, we found that out the hard way.

 

You've gotten suggestions ranging from adapted loudspeakers to 20" truck beds but you never told us what the payload is or what vibration you are trying to simulate. Makes an enormous difference in what to suggest. Weight? Test object configuration? Single frequency, swept sine wave, random? Frequency range? Have you asked your customer what he hopes to achieve? It doesn't do anybody any good to run a test that is unrelated to whatever the DUT will see in the real world.

Remember, there is a reason commercial vibration test equipment is expensive.

Eccentric weights on rotating shafts are very practical if you only need quasi-sinusoidal excitation. It's a common method of generating vibration at relatively low cost but is an extremely poor simulation of real-life transportation environments. Also, for a given weight configuration, you adjust excitation level by varying motor speed so you don't have independent control of frequency and excitation level unless you stop the machine and change weight or eccentricity. You also get circular, not linear motion, which may or may not satisfy whatever test criteria you have. You can get linear axcitation with two identical eccentric weights on two shafts geared together with spur gears, one driven by a motor. Depending upon the energy absorption of your payload and where you may be on any resonances of the moving system, it may take more motor power than you would imagine to drive the system.

Woofers are versatile vibration drivers butwill self-destruct if overpowered or if eccentric loads are applied to the voice coil (not always easy to avoid). Commercial test shakers use leaf flexure springs or sometimes roller guides at both ends of the voice coil structure to resist side loads.

Can you tell us more about your test object and criteria?

awright

 

Continuing:

I'm not sure which McMaster-Carr actuators you are thinking of, but the motorized ones will only give you extremely low frequencies and the "slide actuators" appear expensive enough that real used vibration shakers would be competitive.

In the early days of large disk drive technology you could find in surplus (also known as junk) stores large magnet assemblies with mechanically guided "voice coils" perhaps 3 or 4 inches in diameter with several inch strokes. They positioned large "combs" of disk reading heads along the radius of the 12 inch to 18 inch hard disk stacks. I haven't seen any for a while, but ask around. Probably cost more to ship than to buy.

Can you tell us more about your test object and criteria?

 

Can be done with a motor firmly attached to the table suspended on springs.
The shaft of the motor with an excentric weight on its shaft.
Depending on the motor, a plain dimmer will adjust the frequency.
Thay is exactly what I had on an electronics service center. Cannot remember brand. Had nylon straps to secure the item under test.
For a more elaborated shake table, a second motor and dimmer mounted at 90 degrees from the above will yield complex vibrations.
Hope it helps.

 

Some great ideas.
bluebrakes
I LOVE the sub woofer idea. I need to try to find what force could realistically be generated by one. Could I get some more specifics on how I could drive this from a compute?

>Eccentric weights on rotating shafts are very practical if you only need >quasi-sinusoidal excitation. It's a common method of generating vibration at >relatively low cost but is an extremely poor simulation of real-life >transportation environments.

Yes exactly. This has to be a tad more sophisticated than a simple repeated motion.

The
as mentioned
>Weight?
<50Lbs (less than 50 pounds)
>Test object configuration?
sprung platform
>Single frequency, swept sine wave, random?
A computer generated signal that is neither repeating nor regular.
>Frequency range?
Hard to know at this point as it depends on a subjective factor "how it feels"
>Have you asked your customer what he hopes to achieve?
Simulation of simple movements in two planes like driving down a bumpy road

I have some responses from actuator manfrs and it may be possible to pick a model up for $200ea. I will have to see what the Mean time to failure might be with a mechanical device like this.

Added:
They are sending me a sample actuator next week. They said I will need to build an H-Bridge controller. What is that? Can I control it digitally from the computer?