Vibration shaker controller

[PatMa]

I am looking for a simple Vibration shaker controller working in closed loop, for experimental testing.

I am looking for a simple Vibration shaker controller working in closed loop, for experimental testing.

The controller outputs a signal to the shaker's amplifier.
An accelerometer (connected to the controller) is fixed on the shaker's plate or on the load.

The controller generates a noise for the shaker and adapts dynamically that noise until it corresponds to the operator's input.

I heard of expansive controllers for satellite vibrations measurements, but I am looking for a simple system with only one or two inputs.

Any suggestion?

 

[dgallagher]

http://www.reliabilitydirect.com/shakerproducts/CMCPSHAKER.htm

 

[PatMa]

Dgallagher, the link you sent does not point to a controller but to a calibrator.
That -added to the fact that there is only one answer to my post- makes me think that I did not clearly stated my question. So, lets reformulate it:

Lets connect a white noise source to a shaker -open loop-:
- if the shaker was ideal the acceleration spectra would be also a white noise
- in real, the shaker modifies (as a filter) the signal and the acceleration spectra what we obtain is not flat anymore.

What I am looking for is an active controller that -in closed loop, sensing the acceleration of the load- generates a noise such as the acceleration spectra corresponds precisely to the desired one. With it, an acceleration spectra flat as a white noise is obtainable.

Any hint for such a controller?

Is someone facing of the same problem?

I am very interested in your feedback!

 

[MikeyP]

G+W

http://www.gearing-watson.com/siras.htm

LMS

http://www.lmsintl.com/index.asp?http://www.lmsintl.com/solutions_home/niveau.asp?category=301

Ling

http://www.lds-group.com/products/005.asp

A search on google for "vibration contoller 4 channel" yielded the following from just the first 10 hits

http://www.derritron.com/dvc-4.htm

http://www.vibrationworld.com/

http://www.mradcorp.com/prod_instrumentation.html

http://www.dataphysics.com/B&W Spec Sheets/

350_specs.pdf

M

 

[GregLocock]

And be very very careful! You may get what you wished for, which often leads to tears before bedtime.

Make very sure that your shaker controller does not overload the shaker, and make sure that your shaker doesn't overload the structure.

If you are happy to use swept sine then the wholly wonderful HP 3562 dual channel FFT analyser will do the job very nicely. It will even sample at a frequency until certain conditions are met (eg coherence).

The way we do this for random inputs (simulating road inputs to a car, starting with forces and correlation channels taped) is to

(a) run the taped force signal into the rig

(b) measure the vibration response of the car

(c) compare (b) with the real vibration repsonse of the car. Scatch head.

(d) mangle the taped force signals into a new drive file (secret engineer's business)

(e) go to (a)

Eventually it is as close as we're going to get and we run the thing.

There is a fundamental control issue here - the bandwidth of your shaker setup has to be a great deal higher than that of the system you are exciting (so that it can apply feedback in time), yet the structure itself won't respond especially quickly.

That is why we use an open loop controller. Cheers

Greg Locock

 

[drivingmissdaisy]

Try m+p international at

http://www.mpiuk.co.uk

and look into their vibration control products.

 

[Smasher]

We use data physics controllers, they start off as simple swept sine but can be upgraded to sine on random etc as the need arrises. They build systems with multiple channels and are quite comprehensive. I would start at a DP550 as it provides the best starting point for upgrades etc.

I've had M+P in to demo their system and it was blown out of the water, it was based on the HP AGilent front end. As for the HP3562, we've got 3 of them and even though thay are good thay are beat by a Data Physics 550 when configured correctly.

 

[blackeye]

Try m+p. What country are you in?

 

[drivingmissdaisy]

Have to agree with Blackeye,
Try m+p, they are the current leaders, they can use different types of hardware so the entry level systems are cheaper but still with high spec. Well ahead of Data Physics these days, we used to use them but have moved on to m+p.