Resonence probem in machine

[steve30]

I am experiencing a system instability problem due to a resonence. The system is higly complex consisted of drives, voice colis, piezos,
springs and compex mechanics. It performes precision movement in 6 direction in 100 mm range (motion resolution is in nanometers).

Some times (once in a day) a present system resonance band around 300 Hz (bandwidth aroung 20 Hz) is excited to its 4-30 times
amplitude (upto few µm) due to resonance. It causes the system to be unstable for 10-20 seconds.

How I can solve this problem?
How I can find the source of resonence freuency?
Thanks

 

[izax1]

Well, it seems like you already know some of the main parameters. How do you find it is 300 Hz.? How do you know the amplification of 4-30 times? What kind of measurements and instrumentations do you have? Depending of what instruments you have, you kan perhaps use them to get deeper into the root cause.

 

[steve30]

Interferometrs fixed on the systeme are used to measure and calculate a FFT. But I can not pick up the resposible element of the machine out of the FFT.

 

[IRstuff]

Sounds like you need to start taking it apart. Split it in two and run a sine sweep on each half. Hopefuly, it's at the split or only on one half. Then keep splitting and sine sweeping. If you can't or won't run a sine sweep, you could possibly do an impact test:

http://www.reliableplant.com/Read/22455/identify-resonance-condition-vibration

TTFN

FAQ731-376
Chinese prisoner wins Nobel Peace Prize

 

[BrianE22]

A strobe light firing at the offending frequency can often times show you what's vibrating and even the mode shape.

 

[izax1]

You can usually work wonders with a proper FE Model and analysis of the system. The Dynamic analysis will give you frequencies, mode shapes and areas of "weak links". Will require some work and the relevant competent engineer. But it might be worth it.

 

[Strong]

Brian,

How can you use a strobe to see motion at 300 Hz in the micrometer range?

Walt

 

[zzzjim]

Just a few fault finding tips ,
I have a few suggestions based on Assumption that a
closed loop positional feedback system is present.

Some form of wear and tear , upsetting feedback systems ?
a favourite action , position that starts process off ?
external signals getting into feedback systems ?
Verify that external vibration is not an issue Chicken/Egg !

Experience in this field -None ,But Love a interesting Fault !

Good Luck