Natural frequencies vs real life? fixed-free tuning fork

[spacether]

Hello,
I'm trying to run a test case correlating Elmer's predicted natural frequencies of a part to real life.
I'm using a aluminum tuning fork that I bought off amazon.com.

---------------------------------------------
FREE-FREE STATE, Agrees well
---------------------------------------------
In free-free state (suspended by a thread), the first 9 real modes Elmer predicts are within 3% of real life values.

--------------------------------------------------------
FIXED-FREE STATE, Some modes don't agree
--------------------------------------------------------
My real life tuning fork fixed-free case (holding it in my hand and hitting the prongs on a table) was only showing two modes when I recorded the audio.
One mode is at 513 hz and is the tuning fork fundamental frequency.
The higher modes is at 3147 hz and is the second order sinusoid of the fundamental.
When I run a fixed-free case in Elmer, I am getting modes that are slightly lower and higher than the tuning fork fundamental.

See my attached video showing the elmer natural frequency animations, and my attached sif file.

tuning_fork_fixed_free_7modes.mpg

(398 KiB) Downloaded 379 times

case.sif

(1.61 KiB) Downloaded 331 times

--------------------------------------------------------
Questions:
--------------------------------------------------------
I'm having it report the smallest magnitude eigenvalues for now.
Should I be using largest magnitude, or largest real magnitude?
Do people know why I'm seeing these slightly higher and lower modes in Elmer but not in real life?
Could they be damping out?
How could I have Elmer show me that some modes will dissipate quickly, but other modes will vibrate for longer?

[raback]

Hi

Nice results you have there. Are you using linear or quadratic elements

I would think that the smallest magnitude is ok. Without damping the eigenvalues will always be real.

The fact that you have high modes does not imply that they would be observable in real life. Yes, the higher modes will die out more quickly. There is some damping model that might be perhaps used. Then the eigenvalue would be complex and the imaginary part would show the decay. And the higher modes will also be more difficult to wake. I would think that the tuning fork is designed to have a dominating lowest eigenmode.

I wonder whether the frequency depends somewhat also an air pressure?

-Peter

[spacether]

I used second order tetrahedra for the model (quadratic).
I don't think the air impacts these frequencies.
The fundamental mode for free-free was the tone the tuning fork was supposed to generate. (3% error real life vs analysis)
So I don't think the air impacts it much.