Tri-ax accelerometer cross couple

[JordonMusser]

we recently ran a sine sweep and several random sweeps on a new vibration fixture to verify that we can pass through the various excitations frequencies of the fixture and still meet the customer req for deviation from expected input data.

We used a tri-ax accelerometer as our control, and the axis being tested passed all testing with flying colors. We did however notice at high freq, around 1khz (where we believe our first naturual freq of this fixture to be) the other two axis went from ~0Grms to levels similar to the control axis, even tho the control axis was spot on what it should be. The accelerometer was secured with wax near the product interface.

our customer is fine with this, as they have no requirements for cross coupling.

Should I be concerned? anything else to look at? i've never used a tri-ax like this so this was unexpected and want to explore further?

 

[BobM3]

If your input was vertical, could the resonance have been rolling or pitching the acceleromter?

 

[GregLocock]

I suggest you perform a modal analysis of the rig. As you suggest, if there is a rig resonance at 1 khz then it is not surprising that there would be deflections in all 3 directions.



Cheers

Greg Locock

SIGlease see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[JordonMusser]

I have done a modal analysis, and of course there is movement in all 3 directions.

However, since the axis in question is the control, the closed loop nature of the table handles this quite well and eliminate the increased amplitude/accelerations of the resonance.. so i was suprised to see some of it in the other axis. it sounds like this is just the way it is... and we should move on?

 

[GregLocock]

Either we are talking at cross purposes or you don't understand the implications of your modal- how can the controller be expected to control an axis it can't see?

Anyway, yes it probably doesn't matter.

Cheers

Greg Locock

SIGlease see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[BobM3]

I'm not surprised that a test fixture would exhibit resonances at 1000 hz (I'm surprised you don't see any at lower frequencies). There are two considerations:

1) Are the resonances causing damage to the fixturing?
2) Do the resonances cause unrealistic damage to the specimen?

If the customer isn't bothered (he/she may not know better) then question one is your only concern.

 

[JordonMusser]

bob-
The UUT is relatively light and the table/fixture is rather "serious".

I do not have a problem with it, and we do not see failures. We are moving right along, but I was able to talk upper management into giving me time(money) to at least look at it- I would like to learn something if there is anything to learn. Sounds like this is expected/normal. My Q in one of the axis is less than minimal, the other is around 4.5 which is still low.

 

[BobM3]

My worry is that the customer, not knowing what is bad, might be o.k. for now but comes back in 3 months and says "my UUT is failing unrealistically!". Make sure you tell them what is happening so they can make an informed decision and sign off on it.

 

[Tmoose]

Would it difficult to attach the accelerometer mechanically or with a larger wax footprint and re-test?

I recall I stumbled across a tech article recently about the limitations of triaxial accelerometers for machine monitoring. I could not find it when I just re-Googled. My recollection is 2 axes were often best limited to a few orders of rotation (few hundred Hz)