Sine Sweep Anomaly?

[Transient1]

Hi,

It's not my area of expertise, but I was asked if I had any comments on a 500 Hz resonance that appeared on the .5G sine sweep (20Hz-2000Hz) of four different electronics chassis. I didn't perform the FEA models for these units, but the analyses didn’t predict these resonances. In one case the analysis predicted no resonances in the 20Hz to 2000Hz range. The common natural frequencies between the four units seems faulty (they have different geometries and masses). Any suggestions on what to have the techs look for in regards to this anomaly?

Thanks.

 

[IRstuff]

Is there anything unusual about these chassis'? Are they particularly heavy or tall?

Objects that are top-heavy can sometime affect typical floating slip tables, causing the table to rock on the oil film, creating a false resonance.

Our system sounds to be much heavier, e.g., about 800 pounds and 5 ft tall. A hydrostatic bearing table doesn't have this sort of problem.

TTFN

FAQ731-376

 

[Transient1]

The approximate descriptions of the units are below:

Box (10" x 8") ~ 18 lbs
Box (10" x 9") w/2 rectangular modules that are side by side with about a 1.5 inch gap in between. ~21lbs
Box (10" x 9") w/2 rectangular modules that are in corners diagonal from each other. ~21 lbs
Box (13" x 10") w/ 8 slices on top sandwiched together and supported by long brackets on the top and sides. It is a little shorter than the units with just two modules on top. ~27lbs.

It would be interesting to note that this resonance will show up in each axis. Actually, its about 600Hz +/-50 but there are some that are closer to 500Hz. It just seems strange. I guess due to the spread it is possible, but it seems to indicate some resonance in the system. Could a faulty accelerometer have a resonance that contributes to the measured acceleration?


<<Is there anything unusual about these chassis'? >>
Not really. Two of them are very similar. I wouldn't be bothered if they showed very similar resonances, but the light unit and the heavy unit share the resonances as well.

 

[Twoballcane]

Maybe your response accel is sittng on somthing that is vibing at 500hz like a cover. The accel is only measuring on what it is attached too, so the 500hz is not necessorily the whole box, but the oil canning affect on what it is sitting on.



Tobalcane
"If you avoid failure, you also avoid success."

 

[IRstuff]

Each axis, as installed on the slip table? or is vertical done with just the driver turned vertical?

If the latter, then oil canning sounds more plausible.

TTFN

FAQ731-376

 

[GregLocock]

Have you measured the force INTO the table?



Cheers

Greg Locock

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

 

[Transient1]

IRstuff: It is the latter case. I have to check the location of the accelerometers on the other three units. The oil canning effect seems plausible for the one of the units. By the way, the units are between 10 and 15 inches tall.

GregLocock: For between 20Hz and 200Hz we see the .5 G acceleration that is supposed to be the input. Would I be looking for a force input change at that resonance?

Thanks everyone for their help.

 

[GregLocock]

Yes, probably a dip in the force psd, if everything is OK, or a peak if you have a shaker/stinger problem.



Cheers

Greg Locock

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

 

[Transient1]

Greg,

I have a couple questions:

(1) Does a dip in the force psd correspond to a unit resonance?

(2) A force peak above the base input indicates a shaker issue. The issue could be a number of things right? It could be something in the shaker resonating or a problem with our force input system?

Best Regards.

 

[GregLocock]

(1) yes, the impedance drops

(2) yes, obviously it depends on the exact layout of your shaker system. You need to measure the force where it is applied to the table, not just the inferred force from the drive signal.

Cheers

Greg Locock

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

 

[Transient1]

Well, the shaker table input accelerometer reads an almost perfect .5 G input even thru the resonance peaks. I'll see about getting a read out on how the force changed to maintain the input.

 

[IRstuff]

The base accel still might be misleading. When we had problems with our slip table, the force accel was located at the base of the DUT, and showed nothing abnormal, even though the slip table was rocking and rolling.

Is your base accel a 3-axis accel or only a single axis?

TTFN

FAQ731-376

 

[Transient1]

I still have not recieved word back from the vibe lab on the force input.

IRstuff: The base accel is 1 axis at a time. This resonance appears in each axis. When one axis is excited it is apparent in the cross talk measurements of the other two axis.

 

[IRstuff]

Right, but the point is that that the base accel doesn't measure whether the base itself is seeing the cross-talk.

TTFN

FAQ731-376

 

[Transient1]

Understood. A good course of action would then be to get accel measurements off the shaker table, correct?

 

[IRstuff]

you'll need to be more generous with the accels.

to eliminate any questions about the slip table itself, you'd need a vertical accel close to the end of the table to measure any rocking motion.

Are the accels that were showing resonance attached to the face of a large area? Were there any accels attached to a presumably stiffer structure, like an edge? Were there differences in response?

TTFN

FAQ731-376

 

[Transient1]

The issue has temporarily (maybe permanently) been dropped. Thanks for your help.

The accels were attached to a variety of surfaces on the units: Small panels and Large panels. The accel responses were very uniform in each axis. The unit as a hole is stiff.

 

[IRstuff]

Hmm..

Sounds like a table problem, then.

TTFN

FAQ731-376