Low freq shaker motion subtract from unit motion

[floattuber]

I've been told that at low frequencies, <3ish Hz, that the motion of the whole airbag supported shaker subtracts from the motion of the unit under test, but I'm having a hard time wrapping my head around it.

For example, when we ran a dwell at 2 Hz the whole shaker moved, about 1/2", roughly out of phase of the slip plate. This makes sense because the frequency is close to the natural frequency of the table which is supported by airbags. I've been told the 1/2" of shaker motion should subtract from the motion that the unit sees, but that doesn't make sense to me.

Assuming it's under control, the control accel still sees the required acceleration. This acceleration translates into the required displacement and therefore, it seems to me that the shaker motion does not subtract from the unit displacement.

Am I missing something?

 

[MikeHalloran]

The displacement of the UUT relative to the floor/ building/ local 'ground' is what's important.

To the extent that the shaker base moves, the UUT is not moving, and not being tested.

Similarly, I have been in a commercial test lab where the UUT was pretty much stationary, and all the shaker's activity went into flexing the spindly fixture between the UUT and the shaker. A totally bogus test, producing totally worthless data. The poor schmuck customer didn't have a witness present, so probably thought the UUT design was good enough.





Mike Halloran
Pembroke Pines, FL, USA

 

[GregLocock]

But if the accel is on the UUT there is no problem.

Incidentally we often hang small shakers from vehicle bodies via their stingers, using their inertia to react the driver force. Works very well.

Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[Twoballcane]

Are you doing random or sin? If random, it does not matter because you’re doing such a broad band that all frequencies will be excited. If sin and your spec is 2 Hz than that is what your UUT has to endure. If your Fn of the UUT is near 2 Hz, you will have problems, but if it is two or three octave away from Fn, then your ok. If your controlling accelerometer says it is moving 2 Hz then the table is moving at 2 Hz.

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

 

[Twoballcane]

To add, if you programed in 2 Hz at say 5g and your controling accel says it is moving at 2 Hz at 5g then your table is moving at 2 Hz at 5g which your UUT will see.

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

 

[floattuber]

Mike - "The displacement of the UUT relative to the floor/ building/ local 'ground' is what's important."

Greg - "But if the accel is on the UUT there is no problem."

I think this is where I'm getting confused. Our control accel is on the unit. It seems to me that the displacement relative to the floor is not important and what is important is what the control accel sees.

Like Twoball's example, the unit should be seeing 2Hz and 5g's.

 

[MikeHalloran]

Greg is correct. If an accelerometer attached to the UUT sees 5g, then that's what the UUT is seeing.

I'm just not used to having accelerometers.



Mike Halloran
Pembroke Pines, FL, USA

 

[IRstuff]

Unless you are absolutely certain about your fixturing and equipment, accels are pretty much mandatory. There are at least two sets of accels, the control accel, and a set of measurement accels. The control accel is what verifies the shaker output complies with the requirements.

HOWEVER, one should also be aware of any potential influences caused by the table itself. We had a UUT that experienced an unpredicted resonance around 35 Hz on a slip-table shaker. After months of this and that, we determined that the resonance was actually a shaker table mode excited by our UUT. Turns out the slip-table rides on a film of oil, and if the UUT is sufficient top heavy or tall, it can cause the granite slab to rock on the oil film, resulting in a rocking mode resonance. We wound up changing vendors to get a hydrostatic table shaker, which eliminated the problem.

So, if your shaker setup or fixturing is dubious, they may introduce modes that aren't actually coming from the UUT.

TTFN

FAQ731-376

 

[Twoballcane]

Sorry, let me be clear, the control accel should be on the table where the UUT is fastened "not" on the UUT. You’re trying to find out how the UUT responds to your forced excitation. Better yet, if you can get two or three control accels on the table and take the avarage would be better. Then you should have some monitoring accels "on" the UUT to get your responses.

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

 

[GregLocock]

Engine ancillary specs, for example, are not written that way, the test is defined in terms of what the UUT sees, not the base excitation. Reason for that is that the bracket design is critical.

Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[Twoballcane]

I see what you are doing, but if you want to design something new or swap a different part that is not the same, how do you test it if your input is the old UUT's profile? The new UUT will have it's own profile. The force frequency will be the same at the base, but how two different parts react to it will be different. Just asking

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

 

[IRstuff]

That's why you have to start with a control accel that the controller uses to match the input stimulus, followed by however many measurement accels are deemed necessary, which is dependent on whether the UUT is complex or not. The larger the UUT, the more likely that there are complex modes occurring and the more likely you'll need additional accels.

Seems to me, part of the discussion here is an issue of terminology. When some people refer to "what the UUT sees," they're referring to the input stimulus, ala MIL-STD-167, or the like, not the end result (what the UUT does when it "sees" a particular stimulus.

TTFN

FAQ731-376

 

[floattuber]

Thanks all, looks like I was right.

We've done it both ways here, with the vast majority having the control accel(s) right next to the unit on the fixture. For the sake of this discussion, I just said "on the unit" for simplicity.