Design of Vibration Fixture Plate 1

[EnginerdNate]

Hi All,

I've been tasked with heading the design a effort for a mechanical vibration/shock test (Following MIL-STD-810G methods) and I'm having some difficulty with the fixturing. I'm a structural analyst/designer by trade so I'm familiar with the programs, but my bread and butter design wise is mostly design/analysis to a static strength spec. I'm well out of my depth on this dynamics stuff.

The part attaches to the fixture at four points and the CG is quite high--~20" from the mounting face. This is not something I have the ability to change. Currently the fixture mounts to the table with a large number of 3/8" bolts on 2" centers.

I have run several variations of the fixture through modal analysis in NX Nastran. I represented the unit under test by a point mass (Including inertial properties) linked to the four test mounting points via RBE2 rigid elements. I have modeled the table interface with a combo of cbush elements using Houth's stiffness model to model the mounting bolts and cgap elements with arbitrarily high stiffness to model the 'no penetration' condition of the interface between my mounting fixture and the shaker table.

It seems no matter what I do, I can't push the modes related to the unit under test 'wagging' in the three primary directions (Side to side in the two axis and in/out of the plate) out of the test frequency range. I've successfully doubled the resonant frequencies with design changes but it started at such a poor point that doubling it is still right smack in the middle of the applied spectrum.

If I analyze the fixture by itself the modes are all 2-3x higher than the test frequency range. I am concerned that I am going to run into problems with total mass if I just keep making the fixture thicker at the mounting locations. Any input?

Some specific questions:

1. Should I be including a fascimile of the table itself in my analysis? I could get the dimensions of the plate from my test vendor.

2. Is there ever a situation where the design of the device under test hamstrings you in designing a fixture? The actual part design is bassed of off work that was carried out well before vibration testing was a standard thing for this class of component.

3. If my spectrum has a 6db/octave taper at each end of the test spectrum, do I need to worry about resonances that happen in that taper or am I primarily worried about resonances in the "full intensity" portion of the spectrum?

4. Adding thickness at the mounting points and increasing the support constraint definitely seems to help, but there seems to be diminishing returns. In my last model, the lowest mode behavior in the fixture seemed pretty much constrained to an area immediately surrounding the attach points. Is this a situation where switching to steel for it's higher stiffness would be acceptable (currently using aluminum for everything) or is ringing etc such a problem with steel that that's a bad idea? I'm running out of ways to stiffen things without looking at material changes.

Thanks for any input you can provide.

Best,
Nathan

 

[Tom Irvine]

You should run your fixture design by the engineers or technicians who will actually be running the vibration test for a sanity check. Modal frequencies within the test frequency range are somewhat inevitable for common test items.

There are two concerns: test control stability and over-testing the component at its natural frequencies.

Component natural frequencies can interfere with the test control stability in some cases. The voltage drive signal to the shaker can be adjusted automatically by the control computer by "notching" the input at the natural frequency or frequencies. This usually works if there is sufficient damping present.

Large, unwieldy test items can be tested using "extremal" or "max line" control to prevent over-testing at the natural frequencies. This involves mounting several control accelerometers on the fixture and possibly even the test item itself. The control is made to whichever accelerometer gives the highest response at each frequency.

I have posted some notes at:

https://vibrationdata.wordpress.com/2020/08/11/shock-vibration-test-fixture-design/

But this is a "work-in-progress" blog.

Best wishes,
Tom Irvine

 

[IRstuff]

Is the UUT top-heavy as installed on the fixture? This can cause a conventional slip table to wobble on its oil film.

A hydrostatic bearing table or

https://teamcorporation.com/images/brochures/TFilm_Slip_Tables__2.11.19.pdf

should be specified.

TTFN (ta ta for now)
I can do absolutely anything. I'm an expert!

https://www.youtube.com/watch?v=BKorP55Aqvg

faq731-376 forum1529 Entire Forum list

http://www.eng-tips.com/forumlist.cfm

 

[EnginerdNate]

Thanks for the inputs guys,

I participated in a successful test of a similar test article last year but was not involved in the design of the fixturing, just the test execution. We had some trouble on that test due to the table to adapter plate attach bolts being excessively long but we were able to get around it. I have incorporated what I learned from that experience into this design (and spoke with the designer of that fixture as well) and specced deep counterbores in the attach plate to keep the bolt length down.

Tom-I did exactly what you recommended and looped the guys at the test-house into our design process. They liked my final design well enough to OK moving forward with fabrication of the test fixture so this one is 'case closed' until we run the test and find otherwise. FWIW, the test engineers predicted based on what they saw of my design and their experience that when we ran the resonant sweep in real life that the system would perform better than my analysis was predicting.

Prior to involving the test-house, I was able to push the predicted modes out of the main test range by using a combo of stainless steel and aluminum for the test fixturing and redesigning the fixturing to provide a direct loadpath between the primary UUT mounting points and the vibration table surface. Accounting for a small amount of friction between the adapter plate and the table surface also helped. I set the coefficient of friction to .2 as a conservative estimate, figuring it was likely some oil from the slip plate would be present at the interface.

IRStuff--I'd say it's somewhat top heavy but not once the attach fixture is added. The fixture weighs 2-3x what the UUT weighs. The test house has full details of our UUT and has approved the design of the adapter plate, so I leave it up to them at this point to know the capability of their equipment.

 

[IRstuff]

If you see an unexpected resonance in the 20-30 Hz range, you'll know where that's coming from ;-)

TTFN (ta ta for now)
I can do absolutely anything. I'm an expert!

https://www.youtube.com/watch?v=BKorP55Aqvg

faq731-376 forum1529 Entire Forum list

http://www.eng-tips.com/forumlist.cfm

 

[EnginerdNate]

Good to know!

 

[Tmoose]

Got a general arrangement of the UUT , fixture, and table you can share?

I promise not to sell it to the Ruskies, China, DPR Korea, Iran, etc.

 

[EnginerdNate]

Hey TMoose,

I'll upload a quick sketch in a moment.

UUT has an attach feature about 20" from it's CG. Attachment is via a number of pins, and bolts. If the x-y plane is lying flat and the z-axis is pointing up out of the table, the CG is near zero in the x,y plane and ~20" up the z-axis. (0,0,20).

Vibration fixture has an beefy aluminum ring (~30" diameter) that interfaces with the mentioned pins and restrains the UUT in 'shear' plane only (in plane with the top surface of the fixture). This ring is bolted to the slip plate on a 2" rectangular pattern with fastener locations chosen strategically to prevent making the test tech hate me more than necessary.

The ring has several full thickness slots in it's outer perimeter roughly 3"x3" that steel posts bolt to in shear. These are slip fit into the slots. These steel posts are match-drilled to the aluminum ring such that their bottom face maintains contact with the vibration slip plate after bolt up to the ring. This is to help constrain the blocks in the z direction.

These blocks pick up the bolted connections on the UUT and are preassembled to it prior to mounting to the vibration table. They get lowered with the UUT to mate to the vibration table/aluminum ring. I was going to add some fasteners in these blocks directly to the vibration table, but they had to be outboard of the aluminum ring and seemed to provide little benefit in my analysis. The test-house tech agreed with that assessment and recommended removal of these fasteners from the design prior to my doing so.

Fastener pattern between the steel blocks and the aluminum ring has about 10x the static strength capability of the joint between the UUT and the steel blocks. The latter is dictated by the UUT design. I wanted to make sure that that shear joint had an absolutely prodigious amount of clamp up during test to augment the no-penetration condition between the blocks and the table/sides of the slot sin the aluminum ring.

I came to this solution after doing about 7 rounds of iteration on a one-piece design with clearanced 'slots' between the table and the aluminum ring to allow for tightening of the mounting bolts on the UUT. They have to come in from the backside. These areas were experiencing a 'drumhead' like excitement in my modal analysis that was not improving simply by making the cutout smaller or the aluminum ring taller. I was running into a 'diminishing returns' situation with that setup.

Cheers,
Nathan

 

[EnginerdNate]

Apologies for smudge. I'm a lefty. Bolt pattern is just to convey the idea, Didn't have time to make it all perfect.

 

[EnginerdNate]

Just an update for you all, with the thumbs up from the test engineers at our chosen test lab, we went ahead and had the fixture as I designed it fabricated. The testing was completed without a hitch and the test technician stated he was impressed by the performance of both our fixturing and our test article.

All's well that ends well!

 

[IRstuff]

Great! Wish I had as good luck; been at test facility for over a month, and managed to break all sort of things.

TTFN (ta ta for now)
I can do absolutely anything. I'm an expert!

https://www.youtube.com/watch?v=BKorP55Aqvg

faq731-376 forum1529 Entire Forum list

http://www.eng-tips.com/forumlist.cfm

 

[EnginerdNate]

The last vibration test I attended the lab was very laid back about letting us help get our fixturing etc in place on their machines. Right up until I stripped a keensert out of their magnesium slip plate.