Circuit board assembly vibration analysis

[ddddarkbull]

I got this project to perform vibration analysis for an electronic equipment used in the aircraft. Inside this equipment, there are stacks of circuit board. These board has mix of surface mount as well as through hole components. The analysis will pin point if there is any problem in the circuit boards to meet the vibration requirement.

The question I have is if there is any general guideline to determine the allowable displacement for the circuit board to meet certain number of test hours.
I like to see how stress engineers deal with this type of analysis. Ideally, I would like to point to a generic spec (e.g Mil spec) in my analysis report.
If there is no such spec existed, does it mean I need to model all critical components and their joints to determine the stress level. (I hope this is not the case, as there are thousands of components on the boards and I can easily identified over 100 of them are in some higher risk area--the area subject to highest stress reversal).

I believe Steinberg got some equations to determine the acceptable natural frequency of the board. But the equation is for specific type of components and in specific spot in on the board. I am looking for a generic guideline for board with mix components.

 

[IRstuff]

Allowable displacement is a function of design/construction and specifications.

TTFN
faq731-376

 

[Twoballcane]

You may want to look into Mil STD 810 under vibration for the random vibration profile of your desire. However, I’m a big fan of Steinberg and as indicated in his book the largest deflections are at the middle of the board (if all fours sides are simply supported, never assume rigid support). So there will be high probability that the components in the center will see the highest stress. Also, when you do your random vibration analysis, this will lead into a fatigue analysis which is your goal. So, follow the instructions and examples in chapter 9 in Steinberg’s book and see what you get. To add, don't forget to do a shock analysis too.

Tobalcane
"If you avoid failure, you also avoid success."
“Luck is where preparation meets opportunity”

 

[ddddarkbull]

Thanks for the reply.

As written in my original thread, Steingberg's calculation is mainly based on specific component type and board. My challenge is for high complexity board with many components. The question will remain, under certain vibration environment with given physical data (e.g. board size, thickness, component, mounting configuration....etc) what will be the criteria to determine if the board is at risk of failure without modeling each individual components before prototypes are made.

Mil-STD 810 is an environmental test spec. I couldn't remember seeing the kind of information I am looking for.

 

[Twoballcane]

If you look up “Figure 514.6C-5. Category 7 - Jet aircraft cargo vibration exposure” you will see a generic requirement for aircraft vibration. Pick the aircraft that best suits your situation and then you will know what PSD profile your electronic equipment may be exposed. There is no magic bullet equation that will give you the final answer. You will have to break out your pencil and calculator and do the fatigue analysis. Of course you cannot do all, but you do a few in areas where you will see the most deflection on your CCA. Start with a few in the middle and then in one inch radius calculate a few more and so on. Now be careful with degrees of freedom. The PSD profile is at the fasting points of your box. You will have to calculate out to the CCA fasting points to see what transmissibility (Q) your CCA will see. It may be amplified or best case damped.

Tobalcane
"If you avoid failure, you also avoid success."
“Luck is where preparation meets opportunity”

 

[ddddarkbull]

Thanks for the response. I wish there is "safe" acceptable level of board displacement based on physical data of the board and vibration level of the environment. Sounds like there is no quick and dirty solution to this.

 

[Twoballcane]

There are too many variations in CCA designs to come up with one solution. It all comes down to the failure theories and in this case of random vibration its fatigue analysis. If you want to get a quick feel of what your device can handle, just calculate out Grms and multiply it by 3 and that will be your worst case G level. And then from there you can calculate out the max priniciple or von Mises and compare this with yield strength or ultimate strength of the material.

Tobalcane
"If you avoid failure, you also avoid success."
“Luck is where preparation meets opportunity”

 

[ddddarkbull]

Thanks Twoballcane. A bit painful, but, I guess it's the only way.

 

[Twoballcane]

Wait till you have to do the thermal anlaysis for junction tempertures for each BGA, Flat Pack, etc...

Tobalcane
"If you avoid failure, you also avoid success."
“Luck is where preparation meets opportunity”