Circuit Card Strength

[cbk14]

Hi all,

I have tasks to analyze some mechanical assemblies which include circuit cards (PCB / PCA / CCA whichever terminology you prefer) under particular load conditions. For the purposes of this discussion, let's stick with common FR-4.

What are the recommended best practices for mechanical stress analysis of circuit cards? I have done similar jobs and had success, but I have always felt my methods could be improved.

Specifically, I am struggling to understand the failure modes and strength figures for circuit cards. The literature is widely varied on strength limits (10-45 ksi) and some even exceed the top of this range (greater than 6061-T6!?). I find that pretty hard to believe. I have also spoken with some more experienced colleagues who have relayed stories of circuit card failures at levels much lower than expected (~10 ksi).

I have used FEA simulations on the actual board layer stack-ups (known thicknesses of all layers, material, etc) to determine effective modulus of elasticity, which certainly helps reduce deflections, but the material strength limits are not influenced by assembly details (or are they?).

So what are the recommended methods for mechanical stress analysis of circuit cards? Any pointers or resources would be appreciated.

 

[3DDave]

Not really a spacecraft question, but I have not seen any circuit boards mechanically fail in the board material. It's always been failure of the component attachments to the boards. Usually it is a failure in the solder, but I have seen fatigue failures in the leads of integrated circuits or within small, rigid components that are insufficiently compliant. Corrosion and electrical overload is what damages circuit boards themselves.

"I have done similar jobs and had success, but I have always felt my methods could be improved."

Start with telling what methods you used that failed where you never understood why.

 

[rb1957]

I think there are several issues ...

1) thermal loads (+100 to -100 every 100 minutes)
2) mechanical vibration (at launch).

The vibration loads are probably passed by test, but no doubt someone wants an analysis , and no design experience (to say "like the last design, that passed").
Sure you can do vibration with FEA, but lord knows how accurate it'll be. Maybe you can do detail tests, but you'll need a full scale FEM to understand the local vibrations. Not really my wheel-house.

Thermal are every worse (getting the local effects). Also not my wheel-house.

another day in paradise, or is paradise one day closer ?

 

[3DDave]

That would figure in determining the loads, not in defining the model.

 

[Compositepro]

You will get electrical failure of a circuit board long before you have a structural failure. As 3DDave noted most circuit boards fail due to thermal fatigue of solder joints or copper foil traces.

 

[WKTaylor]

MSFC-STD-3425 DESIGN STANDARD FOR RIGID PRINTED CIRCUIT BOARDS AND ASSEMBLIES

Its possible that the IPC may this procedure/process.

https://www.ipc.org/

Regards, Wil Taylor
o Trust - But Verify!
o We believe to be true what we prefer to be true. [Unknown]
o For those who believe, no proof is required; for those who cannot believe, no proof is possible. [variation,Stuart Chase]
o Unfortunately, in science what You 'believe' is irrelevant. ["Orion", Homebuiltairplanes.com forum]

 

[3DDave]

Possibly useful background:

https://res.mdpi.com/d_attachment/applsci/applsci-11-02679/article_deploy/applsci-11-02679-v2.pdf

 

[MacGyverS2000]

For extreme thermal environment excursions, the board either goes into glass transition or the components fail directly, whichever happens first. For continual excursions back and forth, vias crack or terminations (particularly MLCC's) crack. Never have I seen/heard of the board itself breaking before a component was vibrated or cracked off.

Dan - Owner

http://www.Hi-TecDesigns.com

 

[WKTaylor]

NOTES1.
Every MIL Acft I've worked on has a 'Vibration spectrum' document used for component analysis and testing... since vibrations tend to be the scourge [arrrrgggghh] of small/low-stiffness items. This is incompliance with RTCA DO-160 ENVIRONMENTAL CONDITIONS AND TEST PROCEDURES FOR AIRBORNE EQUIPMENT.
This 'spectrum' is specific for each aircraft model and is typically related to the installed location in the aircraft plus a generic Identification, such as 'mounted electronics box', 'wire harness', 'fluid tubing', 'accumulator', 'pump', 'seat', etc...

Small items are often impossible/difficult/expensive to analyze... but DO-160 testing is relatively low cost and very informative. Survive: to PASS! KISS.

EXAMPLE. On the antique jet that I work on, many aluminum honeycomb parts were designed not to fail at an artificially high limit load... ~3X the expected loading... then tested in a loading fixture to the limit load with NO damage allowed for production certification! NO damage or distortion/etc = PASS!!! IF there was any failure... de-bonds, cracks, permanent-distortion, etc... then the new part and all parts in it's production-run were condemned. KISS.

Regards, Wil Taylor
o Trust - But Verify!
o We believe to be true what we prefer to be true. [Unknown]
o For those who believe, no proof is required; for those who cannot believe, no proof is possible. [variation,Stuart Chase]
o Unfortunately, in science what You 'believe' is irrelevant. ["Orion", Homebuiltairplanes.com forum]

 

[hendersdc]

The rule of thumb I have heard/used is that around 500 microstrains for unleaded solder joints there is a significant chance of fracture over time. 1000 microstrains for leaded solder.

 

[cbk14]

Thanks for the replies, all.

I guess my primary question was something like "what is a reasonable yield/ultimate strength for typical FR-4 PCBs?" but that question now seems moot. PCB construction (# of layers, copper thicknesses, etc) varies widely so there's probably not a convenient figure like there is for a common material like 6061-T6.

I agree solder joints are likely the limiting factor in assessing circuit board survivability. Towards that end, the method I've used in the past is Steinberg fatigue analysis. This requires the expected displacements from an FEA, which is where I currently am. I have completed some FEAs to determine modes, random vibe response, srs, etc. But now I am left trying to compute MS to an unknown material strength limit.

My customer doesn't want a CDI / Steinberg analysis, so I am somewhat stuck trying to calculate MS against a material with a poorly defined strength.