How much mechanical support for pcb's is required

[Geoff]

My pcb (in its housing) is likely to be hit by whipper snipers, ripped off its mounting by being hit by wheelbarrows, etc. And I would like the pcb to remain undamaged by such events.

Is there a good guide on how much mechanical support for pcb's is required or recommended?

Wikipedia has the following information:
Flexural strength (A; 0.125 in) – LW   > 415 MPa (60,200 psi)
Flexural strength (A; 0.125 in) – CW   > 345 MPa (50,000 psi)
https://en.wikipedia.org/wiki/FR-4

How do I take this information and determine how far apart the mounting holes need to be, how do I adjust for other factors like the mass of the components?


The pcb is 85mm x 95mm rectangle and will be one of JLCPCB's nice 4 layer FR4 boards. Currently, I have mounting holes in the four corners, but is this enough?

[tooki]

Bearing in mind that I have zero formal training in reliability engineering, my suspicion is that if it’s going to be subjected to mechanical abuse, then you’ll need to do testing to find out where your weak points are.

For sure, the effect of board flex depends massively on a) what’s on the board (a PCB with just some ICs and small chip resistors and caps will fare very differently from one with a big transformer on it), and b) what non-shock forces it’s subjected to in use (for example, a board with buttons or external connectors on it needs more support than a purely internal board, since every button press and connector insertion causes flex).

Flex aside, mechanical shock can cause SMD components with a high weight-to-pad-size ratio to be torn off the board; a classic example of this is larger inductors.

Thinking aloud here: depending on the types of shock, maybe a compliant shock mounting would help? Something like mounting the board with rubber grommets.


The upshot being that I doubt there’s any simple formula you can use. Maybe there exists software that can perform such physical simulations?

[Kleinstein]

The FR4 board alone can stand quite some mechanical forces.
The problem is more with SMD parts soldered to the board.  MLCCs can already fail at relatively low deformation.
So the ultimate strength of the board is not the relevant limit, but for some parts the bending of the board and for heavy SMD parts possibly the pure acceleration / shock.

[nfmax]

In high vibration environments, the big danger is mechanical resonance. The wrong frequency of shaking can cause large deflections, not just of PCBs but also of wiring harnesses. In these situations you need mounting points close enough together so that the resonant frequencies are raised above the spectrum of the environmental vibration.

[Haenk]

I'd say the easiest way is to create a resin block and maybe some rubber shock absorbers.
(Unfortunately, that's a non-repairable solution...)

[Geoff]

.
.

The upshot being that I doubt there’s any simple formula you can use. Maybe there exists software that can perform such physical simulations?

You may be right 

[Geoff]

..., not just of PCBs but also of wiring harnesses. In these situations you need mounting points close enough together so that the resonant frequencies are raised above the spectrum of the environmental vibration.

Thank you for pointing out the need to look at the wiring harness.