Board bends when lay on rails

[Faringdon]

Hi,
I put a post here on this some time back, but cant find it now.
A customer sent our company a pcb to do some tests on.

The PCB sits on a ledge on each of the 150mm sides (pcb 150mm x 100mm)

It wont sit on the rails, because rails are narrower than the pcb.... you have to stand up and press it into the rails using full downward body pressure through your thumbs., and then its bent between the rails. Then you screw it in with the earth screw.
Is there any situation you have seen where this is OK?

They have only caps 0805 and less, and res 1206 and less. + 100 pin micro, soics, smd tants etc

[The Soulman]

Is there any situation you have seen where this is OK?

No, unless mechanical abuse is part of the test you need to perform.

[selcuk]

I occasionally run into issues when someone breaks vcuts recklessly and bend the boards. It is similar to your case. Once some DFN packaged ICs refused to work. MLCC capacitors may crack and you cannot see the result with a meter. I observed that after separating the boards. Some 50V capacitors can work at 5V but releases smoke when run at 24V supply. Cracking lowers the breakdown voltage.

[Gyro]

Hi,
I put a post here on this some time back, but cant find it now.
A customer sent our company a pcb to do some tests on.

The PCB sits on a ledge on each of the 150mm sides (pcb 150mm x 100mm)

It wont sit on the rails, because rails are narrower than the pcb.... you have to stand up and press it into the rails using full downward body pressure through your thumbs., and then its bent between the rails. Then you screw it in with the earth screw.
Is there any situation you have seen where this is OK?

They have only caps 0805 and less, and res 1206 and less. + 100 pin micro, soics, smd tants etc

Go back and look at your topic history - you've started enough threads on mysteriously cracked and broken smd parts. Of course there is no situation where this is going to be OK. What a stupid question.

[jpanhalt]

It's quite easy to find posts with Google.  Maybe one of these is on point:
https://www.eevblog.com/forum/manufacture/how-might-pcb-be-damaged-in-production/
https://www.eevblog.com/forum/manufacture/daughter-pcb-make-it-like-an-soic8-(the-legs)//

In any event, I believe the author is well aware of the risks from bending a PCB with SMD components that is not intended to be bent.  Maybe what the carpenter needs is not a bigger hammer, but a scraper, plane or sandpaper.  Make the board fit comfortably between the rails.  In the alternative, get a quick-set clamp, set it to spread, and make the box wider.

[SMTech]

This does all depend a bit on how much it is bending and how that bending relates to the components themselves. It is a risk, and if you can reduce that risk, you should.

We make extensive use of v-scored panels and we break out manually. The issues seen are truly minimal (unmeasurably uncommon) on the kind of bending that process exerts on typical 1.6mm boards.

 We also have a client for whom we build boards that have a massive plane of copper on the underside of a 1mm FR4 or Rogers and nothing but components and traces on the top (no fill/flood). These boards warp pretty much as soon as you open the packet and are decidedly convex post-reflow.  The client then straps these horrible things to big heatsinks with many fixing points forcing them flat again. To my knowledge they are not then reworking and replacing broken SM passives on any scales, nor would it be very easy..

Get larger chips or BGAs tho' and no doubt you'd see issues very quickly.