5

I was told that electronic components and printed circuit boards have a limited lifetime. This is backed by several questions and answers:

https://www.quora.com/Do-circuit-boards-have-a-useful-life https://www.quora.com/How-long-do-circuit-boards-last

It is however unclear to me if this is because of electronic wearing of the components, or environmental influences causing the connections to break, parts to overheat, etc.

So, considering an average printed circuit board with several IC's, a bunch of resistors and some capacitors that is connected and working under ideal circumstances (controlled temperature, humidity, etc). Will this PCB or its components eventually break because of the electrons that are moving though it, or will it fail because of other reasons?

Does the same failure occur if the PCB is running in a less controlled environment like a regular office or is it likely to be something different? And is it in the same time span, or significantly earlier?

Neograph734
  • 153
  • 6

2 Answers2

5

If water condenses on your PCB, then salts left over from manufacturing processes will turn it slightly conductive. If the PCB is powered, copper from exposed pads and solder will be corroded by electrolysis. This is what kills your phone if it gets wet and you leave the battery in. It goes a lot faster than you think.

You can replace "water" with sweat, dead bug juice, rodent pee, whatever...

On a very humid day, you may hear a loud bang followed by a puff of smoke: that used to be a chineeese USB charger with 0.5mm mains creepage distance.

Now, under normal circumstances, everything has a finite lifetime...

The first thing to fail is usually a capacitor which dries up. Heat accelerates the process. If the capacitor overheats, bulges, and leaks electrolyte all over your board, this will corrode everything it touches.

Solder joints (and semiconductors) can crack under heavy vibration or due to thermal cycling (the famous XBOX ring of death).

Mechanical shock (ie, dropping the stuff on the floor) will cause heavy components to lift their pads, or just break off the board. Shock, or board flex from an over-eager user sticking a huge CPU cooler in their PC will crack brittle stuff like MLCCs or ferrites... or BGA solder balls.

Connectors which are manipulated often will wear out, or they will break off the board if they're pure SMD.

Vibration will loosen screws. Murphy's law ensures it'll fall off in the worst possible place.

Cable which is flexed too often will break.

Electrical contacts which are secured by screws can eventually loosen, then overheat and burn if high current.

Multi-strand wire will corrode over the years if the air is contaminated.

Fans fail. Air filters clog up with dust.

Software is buggy.

Then...

Semiconductors under high heat/voltage/current stress can fail due to thermal cycling or electromigration.

Eventually, flash memories will forget their contents.

Lead-free solder grows whiskers which will cause shorts.

However, unless you're designing a guidance board for a Minuteman missile, your stuff will probably be obsolete before it fails!...

What I mean is that the most likely causes of failure should be addressed first.

bobflux
  • 70,433
  • 3
  • 83
  • 203
  • _However, unless you're designing a guidance board for a Minuteman missile, your stuff will probably be obsolete before it fails!..._ So you mean that all failures you have described will occur after multiple years of use, while technology evolves at a much higher pace? And the main cause of failure is because of external influences and not the actual moving of the electrons? – Neograph734 Mar 02 '17 at 13:55
  • Electromigration is electrons pushing atoms out of the way when current density is too high. Chip designers know about this and design semiconductors accordingly. It is well-studied, you shouldn't worry about it too much. The most likely cause of failure is usually a capacitor, a connector, aging of batteries, stuff like that. Or simply obsolescence for consumer gear. Now, IF you design a system that must both work for a VERY long time, and MUST have extremely high reliability, then you can think about super expensive mil-spec parts, ceramic packages, etc. Cost is extremely high. – bobflux Mar 02 '17 at 14:10
1

From good old MIL-HDBK-217 we get this failure rate formula, which accounts for electrical as well as environmental factors (and others):

Failure rate formula

Find here a brief but complete overview on reliability engineering.

Enric Blanco
  • 5,741
  • 6
  • 22
  • 40