[RicoElectrico]

> No one has been able to eliminate whiskering, as the phenomenon is not yet fully understood.

It's the kind of stuff we should be embarrassed not to understand. I get that understanding living things can be tricky due to complexity and issues with controlling conditions, but a lump of metal? Whatever we find out will at least save us money in damaged devices, and hopefully drive some progress in metallurgy as well.

We had an era in semiconductor manufacturing when despite the relative simplicity the process was not understood/controlled fully, which took the toll on yield. E.g. CMOS was super fussy due to difficulties in creating gate oxide - impurities in air like halogens made the yield seasonal [1]. But now I assume that if any problems arise, they're due to bona-fide complexity.

[1] https://news.ycombinator.com/item?id=28178612

[semi-extrinsic]

A lot of the challenge here AFAIK is that the process needs to be understood at the molecular level, where we measure time in picoseconds (10^-12), while this process takes something like 10^6 seconds. The disparity is an absolutely astronomical factor of 10^18.

[robomartin]

> It's the kind of stuff we should be embarrassed not to understand.

We understand it. Onset is stochastic. Mitigation is impossible given current regulations in consumer-land. Read my longer comment for further details.

EDIT: What I mean by "we understand it" is that we know that lead-free solder chemistry leads to tin whisker growth. When I was taking a deep dive into this many years back, the researchers I was working with at NASA told me "Growth onset can be 0 days to 3 years after manufacturing. Your guess is as good as mine.". And, BTW, you can have growth start in a few days in one corner of the PCB and a few months later elsewhere. It's a complex relationship of materials properties.

We know that tin whisker growth in lead-free solder is as much of a reality as gravity is between two celestial bodies. In other words, it will happen. We simply have no way to predict when or how quickly they will grow. It might just be too complex to compute/predict given the variables involved.

[gowld]

That sounds like not understanding it.

[robomartin]

Not quite. We understand that we can't build an anti-gravity device and don't even know how to go about thinking of one. Understanding that something is impossible (or likely impossible) is understanding. We might not like the answers (I sure didn't at the time) yet they are a based on knowledge and decades of research by some of the smartest scientists I have ever met.

[hulitu]

There are parts of the process which we do not underestand. Testing is expensive especially at this level and nobody wants to pay for things which _could_ happen.