Robert Cringely blogs on technology (mostly computers) for PBS. The subject of this week's column is a bit different: tin whiskers. Anybody heard of them or had experience with them? Here's this week's column.

Skip Montanaro

Its real, and it's a huge problem. Here's a link to NASA-Goddard's tin whisker home page. At the top, click on "Photo Gallery", and then come back and check out "Whisker Failures". There's a list of satellite, missle, nuclear power plant failures that have been attributed to tin whiskering.

http://nepp.nasa.gov/WHISKER/index.html

The problem has been known about for decades. It was seen in WW2, when radios were failing due to tin whisker growth shorting out the air capacitors in radio tuners. The fix was simple- add 3% lead to the tin, and voila, no whiskers. Now that the European Community has banned lead, the component manufacturers are turning back to pure tin coating on their components' (meaning IC, resistors, capacitors, etc) leads.

Military, medical, computer servers are exempt, but the chip manufacturers don't want to stock two types, so they're converting over 100%, at least many of them. There's no such law here in the US, but we have to buy our parts from the same sources. Lead times for old-fashoned leaded parts have gone from overnight to 3-4 months, if you can get them to run a lot for you at all.

It isn't just a problem with electronic components. Anything plated with pure tin will whisker. A ring lug, a bracket, you name it.

And it isn't just tin. Silver will whisker, as well as other metals...

There are very few options to mitigate the problem. MIL spec printed circuit boards are coated with a clear insulator- conformally coated- but the stuff grows right up through it. Replacing the pure tin with tin/lead alloy works 100%, but you usually can't get to every nook and cranny of the part, and you can't get inside.

Dryness, humidity, or even vacuum in space don't speed it up or slow it down. It has to do with the intermetallic layer between the base metel that you are coating, and the tin plating. There's a residual stress that over time literally extrudes tiny whiskers of tin out through the tin metal of the coating.....

Yes. Big problem when I worked at IBM.

I think maybe I have had problems from this. I never heard of tin whiskers but it may explain the problems I have seen several times. On the equipment I service we use a large wire wrap nest. This is multi layered frame. It has both +5 volt and ground layers. Sometimes a short develops to a signal wire. I cannot see them but the short is there. I blast the short clear using a 12 light bulb to either ground or to +5 and this clears the short. The bulb never lights so the short is very small. Probably one of these whiskers.

Thanks for sharing