Gordan, having done some research in this area myself, I saw that part of this problem was derived from mfg'rs trying to keep costs low. A nickel under plate between the copper lead frame and pure tin plate would essentially stop the intermixing of tin-Cu intermettalics which drive the formation of tin whiskers.
An over plate of electroless ni is another added step that producers do not want. Why? Their installed plating systems were designed for one pass tin lead alloy plating - not a two step plating process. It may work, but getting producers to adopt it is another question.
"Both the RoHS directive and the tin-whisker issue present challenges, but they are not insurmountable. With proper choice of solder, manufacturers of high-reliability devices can achieve RoHS compliance while still maintaining reliability, performance, and cost objectives."
Funny he would end the article with this rosy summary, when the bulk of the article is describing how there is really very little that can be done to avoid reliability problems with lead free circuit boards.
Exactly why is everyone still falling in line with this dumb, PC idea? Talk about diminishing returns. 80% of lead is in vehicle batteries which are total exempt, but we must stamp out the 0.03% usage of lead in ICs. What a joke.
Conformal coatings seen to have the greatest impact in tin whisker growth mitigation, with Parylenes applied at 1 to 1.5 mil showing great ability to inhibit their growth.
Parylenes are solvent and catalyst-free thin organic coatings that offer solutions to many existing challenges of the electronic industry.
Parylenes truly conform to the parts due to their molecular level deposition characteristics.
Contact SCS Coatings for more details.
Most of the manufacturers are using pure tin coatings for RoHS components. However, those of us assembling these parts are using Tin/Silver/Copper alloy solders to maintain RoHS status. The main impetus behind Tin/Silver/Copper alloys as we have been told is to avoid the beta-tin to alpha-tin transformation in cold-exposed applications. This paper addresses pure tin. What about these Tin/Silver/Copper alloys and whisker phenomena?
The coating is so thin on terminations because it is applied as a liquid. Liquids do not coat the sharp edges so characteristic of terminations. The best that can be hoped for is that the whiskers that penetrate the coating from below (i.e., from the tin termination finish) do not penetrate the coating on adjacent terminations.
A new way of coating the tin, with a metal cap layer instead of a polymer, was presented in a paper at the most recent CALCE Tin Whiskers Symposium (available at http://www.ldfcoatings.com/slides/Day1-13-Davy.pdf). The most practical metal appears to be nickel, and it can be deposited exclusively on metal surfaces by what is known as "electroless" metal deposition. Whiskers penetrate some metals in days, but a nickel cap only about a hundred atoms thick (35 nm) has been found to be virtually impenetrable. The paper argues that a nickel cap thicker than this does not simply _retard_ penetration, but truly prevents it - permanently.
Gordon Davy, Ph.D.
Mr. O'Boyle has written a superb article on the subject of tin whiskers, and documented it well, too. He states that a conformal coating "mitigates," not "eliminates," the risk of a whisker-caused short circuit.
The original purpose of conformal coatings is to prevent not whisker shorts but dendritic growth shorts. The coating covers a flat board and its lands and solder connections with an insulating layer that prevents the formation of a continuous layer of water between adjacent areas of metals with different electrical potentials, thereby preventing dendritic growth. (Among the many differences between dendrites and whiskers is that the latter grow from component terminations. They grow away from any surface, and if long enough can cause a short circuit.)
Getting a thickness on component terminations anywhere close to the Table 2 values, using _any_ of the commercially available conformal coatings has proved to be impracticable. (This impracticability can be seen in cross section by measuring the coating thickness on _any_ termination of _any_ component removed from _any_ assembly that has had a conformal coating thickness applied to meet the J-STD-001 requirements.)
To be concluded in next posting.
Blog Doing Math in FPGAs Tom Burke 8 comments For a recent project, I explored doing "real" (that is, non-integer) math on a Spartan 3 FPGA. FPGAs, by their nature, do integer math. That is, there's no floating-point ...