TechNetters,
The following is my original message, which is followed by conclusions which
I have drawn based on the information received from fellow TechNetters.
Original Message:
I am writing with respect to a 4 layer (mixed technology) PCB with the
following layering:
Component side traces
Ground plane
Vcc plane
Solder side traces.
We have, on three occasion, had boards burn out after the boards were in
service for some time. The burns are between the 12V battery traces and
other traces/planes. The 12V is fused with a 1.5V quick fuse. The post
mortem on one occasion indicated that the short occurred between a component
pin hole and the internal ground plane. The other two cases are not so
clear, as the short occurred on an inside layer along an edge of the board.
The spacing generally is 0.015".
I recall to have read in the past about wisker growth. Can anyone point me
to such literature, preferably in EDN, Electronic Design. Any help would be
appreciated.
End of original text.
Comment:
I have received numerous replies with useful direct information or
references, which made me understand the possible failure mechanisms, they
being either 'wisker growth' or 'dendritic growth', the latter also being
known under electromigration .
WISKERS, which are in the range of 1 to 5 um in diameter and generally up to
a few mm in length, can grow without applied voltage. They are normally
straight and of single crystal structure. Pure tin is a favorite material
for wisker growth. Adding a small amount of lead or other material reduces
the likely hood of wiskers. Stress is another promoter of wisker growth.
Wiskers can grow away from a surface or between a board surface and the
conformal coating.
DENDRITES are formed in an identical process as electro-plating
(electromigration), requiring an electrolyte and a potential difference of
only a few Volts., eg. some salts, moisture and current. Salts are at times
the unwanted results of board manufacturing residues. Dendritic growth
usually shows branched out configurations of larger cross sections.
Conclusion:
Contrary to my original suspicion, wiskers do not seem to be the culprit in
our case. The shorts occurred inside the 4 layer board where there is no
pure tin present. Copper is much less likely to form wiskers. Also wiskers
are so thin, so that I doubt that they could have sustained so much current
to burn the board before acting a as fuse.
Dendritic growth is more likely, as I see more likely hood of larger
currents due to the larger cross section of the short.
In the discussed board the problem occurred in two cases along the edge of
the board for about 2-3". The 12V battery (VBAT) trace is along this edge on
the Vcc plane. There are also break-away tabs for the carrier strips. The
close proximity of the VBAT trace to the edge may have prevented a proper
edge seal which was at places further diminished by the fraying and partial
delamination at the break-away, thereby allowing moisture to enter the
lamination.
Two reports pointed to fraying of the glass fibers along the component hole
due to aggressive drilling, thereby allowing plating salts to be entrapped.
These in turn progress along the fibers and may eventually result in shorts.
I do not want to include all references in this report, but I am glad to
e-mail directly to interested parties. I hope I do not put my foot in the
mouth and open a flood gate of requests.
Rolf Ehrat, Sen. Des. Eng. Tel. (905) 669-2280 ext. 320
Scintrex Ltd. FAX (905) 669-5132
222 Snidercroft Rd.
Concord, Ont.
L4K 1B5
***************************************************************************
* TechNet mail list is provided as a service by IPC using SmartList v3.05 *
***************************************************************************
* To unsubscribe from this list at any time, send a message to: *
* [log in to unmask] with <subject: unsubscribe> and no text. *
***************************************************************************
|
