The definition of electromigration is migration of conductor material in a
circuit trace due to the momentum transfer from charge carriers (usually
electrons) flowing in the conductor traces and the phenomenon bears
characteristics of diffusion. The species undergoing electro migration need not
be ionic, or charged in any way. The electron flow is simply pushing the
conductor metal out of the way.
Electromigration CAN NOT TAKE PLACE in the absence of significant current
density flow in the conductor in question.
In electromigration, conductors migrate atomistically in a direction orthogonal
to a current flow, causing flow of conductor material in a path that leads away
from the conducting trace. This phenomenon can cause either a short circuit to
an adjacent trace or an open circuit in the trace if sufficient material
migrates. Much effort to reduce this effect, especially in Aluminum
metallization in IC chips, was undertaken in the late 1970’s. Electromigration
on the macro scale, the size of PC boards and associated parts, has never been
documented.
More likely what you saw was solder whisker formation, a recurrence of what
took place decades ago with low Lead solders and which is coming to the
foreground once again with Lead-free formulations. These whiskers can grow due
to heat, stress and other environmental stresses, mostly in purer metals, but
can also take place in certain alloys (no need for current flow) and cause the
formation of short circuiting bridges as you describe. The small sizes of the
whiskers, on the order of the wavelength of light, is what gives rise to the
whitish appearance. These types of whiskers caused failures in satellites and
in some nuclear power plant metering systems. The term “electromigration” is
incorrectly applied to this phenomenon even in Wikipedia. They correct their
mistake later in their discussion when they point out that no current flow is
required for whisker formation to take place. Current flow is required for
electromigration to take place, though the current can be a corrosion current
in an appropriate corrosion cell.
They do show some pretty cool images of such whiskers, many of them several
millimeters in length, more than enough to cause the problems you experienced.
If you had a small sample of the stuff you removed, I could send you some nice
scanning electron microscope images of them along with their compositions… That
would take about 15 to 30 minute of my time.
73
Gary
> On Oct 16, 2016, at 10:48 PM, A R <raf_3@msn.com> wrote:
>
> Just recovered from a failure of my Orion 565 that looks to be a result of
> ionic solder migration.
>
> Orion was purchased two years ago from original owner's estate. Details of
> his operation/use are unknown.
>
> Replaced snaphat battery, A9 caps, and intermittent VFO encoder. Orion
> performed flawlessly for over a year following those measures, and was
> relegated to back-up status about 3 months ago.
>
> When it was brought back into service, receiver worked fine, but no
> transmitter output on either ANT 1 or ANT 2 outputs. Two master resets were
> performed, but no joy. Covers were removed, and interior subjected to the
> burnt component "smell test", and visually inspected for any obvious causes
> (loose cable connectors, etc.). Discovered powdery substance on the I/O board
> that surrounded the entire periphery of TXEN 1 rca jack solder pad (perfectly
> circular around solder pad) intersecting the adjacent +13vdc wire conductor
> solder pad, with a "track" extending to ground side (anode) solder pad of
> adjacent 5KP15A surge suppressing diode. Removed the powdery deposit using
> Q-tips and 90% isopropyl, followed by judicial flushing with distilled
> water.Gently and slowly dried with warm (not hot) air, and after 10 minutes
> of ambient air dry time, applied power. All sytems are "go", and the Orion
> lives again.
>
> I have never used the TXEN or TXOUT jacks, but don't know whether the
> original owner did. At any rate, the condition (which I attribute to
> electromigration of the solder) had to have resulted/propagated with no
> externally applied potential to the TXEN jacks, since the condition only
> became "fatal" after two years of problem-free operation/life. The internally
> applied potential to the TXEN pad/lead is just +3.3vdc. And, the distance
> from the TXEN's pad edge to the diode's anode (ground) pad edge is more than
> 1/2 inch, and distance from the +13vdc pad edge is approx 3/16 inch. Since
> the deposit was uniformly centered around the TXEN's pad (and, not the +13vdc
> pad edge), it would seem that the donor was the TXEN pad/solder.
>
> Given the wide spacing between these pads/solder and the low potentials
> involved in this case, the potential (no pun intended) for similar failure
> conditions elsewhere (with closer pad separation) would seem quite possible.
> So...I guess my counsel is to do a close visual inspection (under
> magnification) of boards when unexplain/unprovoked failures occur. Or, maybe
> better...periodic visual board inspections...or, even cleaning...BEFORE
> electromigration causes a failure. Fortunately, in this case, the condition
> didn't result in permanent (and catastrophic) damage. An ounce of
> prevention....etc.
>
> 73,
> Allen--W7GIF
> _______________________________________________
> TenTec mailing list
> TenTec@contesting.com
> http://lists.contesting.com/mailman/listinfo/tentec
_______________________________________________
TenTec mailing list
TenTec@contesting.com
http://lists.contesting.com/mailman/listinfo/tentec
|