Re: Gold plating the entire PCB
- From: MooseFET <kensmith@xxxxxxxxx>
- Date: Sun, 30 Mar 2008 09:38:13 -0700 (PDT)
On Mar 30, 1:39 am, Guy Macon <http://www.guymacon.com/> wrote:
Content-Transfer-Encoding: 8Bit
Thanks for the reference materials.
MooseFET wrote:
Lately, I've been looking again at gold plating the entire PCB. It
makes a nice chemically stable layer on top of the copper. With the
price of copper these days, changes are the copper on the PCB will
cost more than the small amount of gold involved.
[....]
The method I was thinking about was:
"With direct gold-on-copper plating, the copper atoms tend to
diffuse through the gold layer, causing tarnishing of its surface
and formation of an oxide and/or sulfide layer. A layer of a
suitable barrier metal, usually nickel, is usually deposited on the
copper substrate before the gold plating.
I was not considering nickel under the gold. I was thinking of
strickly a copper-gold surface.
"Gold reacts with both tin and lead in their liquid state, forming
brittle intermetallics. When eutectic 63% Sn - 37% Pb solder is
used, no lead-gold compounds are formed, because gold preferentially
reacts with tin, forming the AuSn4 compound. Particles of AuSn4
disperse in the solder matrix, forming preferential cleavage planes,
significantly lowering the mechanical strength and therefore
reliability of the resulting solder joints.
Many years back Varian did gold plated PCBs. Some of these boards
still exist. Some may still even be in use somewhere. They were used
in early atomic clocks. It seems that they had a process that worked
because they had a very low failure rate on them.
"If the gold layer does not completely dissolve into the solder,
then slow intermetallic reactions can proceed in the solid state as
the tin and gold atoms cross-migrate. Intermetallics have poor
electrical conductivity and low strength. The ongoing intermetallic
reactions also cause Kirkendall voiding, leading to mechanical
failure of the joint, similar to the degradation of gold-aluminum
bonds known as purple plague.
"A 2-3 µm layer of gold dissolves completely within one second
during typical wave soldering conditions. Layers of gold thinner
than 0.5 µm (20 microinches) also dissolve completely into the
solder, exposing the underlying metal (usually nickel) to the
solder. Impurities in the nickel layer can prevent the solder from
bonding to it. Electroless nickel plating contains phosphorus.
Nickel with more than 8% phosphorus is not solderable.
Electrodeposited nickel may contain nickel hydroxide. An acid bath
is required to remove the passivation layer before applying the gold
layer; improper cleaning leads to a nickel surface difficult to
solder. A stronger flux can help, as it aids dissolving the oxide
deposits. Carbon is another nickel contaminant that hinders
solderability."
http://en.wikipedia.org/wiki/Gold_plating
Note: with an entire trace plated in gold and solder resist
stopping most of the trace being wetted with solder, there
exists a region at the edge of the solder with a high
concentration of gold, followed by a region with solder
over undissolved gold.
Also see:http://www.smtinfo.net/docs/Electronic%20Production/9.htmhttp://www.polarinstruments.com/support/cits/AP171.html
--
Guy Macon
<http://www.guymacon.com/>
.
- Follow-Ups:
- Re: Gold plating the entire PCB
- From: John Larkin
- Re: Gold plating the entire PCB
- References:
- Gold plating the entire PCB
- From: Guy Macon
- Gold plating the entire PCB
- Prev by Date: Re: Tek scope service manuals (downloads)
- Next by Date: Re: OT: American gun nuts score a classic own goal
- Previous by thread: Gold plating the entire PCB
- Next by thread: Re: Gold plating the entire PCB
- Index(es):
Relevant Pages
|
Loading
