[Amps] current threads about silver solder and resoldering

[Michael Tope]

In my experience high-rel IC pins are often passivated with a thin layer 
of gold plating. Gold and tin-lead solder are a bad combination (Indium 
solder and gold is even worse) due to the poor characteristics of the 
resulting alloys, so high-rel assemblers are often instructed to 
"tin-wick-tin" gold plated leads before soldering. This deposits the 
bulk of the gold flashing into the discarded solder, thus minimizing the 
risk of gold "inter-metallics".

I've seen cases first hand where small amounts of contamination in the 
solder can radically alter its characteristics causing unwanted re-flow 
during equipment burn-in conditioning (at temperatures well below the 
intrinsic melting point of the uncontaminated solder).

73, Mike W4EF.............

On 8/28/2012 2:56 PM, Manfred Mornhinweg wrote:
> When I repair old equipment, indeed I often find myself removing old
> solder from a joint as completely as I can, and soldering it with new
> solder.
>
> I can come with several plausible reasons why the old solder works
> terribly if I just try to reflow it applying some flux:
>
> - Antique solder is often 30/70 tin/lead! That solder goes through a
> very long paste phase when hardening, causing a great risk of
> cristallizing and creating bad joints. 40/60 solder was common in the
> 70's, and is only slightly better. Today the most common (unless RoHS
> compliance is needed) is 60/40, which works fine, and the eutectic (no
> paste phase) 63/37 solder is easily available too.
>
> - Old solder can be heavily oxydized or sulfated, and when reflowed,
> these oxides and sulfates mix with it.
>
> - Whenever solder is heated, some of its tin evaporates. Even if it
> originally was 60/40, after melting it a few times it can be 50/50 or
> even worse.
>
> - Soldering consists in creating a surface alloy between the solder and
> the base material. The longer the solder is liquid, the hotter it gets,
> and the more times it is heated, the more base metal it dissolves. So
> the solder can end up containing enough of the base metal to give it
> poor characteristics.
>
> A very common cause of failure in modern equipment is that a
> surface-mount device develops an intermittent connection to the board,
> because the solder under one or several of its pins cracked. Typically
> the microscopic arcing and the localized heating inside the crack will
> cause oxidation. If such a damaged joint is simply fluxed and heated,
> the solder will beautifully re-arrange itself, one half on the pin, the
> other half on the board pad. And between them there will still be the
> oxide layer! The joint will look nice and shiny, but the intermittent
> connection is still there! Even close inspection with a strong magnifier
> often doesn't reveal the problem, but a simple ohmmeter does. Another
> trick is to push down the suspect pins with a rubber stick while the
> circuit is powered on, and see if it starts working.
> In these cases, I do the radical solution: Completely unsolder and
> remove the part, flush the pads with fresh fluxed solder, wick it off,
> check with the magnifier to make sure the solder properly wetted ALL of
> the pads, then do the same to the component, then reinstall.
>
> So, yes, I agree, old solder should be removed and replaced, not simply
> reflowed, whenever a joint is remade. Specially if it looks even
> slightly dull.
>
> Manfred
>