Corrosion(?) on solder joints

[tinska]

Greetings,

In our subcontractor manual soldering line we found out random units (maybe 0.1%) that would produce a "phantom" resistance on powerwire-to-PCB joint after few months. By phantom resistance, I mean we see voltage dips when ICs draw current(~20mA, wires are thick) from battery, since U=RI. Then we found out that if we re-melt solder joint the problem was gone.

From visual inspection there seems to be nothing wrong. No cold solder lookalikes. One unit had corrosion stain on the joint, others don't. If you measure with multimeter it says 0.1 - 1Ohm which should be normal. Solder wire is some Chinese brand with clean-free-flux. Solder supplier says that this solder is used by many companies without problems and flux should be left on the PCB. However I wonder if this sounds like a flux problem or solder process problem or PCB surface problem. Solder is RoHS, wire is tin plated copper, PCB is FR-4 HASL RoHS.

All ideas and opinions are very appreciated. Thanks

 

[mcgyvr]

Its this a direct wire soldered to a PCB connection or is there a connector system of sorts..
PTH (plated thru hole) barrel cracks?
% of solder fill in PTH?

We had a few failures once from insufficient thru hole plating during PCB fabrication and inadequate solder fill during soldering.. Simply applying more solder would cause the issue to go away..boards were still scrapped