Our Process
     Isolated tabs in the center of the board to be gold plated are defined 
     by a primary layer of resist.  After copper and tin-lead plating them, 
     we vacuum laminate 3 mil dry film resist.(over the plated tin-lead and 
     over the primary layer or resist)  This top layer of resist is 
     selectively imaged to produce an opening exposing the area to be gold 
     plated (which now has tin-lead on it). Then the tin-lead is stripped, 
     and the nickel and gold are plated using our usual sulfamate 
     nickel/soft gold process. 
     
     The Problem
     The plated nickel peels from the copper surface.  The peeling is 
     always near traces connecting to the tabs.
     Therefore I suspect that tin-lead stripper is being trapped under the 
     top layer of resist and coming out during plating. 
     
     Solutions?
     I don't suspect the nickel/gold process is the problem, as the usual 
     yield on products without this tin-lead stripping step is nearly 100%. 
      
     We have no option to change the design of the board to facilitate 
     manufacturing. ie: No bussing of the tabs is possible therefore base 
     copper must be used as a conductive path prior to etching.
     
     We've tried: 140 degree F water rinse after stripping, conveyorized 
     nitric tin-lead stripper, immersion and conveyorized peroxide tin-lead 
     strip, a second strip cycle, UV bump (160mJ), among other things.
     
     Any help would be appreciated.