I couldn't agree more with Tom. We have established via reliability testing
(air-to-air thermocycling) that solder plugged holes are MORE reliable than
the unfilled yet solder coated holes.

Dave Rooke
Circo Craft

>
>               As I read some of the comments regarding the open vias
>          problem, I think there is some confusion which I'd like to
>          clear up.  Regarding tolerancing microvias with no bottom
>          tolerance (i.e., they may plug shut), the plugging generally
>          occurs during the solder levelling operation, not plating.
>          Board fabricators must walk the tightrope between keeping
>          holes open and not blowing too much solder off of surface
>          mount pads when adjusting air knife placement and pressure.
>          Another issue is using a small enogh drill diameter so that
>          hole breakout is minimized.  Board fabricators generally
>          like some slack with vias in using a smaller drill size than
>          the nominal callout of the via hole diameter in order to
>          keep the hole in the pad.  Aspect ratios (bd. thickness to
>          dill size) of 7:1 or greater are not uncommon in volume
>          environments these days.  My company has just installed new
>          PAL plating lines in our volume shops to meet such
>          challenges.  The PAL folks tell us that we should be able to
>          plate boards with aspect ratios of 10:1!
>               The fallout rate experienced by the original author
>          strikes me as unusually high.  One possible way that they
>          might have escaped the E.T. net is if they were intermittent
>          opens which may show up only when the board is flexed or
>          stressed.  Gary Ferrari's comments about supplying and
>          requesting IPC-D-356 net lists should be well taken.  This
>          is an extra measure which should guarantee 100% electrical
>          test on the part of the board fabricator.  At the very
>          least, it assures that the design intent and the Gerber
>          supplied match.
>               The bottom line is that I'd like to dispel any notion
>          that allowing vias to plug shut causes a less reliable
>          connect.  Hole sections should be done by the board
>          fabricator to get to the root of the real problem.
>          Regards,
>          Tom Coyle
>          Field Services Engineer
>          HADCO Corporation
>          [log in to unmask]
>
>
>
D. Rooke
([log in to unmask])