Printed Circuit Board Failures – Causes and Cures

Estimated reading time: 5 minutes

Outgassing from via hole during testing with oil.

His surveys on behalf of IPC and NPL had indicated that shorts and open circuits accounted for 27% of PCB problems, and via hole failures for 8%. But solderability was by far the principal cause of failure at 39%. He listed a whole series of bare-board solderable finishes: hot-air solder level, immersion silver, immersion tin, organic solderability protective and electroless nickel immersion gold, and reviewed methods for thickness measurement and solderability determination, as well as showing examples of phenomena such as nickel-footing and solder wicking. Whereas the wetting balance gave the most scientific measurement of solderability, it tended to be found mainly in specialist test laboratories. Willis advocated a simple dot-coalescence test as a practical alternative, for which a small test-coupon could straightforwardly be incorporated into the PCB panel design. In the investigation of “black pad” failures, he demonstrated what could be revealed by scanning electron microscopy of electroless nickel surfaces after chemical removal of the gold layer from an ENIG finish.

Short circuits or low-resistance current leakage paths could be formed in service by electrochemical migration of copper, and Willis showed examples of dendritic growth across the PCB surface and conductive anodic filament (CAF) growth from the barrels of plated-through holes along glass-fibre bundles. It was critical that assemblies were free from ionic contamination after the soldering and post-soldering cleaning processes, particularly if the work was to be conformally coated, and Willis described techniques for measuring ionic contamination and interpreting the results.

Two particular defects had been associated with the introduction of lead-free soldering: copper dissolution and pad cratering. In the former case, if there was no electroless nickel barrier layer, there was a tendency for copper to be dissolved by the solder, particularly if the solder had a high silver content, in some cases drastically reducing the effective thickness of surface-mount pads, and in the latter case for a cohesive failure to occur in the resin under the pad as a result of the increased brittleness of laminates formulated to withstand lead-free processing temperatures.

Finally, Willis referred webinar attendees to the NPL Process Defects Database, http://defectsdatabase.npl.co.uk, a searchable on-line compilation of actual soldering process defects. For each defect entry, there was a link to a relevant report where appropriate, with free access to over 100 NPL reports on various aspects of reliability characterisation.

There was great interest in this webinar. An enormous amount of relevant information was shared, and the question and answer session ran on well over the nominal finish time in response to attendees’ specific queries. A great credit to SMART Group, whose aim has always been to promote the advancement of the electronics manufacturing industry through the education, training and notification of its members in surface mount and related assembly technologies, and to Bob Willis himself for generously sharing his know-how.

• < Previous Page