Content area
Full Text
Introduction
In the mid-1970s Bell Laboratory researchers were concerned about potential failures of printed wiring boards (PWBS) intended for high voltage switching applications. [1] Boddy et al. (1976) and [2] Delaney and Lahti (1976) reported on accelerated life testing of flexible PWBs coated with UV cured resin, and identified two new failure modes:
conductive bridges between conductors on the surface; and
conductive shorts through the substrate.
At the same time, [3] Der Marderosian (1976) at Raytheon used accelerated conditions to study measling, crazing and delamination in multilayer boards and reported a failure he termed "punch thru." Both of these research groups independently discovered an unexpected failure mode associated with the growth from the anode of a conductive filament, which today is known as CAF. This paper will highlight the history of CAF.
AT&T Bell Labs' test vehicle (Figure 1 [Figure omitted. See Article Image.]) was a flexible epoxy-glass PWB 0.005-0.007 in. thick with comb patterns of 0.008 in. lines and 0.009 in. spaces. Some combs were biased on the surface and some were biased through the substrate. Processed boards, coated with conformal coating were tested from 35 to 95°C, 25 to 95 percent RH and DC voltages up to 400 V. For accelerated testing at 85°C, 80 percent RH, 78 V bias, failures occurred within 2-5 days.
They identified two major failure modes that they described as causing "catastrophic loss of insulation resistance due to the formation of conductive bridges between conductors." The first failure mode - through substrate shorts - only occurred above 75°C and 85 percent relative humidity, and thus was not considered to be a problem at use conditions (Figure 2 [Figure omitted. See Article Image.]). The second mode involved shorts between conductors on the same side of the board in which conductive material accumulated between the glass bundles and the epoxy (Figure 3 [Figure omitted. See Article Image.]). [2] Delaney and Lahti (1976) noted that the thicker the buttercoat, the more this failure was reduced. They also observed a failure, which they termed - anodic eruption failure mode - in which corrosion products emerged from the anode to the covercoat surface, charring the surface, and then growing back through the covercoat to the cathode where it shorted (Figure 4 [Figure omitted....