| The objective of this research was measuring bit temperatures with an infrared pyrometer while drilling printed circuit board stacks. Cited research indicated drilling temperatures are related to many plated-through-hole defects. This investigation was divided into three phases: developing a suitable on-line measurement system, documenting sensor responses, and evaluating the sensitivity of temperature to a comprehensive set of drilling factors.; Six data sets were generated. Four documented sensor responses and sources of measurement error including; peak entry and exit temperatures for a 2500-hit drill run at constant speed and feed, time traces from individual drill cycles for temperature and a complex pressure foot-thrust force, bit emissivity measurements, and data demonstrating the significant effect of IR probe position on exit temperature.; Sensitivity of temperature to several drilling factors was evaluated in two designed experiments for 18-mil and 42-mil solid carbide designs drilled in FR-4 multilayer panels. Factors included: surface speed, chip load, head height, number of panels, hits, and localized differences in layer geometry. Measured responses included: thrust force, bit temperature, and changes in bit relief areas. A unique blocking methodology was used in these designed experiments. Balanced factor combinations were distributed in small blocks through the linear response region of each bit.; All bits exhibited dynamic temperature cycling with cooling between holes exceeding 70{dollar}spcirc{dollar}C. Measured emissivity changes were minimal. Bits of different design/diameter exhibited different responses. 42-mil designs showed large nonlinear increases in temperature during break-in followed by linearly increasing response in later hits. Within 200-hit blocks 42-mil response was sometimes nonlinearly increasing or decreasing, and exhibited positive and negative changes in slope. 18-mil designs produced relatively linear responses with slightly negative overall slopes. In general, temperatures were affected most by bit design/diameter, stack, height, percent copper layers, chip load, and speed. Temperature was affected least by head height, variations among bits of the same design, and higher order interactions. |
