Bose SW

Subject: Satellite cube loudspeaker mount insert
Description: The satellite cube loudspeaker array, present in many Bose home audio products, was failing at the end-customer level because the mounting brass insert was separating from the cube during installation; Bose management estimated the loss at $2 million dollars per year. The insert was being assembled by an injection moulding subcontractor, and Bose claimed the assembly process after the subcontractor's failure to solve the problem in a 2 year period. At the start of this project, the insert was separating at a torque between 10-15 ft-lbs; the minimum torque-to-failure specified was set at 30 ft-lbs.
Experiment design: Only 3 factors were identified, and test specimens were plentiful, thus a 3-factor 2-level full factorial experiment with 4 replicates was designed. The experimental trials were conducted in 3 work days.
Results: From the ANOVA, it was determined that only 2 factors were significant and the explanation was that the insert was not being heated sufficiently to bond to the body of the plastic cube.
Steepest ascent: Moving the process along the path of steepest ascent, torque-to-failure was maximized to the 35-40 ft-lbs range; however, the process' cycle time was exceeding production specifications.
Process upgrade: A higher capacity heater was installed at a $2,000 dollar investment.
2nd experiment design: A 2 factor 2-level full factorial experiment with 2 replicates was conducted to characterize the improved process and generate the 2nd steepest ascent path. This experiment was conducted in one work day.
2nd steepest ascent: Moving the process along this second steepest ascent path, torque-to-failure was optimized at the 70-75 ft-lbs range with a 4 sec. cycle time.
Conclusion: Torque-to-failure was improved from 15 ft-lbs to 70 ft-lbs. Problem was solved in two work weeks.
Estimated project benefit:
$2 million dollars from field failure product returns elimination.
Honeywell IAC

Subject: Wave solder process quality yield
Description: The Standard Products Factory wave solder process served four independent production lines manufacturing high-value process-control electronic assemblies (PWAs). Before this project, the process was operating at a poor quality yield level requiring extensive post-soldering touch-up and with a significant production loss.
Experiment design: 5 factors were identified; test specimens and process resources were limited due to cost and production constraints; in view of these conditions, a 5-factor 2-level fractional factorial experiment with 1 replicate was designed.
Results: From the effects analysis, it was determined that only 2 out of 3 process factors were significant, pointing to an equipment capability weakness.
Steepest ascent: Moving the process along the steepest ascent path, fluxing related defects, in particular solder balls, were significantly reduced; however, the process' pre-heating rate was exceeding the specifications of some electronic components, raising the concern of production and materials engineering.
Process upgrade: With these experiment results and cognizant about the technology requirements, the best wave solder equipment, with strict design requirements, was researched and procured. The new process was configured with the aid of a temperature profiling device to ensure that the PWAs were exposed to the appropriate volume of heat without exceeding electronic component specifications.
Conclusion: Process defects were eliminated and material defects became easier to identify. The higher quality yields facilitated the removal of the touch-up workstations, enabling the factory to expand productive operations without acquiring additional floor space.
Estimated project benefit:
$600,000 dollars per year from scrap elimination and touch-up costs.
Designed experiments software