Alternate Title

In-Line Quality Assurance for Switch Manufacturing

Contributor(s)

Zeid, Ibrahim, Messac, Achille

Abstract

Because of an undetected problem in Cole Hersee switch manufacturing process, 1,400 defective switches were shipped to customers in late 1999. These switches underwent and passed visual inspection prior to shipment, but the test was not adequate to identify the defective switches. The NEU Capstone design team was asked to design an effective inspection test for the switches. Specifically, the team was to create a fixture for in-line random testing of switches, to establish control limits for the continual monitoring of the process using the testing device, and then to determine functional user limits (ie: the maximum recommended force that the customer can apply to the switch) based on data from the testing. The team designed a small, manual data acquisition system (DAS) to conduct destructive testing of switches. The DAS applies an increasing force to the switch bushing until the pushing shears from the casing, mimicking the primary failure mode of the switches in application. A force gauge measures the force at which the joint fails. Data from a sample of sixty non-defective switches showed that the strength of the switch can be modeled by a normal distribution with a mean of 130 pounds and a standard deviation of 26.5 pounds. Based on this analysis, the upper and lower control limits of 165.55 and 94.45 were determined. Switch failures outside these control limits invoke Cole Hersee quality control procedures. In addition to the DAS, the team developed tables of switch reliability as a function of force, from which the functional user limits are set in an effort to minimize failure due to misuse and improve customer satisfaction.

Notes

Capstone Design Course, Spring 2000

Keywords

switches, manual data acquisition systems (DAS)

Publication Date

2007

Permanent URL

http://hdl.handle.net/2047/d10010610


View Full Text

Click button above to open, or right-click to save.

Share

COinS