Operator performance predictability in electronic circuit board inspection /
| Main Author: | |
|---|---|
| Other Authors: | , , |
| Format: | Thesis Book |
| Language: | English |
| Published: |
1990.
|
| Subjects: | |
| Online Access: | ProQuest, Abstract Link to OAKTrust copy |
| Abstract: | The printed circuit board (PCB) inspection task is a complex visual task requiring the search of many solder points and the identification of the rare defect. The criteria used in the industrial environment are complex and include approximately 160 potential types of visual defects. The hardest to find are solder defects. Persons who inspect PCBs generally require several months experience before they become minimally proficient at inspection and several years before they are considered skilled. The problem addressed by the author is to determine if a visual simulation can be used to predict whether or not a person will become a successful inspector. This has been accomplished by designing an experiment and comparing its results to the prior work experience of 10 currently employed circuit board inspectors and 3 persons who have never been employed in the field. The experiment consisted of a simulated work center where each volunteer inspector viewed a sheet or a video screen, searching for pre-encoded defects and marking the defect location on a clear plastic overlay. Three controlled paces were used. Six series of 20 images were shown; the first contained photos of PCBs, the second had CRT images of the PCBs, the third had CRT images with digitized defects, the fourth contained CRT images of PCB pad drawings, the fifth had PCB pad drawings, and the sixth had abstract pad drawings. The facilities of the Automated Visual Inspection Lab (AVI) of the Industrial Engineering Department at Texas A&M University were utilized to develop CRT presentations. They were made by digitizing color video camera images (560 x 400 pixels) with a Targa Card. The data was analyzed with both ANOVA and nonparametric statistics because of the anticipated lack of normality and high zero content. Significant differences occurred for pace and image series. Two image series were selected as the most appropriate of which the inspectors ranked. These rankings were tested against those of 6-month job performance measures of quality, output, attendance, supervisory evaluation, and test monitor evaluation. The rank order was significant. A ranking of neither the job performance measures nor the experimental performance against an aptitude pretest proved significant. Recommendations for validating the test images are included in a section on test development. |
|---|---|
| Item Description: | Typescript (photocopy). Vita. "Major subject: Industrial engineering." |
| Physical Description: | x, 154 leaves : illustrations ; 29 cm |
| Bibliography: | Includes bibliographical references. |