<![CDATA[A multi-stage discrete manufacturing line performs only as well as its constraining workstation, yet many plants still locate that constraint through periodic manual reports that arrive too late to guide daily action. This study analyzes 60 days of field operational data from a five-station precision automotive-component line, comprising 600 daily-shift overall-equipment-effectiveness (OEE) records, 120 machine-breakdown events, a 45-element cycle-time study, and six months of throughput records before and after a real-time diagnosis practice. The objective is to demonstrate a reproducible diagnostic procedure that decomposes OEE into availability, performance, and quality, locates the constraint, and separates reliability losses from inherent process time. The welding workstation is the persistent constraint, with a mean OEE of 56.54% against roughly 83% at the other four stations, and it concentrates 91 of 120 breakdown events and 91.5% of total downtime. The standard-time study places welding within the 147.9-168.8 second designed-workload range, while its observed cycle time and breakdown burden are disproportionate, which is consistent with a reliability-driven constraint. Between the conventional and real-time periods, mean line OEE rose from 63.62% to 82.12%, scrap rate fell from 3.99% to 1.23%, and good output increased from 24,666 to 32,279 units per month. Because the two periods span different calendar months without a controlled trial, the results are reported as descriptive associations rather than causal effects. The contribution is a transparent bottleneck-diagnosis and review structure that links OEE decomposition, downtime concentration, and throughput monitoring for shop-floor decision making]]>
