Expediting in a Contract Electronics Assembly Line
Northbridge Electronics assembles control boards for industrial pumps on a single SMT line feeding manual through-hole assembly and final test. Sales quotes six working days, yet order-to-ship average is 18 days and on-time delivery has fallen to 82%. There are 3,200 boards in process, including 1,050 waiting at final test and 600 in a rework queue; value-added time is only 11 minutes. Daily demand is 900 boards and final test capacity is about 1,200 boards per day with two testers. First-pass in-circuit test yield is 62%, and 60% of failures trace to solder bridges on one 40-pin connector. The SMT supervisor wants a second automated optical inspection machine costing $310,000, arguing that inspection is the bottleneck. Changeover is 90 minutes and the line runs weekly batches of 250 boards. Two expediters spend most of their day chasing hot orders. The plant manager has asked for a waste-reduction plan that avoids major capital and can show results within 30 days.
1. Which lean waste is the primary driver of the 18-day lead time, and what evidence argues against the AOI purchase? Use the data to support your answer.
Read the worked answer
The dominant waste is waiting/inventory created by batching and a hidden rework loop, with defects as the upstream driver. Final test has roughly 1,200 boards/day of capacity against 900 demand, or 75% utilization, so it is not the constraint. Buying AOI adds inspection capacity and overprocessing without removing the 38% first-pass failure rate.
Little's Law says lead time equals WIP divided by throughput. At 900 boards/day, 3,200 units is about 3.6 days of throughput, not 18 days. The extra time is queueing and rework, not a simple capacity shortage. The 600-board rework queue and 62% first-pass yield show that defective boards are recycled through the line, inflating WIP and confusing the schedule. The 40-pin connector is a concentrated defect location, so targeted root-cause work at printing, placement, or reflow has higher leverage than AOI. AOI could later serve as a detection control, but first stabilize flow and reduce the defect source.
2. Sequence a 30-day countermeasure plan using Lean tools. Explain why you sequence it this way.
Read the worked answer
Sequence by impact and speed:
- Go to the line, map the current-state value stream, and tag every queue. Separate the rework loop from the main flow with a FIFO lane and a visible WIP cap, for example 300 boards at final test.
- Charter a focused 5-why/fishbone at the 40-pin connector. Check stencil aperture, paste volume, placement pressure, and reflow profile. Run a small paired comparison or DOE rather than a broad campaign.
- Attack changeover with SMED. Reduce batch size from 250 to 100 only after changeover is under about 45 minutes; otherwise smaller batches create more changeover loss.
- Establish a daily tiered huddle with board-level metrics: WIP, rework age, first-pass yield, and expedites.
- Defer AOI until the defect rate is stable, then evaluate it as a detection control.
This order is deliberate. Batching and rework create the long lead time. Reducing batch size before changeover improvement would hurt output, and AOI would not fix the defect source.
3. What measures and simple experiment would show whether the plan is working?
Read the worked answer
Track lead time from order release to ship, total WIP, final-test queue, rework age, first-pass in-circuit test yield, expedites per day, and on-time delivery. Use a run chart or I-MR chart rather than a before-and-after average.
To avoid confounding, pilot on one product family and compare it with a similar family that is not changed. Standardize data definitions and record whether the WIP cap is actually followed. Predefine success, for example: first-pass yield up 10 points, WIP down 30%, and lead time down 25% within 30 days. Also check that expedites and rework age fall. If on-time delivery improves only because expediters are working harder, the system has not improved.