The Bottleneck That Is Only Half Busy
A frame cell has cut (2 saws), fit (3 fitters), weld (4 stations), and paint. Demand is 120 frames per week; output is 95. Utilization: weld 98%, fit 70%, cut 60%, paint 45%. Welders work 10-hour shifts, five days a week. A time study of a welder shift gives: 3.2 h arc, 1.8 h fit-up and tack, 1.2 h grinding, 1.1 h searching for parts, fixtures and tools, 0.7 h rework, 1.0 h breaks/meetings/housekeeping, 1.0 h travel and repositioning. The plant manager has approved a $60,000 fifth welding station plus roughly $85,000/yr in wages and fume control, and wants it ordered Monday to "break the bottleneck."
1. Is the weld bay the constraint, and does the data justify the fifth station?
Read the worked answer
Constraint: yes. 98% utilization at weld versus 45-70% everywhere else is the classic signature; the other three areas have slack.
Fifth station: not yet. Capacity is 4 stations × 50 h = 200 h/wk; at 98% that is 196 h for 95 frames = 2.06 h of weld-bay content per frame. 120 frames needs 247 h - a 51.6 h gap that a fifth station (50 h) closes only on paper.
But look inside the shift: only 3.2 of 10 hours is arc time (32%). The other 68% is fit-up, grinding, searching, rework, breaks, and travel. If the 1.1 h of searching dropped to 0.4 h and one hour of grinding were eliminated, content falls to 1.76 h/frame and capacity rises to about 111 frames/wk with no capital at all.
2. Sequence the corrective actions and justify them on cost.
Read the worked answer
Cheapest first, because each one also tells you whether the capital is needed.
- Kit and stage parts per frame - a supermarket and a staging lane, roughly $10-15k one-time, removes most of the 1.1 h search. This is waiting waste, not a welding problem.
- Attack the 1.2 h grinding - grinding is almost always downstream of fit-up variation and oversized welds. Feed gap data back to the fitters and set a weld size standard so nobody builds to maximum leg.
- Standard work and 5S at the station to recover travel and searching time.
Only then price the fifth station: about $60k capital plus $85k/yr recurring, against maybe $20k of one-time spend that captures a third of the gap. Also check whether the 120 figure is level demand - if it is spiky, a buffer plus level scheduling may hit 120 with no capacity change at all.
3. How do you prove the fix worked before releasing the capital?
Read the worked answer
Define the numbers in advance so success isn't decided by anecdote.
Target: to hit 120 frames in 196 station-hours, content must be 1.63 h/frame, which means arc-on of about 40% (0.66 h arc per frame).
Method: measure arc-on percentage and frames out per week for three to four weeks against that target. If arc-on moves from 32% toward 40% and output tracks to 118-120, defer or cancel the purchase.
If it stalls, you now have a quantified, documented case for the capital - and you will know how many stations, because you will know the real content per frame rather than the one distorted by waiting and grinding. Set the decision date at the start, or the purchase will happen Monday anyway.