Setups and resource conflicts

A line can be busy even when it makes no bottles. Northstar must change the label state when it moves from blue to red work. That change takes time on the same line. If we leave it out, our plan will look fast and fail on the floor. In this module we place setup time where it belongs. We then add a shared person to see why checking each machine alone is not enough to prove a plan works. The line starts in blue. A is blue and takes sixty minutes. B is red and takes forty. C is blue and takes fifty. Each color change takes thirty minutes. A same color move takes no setup in this case. These are the rules for this small model. A real line might use a different time in each direction, but we will not add a rule that is not in the data. Always name the starting state before you count changes. Run A, then B, then C. A ends at sixty. Change to red from sixty to ninety, run B to one hundred thirty, then change back to blue until one hundred sixty. C ends at two hundred ten. Run time totals one hundred fifty and setup totals sixty. The sum is two hundred ten, which matches the last finish. That is a useful check. If those totals did not match, we would look for a hidden gap or a missing block. Now group the blue jobs. A ends at sixty and C at one hundred ten. Change to red until one hundred forty, then run B to one hundred eighty. We saved thirty setup minutes and finished the whole set sooner. But B is due at one hundred fifty, so it is thirty minutes late. The first plan met every due time. Grouping families can save time and still harm service. We must show both results to the person who owns that promise. A second cell adds a new kind of clash. D runs on machine one and E on machine two. Each needs a thirty minute setup by the only technician. D then runs sixty minutes and E forty. The runs do not need that person. If D is set up first, its setup ends at thirty and its run ends at ninety. E is set up from thirty to sixty and runs until one hundred. Both jobs meet their due times in this order. What if both setups were drawn from zero to thirty? The machines are different, so their own lines look fine. But the one technician would be in two places at once. That plan is not legal. Put each setup on the person line as well as its machine line. It is still one task, with two needs. Do not double its duration just because it appears on two resource views. Check the overlap on every view that the task requires. Try the other technician order as well. E is set up first and finishes its run at seventy. D is set up next and finishes at one hundred twenty. Both are on time, but the last finish is later than in the D first plan. A total of sixty setup minutes fits either way. The total alone cannot tell us which order is good. We need the placement of those minutes and the lengths of the runs that follow them. In the lab, draw setup blocks, run blocks, and closed time. Audit one resource at a time, then check each job route. Compare the two choices before you write a recommendation. Use the stated service goal and do not turn saved time into invented cost savings. No labor rates were given. A sound memo says how much setup was saved, which job moved, and whether its due time still holds. That is enough to make the tradeoff clear and keep the plan honest. Use the initial state as an input row, not a note you hope everyone recalls. If the line starts red next week, the first change may be different even with the same three jobs. Test that case with the same care as a new due date. A setup matrix is part of the model, just like a run time. When you change it, count the actual moves again. Do not copy the old total and assume that fewer color groups always means the same saving.

1. Count each change

   Original static calculation card for count each change, using the Northstar numbers stated in the narration; show labeled units and no decorative charts.

   • Initial state Blue
   • Two changes for A-B-C
   • Setup total60
2. Compare family groups

   Original static timeline for compare family groups, using the Northstar numbers stated in the narration; show labeled units and no decorative charts.

   • A-C-B saves30
   • Finish180
   • B completes late
3. Trace due effects

   Original static calculation card for trace due effects, using the Northstar numbers stated in the narration; show labeled units and no decorative charts.

   • A-B-C meets dues
   • A-C-B makes B30 late
   • No universal winner
4. Add the technician

   Original static timeline for add the technician, using the Northstar numbers stated in the narration; show labeled units and no decorative charts.

   • D and E share one person
   • Each setup lasts30
   • Runs need no technician
5. Audit every line

   Original static timeline for audit every line, using the Northstar numbers stated in the narration; show labeled units and no decorative charts.

   • D-first ends100
   • E-first ends120
   • Simultaneous setups illegal
6. State the tradeoff

   Original static calculation card for state the tradeoff, using the Northstar numbers stated in the narration; show labeled units and no decorative charts.

   • Check machine and person
   • Count one activity once
   • Report service and time