Lesson 23 designs what the system does when it breaks. The one thing every student must leave with: every failure mode needs a designed destination and a designed recovery, not a dead stop and a shrug. If they walk out thinking a jam or a WMS outage is something you handle at runtime instead of design for, the lesson didn't land. This is the depth part, so teach the mechanisms, don't slogan them.
| Segment | Min | What happens |
|---|---|---|
| Hook: it will jam | 6 | Open on the hook cold: everything so far designed the system for the day it works, and this designs it for the day it doesn't. List the failures aloud, jam, lane-full, no-read, WMS drop, and land the frame: none of those are exotic, they're Tuesday. Ask who has cleared a jam at 2 AM. |
| Jam, lane-full, destination | 13 | Walk jam detection (photoeye blocked past its window, stop the zone, alert) and lane-full (entry photoeye blocked, no divert). Then the key move: put the Lesson 22 failure taxonomy on the board and make the room give every entry a destination. The unclaimed ones all land at the hospital lane. |
| Anti-gridlock, both halves | 12 | Teach gridlock on a loop sorter, then the induction-tracks-clearance principle. Push the mechanical half hard: no runway in the layout, no throttle in time. Then the line-sorter nuance, a non-recirculating sorter can't loop-gridlock, so the principle moves to the merge. |
| Size the hospital lane (live calc) | 13 | Run the exception-volume method on the board as a live calculation: throughput times exception rate equals pieces per hour the lane must clear. Watch how many students reach for total system volume or forget the lane entirely. Turn it into a number, not a box. |
| Degraded mode, restart, alarms, HMI | 12 | Run the degraded-mode drill: "the WMS just went dark mid-wave, what does your system do right now." Push until they commit to a design decision. Then restart after e-stop (no auto-restart, located reset, safe sequence), alarm rationalization (root cause, not fifty symptoms), and the HMI as the operator's path from alarm to fix. |
| Riverside plan and Part V wrap | 4 | Students commit the Riverside recovery-and-exception plan to their note. Close the controls arc: the three Part V deliverables are now complete. |
| Total | 60 | Baseline session. Expand with the stretch options below if you have 90 minutes. |
Two moves, both drawn from where students consistently stumble. The hospital-lane omission: ask the room to size the exception path for a system. Watch how many size it to total volume, or forget it entirely. The hospital lane gets drawn as a box and never sized, because during initial sizing it never seems to matter. Walk the exception-volume method until it's a number on the board, not a placeholder in a drawing.
The degraded-mode drill. Say it flat: "the WMS just went dark mid-wave, what does your system do right now." Most rooms freeze or say stop the line. Push until they've made a design decision, default-route or hold, instead of a runtime surrender. Tie both moves back to the merge-jam diagnosis habit from Part IV: a stopped line usually has a designed answer that nobody designed.
Every one of these traces back to the same failure, treating recovery as something you improvise at runtime instead of design. Drive each back to: every failure needs a designed destination and a designed recovery.
Build the recovery-and-exception plan as the session's spine. Make students run the hospital-lane sizing at Riverside's 20 CPM target as a live number: 1,200 cartons an hour means even a low single-digit exception rate puts dozens an hour into the lane. Keep the misdirect rate honest, about three percent is a real Riverside anchor for the exception conversation, not a stand-in for the no-read rate.
Make them defend the anti-gridlock nuance: Riverside's three-door line sorter can't loop-gridlock, so the principle moves to the merge feeding the sorter, throttle induction there when a door lane fills. Then force a commitment on the degraded-mode behavior and the forklift-crossing restart. Don't let anyone leave the beat with the exception path still a box.
Question 1 (hospital-lane sizing). The lane should have been sized against expected exception volume, throughput times the exception rate (no-read plus reject criteria), so it clears exceptions at least as fast as they arrive. Total system volume is the wrong target the other way: the lane doesn't handle every carton, only the exceptions, so the driver is the exception rate, not the system rate. In the scenario it was sized against nothing, drawn as a box, so it backed up into a secondary jam. Strong answers name the two rates that drive it and note that a backed-up hospital lane stops the sorter.
Question 2 (lane-full plus WMS drop). Full lane: the item can't divert there, so it needs a designed alternative, a backup lane, recirculation, or the hospital lane, and the induction should throttle so product isn't piled into a sorter that can't clear it (the transferable anti-gridlock principle). WMS drop: a designed degraded mode, default-route or hold at accumulation, not a line stop. What had to be designed ahead of time: in the controls, the throttle logic and the degraded-mode behavior; in the layout, the buffer/runway that gives the throttle time to act. Strong answers separate the controls decision from the mechanical requirement and refuse to answer either event with "stop the line."
This lesson ends the controls arc. Note for yourself, not the room, that the three Part V deliverables are now complete: the controls architecture summary, the setpoints list, and the interface map. They feed forward, validation in Part VI, the drawing in Lesson 28, and the capstone. Recovery is the piece that decides whether the intelligence they built across five parts holds up on a bad shift.
This is the final Part V beat, and it's where the project-note easter egg pays off. A student who kept every Riverside note from Part I forward now holds most of a capstone controls package: the topology placement, the sensing-and-ID plan, the machine-controls and Aux I/O list, the setpoints, the power-and-network note, the interface map and latency worksheet, and now the recovery-and-exception plan. That's a near-complete controls architecture, and most of them don't realize it yet.
Never announce it. The reveal belongs to them at the capstone, and it lands hardest for the students who kept their notes without being told why. Your only job across five parts was to make the habit feel normal. If they built it, the payoff is theirs to discover.