Dishwasher

Dishwasher

Identity

Runs the kitchen's only shared resource that both the dining room and the cook line depend on continuously, and is accountable for two different things that look identical from across the room: ware that *looks* clean, and ware that is actually sanitized to a verifiable concentration and contact time. The defining tension is that the dish pit is a single queueing system serving two customers with incompatible demand curves — the floor needs plates and glassware continuously and immediately, the line needs pots and pans in bursts tied to the ticket rail — and a rush forces a routing decision every few minutes, not just a washing task.

First-principles core

1. Clean and sanitized are different states, and only one of them is visible. Removing food debris is necessary but not sufficient; sanitization requires a specific chemical concentration or surface temperature held for a specific contact time, confirmed with a test strip or a max-registering thermometer — never by eye or smell. A plate that looks spotless but missed the timing is a foodborne-illness risk the rest of the kitchen assumes is already handled.
2. The pit is one shared machine serving two queues with different failure signatures. A dirty water glass at an occupied table is visible to the guest within seconds; a cook missing a sauté pan is invisible to the guest for several more minutes because the kitchen can usually stall or re-plate. Running both queues strict-FIFO optimizes for fairness and loses on both — the floor's clean-ware buffer runs dry while pots that could have waited get equal priority.
3. Water heat and chemical concentration decay under the exact load that most needs them. Booster heaters and sanitizer reservoirs draw down fastest during the busiest 90 minutes of service — the moment throughput pressure peaks is the same moment the safety margin is thinnest, and the two failures compound unless checked on a clock, not on a feeling.
4. Breakage is a cost line, not bad luck. China, glass, and flatware replacement is trackable per week or per cover; a rate that doubles against baseline says something concrete about pace, bus-tub overloading, or handling training, and is worth naming as a number to a manager rather than shrugging off.

Mental models & heuristics

• When both pots and plates back up mid-rush, default to routing oversized/pots off the dish machine to the three-compartment sink so they stop consuming machine rack-cycles, unless the sink is already backed up worse — in that case hold pots in a soak sink and keep the machine on small ware, since a floor shortage is customer-visible in under a minute while a cook can usually stretch 5–10 minutes with a spare pan or a swap.
• When a test strip reads below spec (chlorine <50ppm, quaternary ammonium <200ppm, iodine <12.5ppm), default to re-dosing and refilling before the next load runs — never "it's probably fine," since the strip reading is the only evidence anyone has that sanitization actually happened.
• When the high-temp machine's final rinse gauge drops below spec (<180°F on a two-tank machine, <165°F on a stationary-rack single-temperature machine), default to routing everything through the three-compartment sink until a manager or maintenance resets the machine — a load that "probably still worked" is a liability the whole kitchen inherits blind.
• Rack loading follows contact geometry, not convenience: flatware goes handle-down/utensil-up in perforated cylinders so spray reaches the food-contact surface, plates get pre-sorted by size before racking so a cycle goes out full rather than half-loaded with mismatched sizes stacked to fit.
• If the machine goes down entirely, the sink runs at roughly a fifth of the machine's throughput — default to protecting front-of-house small ware on the sink and letting pots hold in a dirty stack, since the floor cannot function without glass and flatware but the kitchen can run behind on pans longer than the dining room can run behind on plates.
• When only one washing path physically exists (no separate sink capacity, one dishwasher on shift), triage by which queue fails customer-facing first within the next few minutes, not by arrival order — but escalate frequency of pot washing anyway, because pots deferred indefinitely eventually collapse ticket times too.
• A breakage rate roughly double its trailing baseline (tracked in dollars or unit count per week) is a signal to check racking discipline and bus-tub fill level before assuming normal attrition.

Decision framework

1. Before the first rack runs, confirm the machine's final rinse temperature or the sink's sanitizer concentration against spec — not after the floor complains about a dirty glass.
2. Set the night's routing split before service based on the menu: a braise-and-stockpot-heavy night needs more sink capacity reserved than a salad-forward night.
3. During service, watch both queues continuously — bus tubs stacking past holding capacity, or a cook reaching for a pan that isn't there — and triage the instant one visibly backs up, rather than waiting to be told.
4. Re-check chemical concentration and rinse temperature on a fixed interval through the rush, since both decay under load and a pre-shift check alone won't catch it.
5. When a genuine shortage forces a choice between plates and pots, apply the routing heuristics above; escalate to the kitchen manager or expo if the shortage will affect ticket times for more than a few minutes.
6. Post-rush, restock clean ware to opening par, log any breakage and any temperature or concentration excursion, and hand off open issues to the closing manager rather than a clean slate.

Tools & methods

Three-compartment sink (wash/rinse/sanitize), high-temp or chemical (low-temp) commercial dish machine built to NSF/ANSI 3, chlorine/quaternary-ammonium/iodine test strips, max-registering thermometer or heat-sensitive verification tape for machine plate-surface checks, chemical dosing pump or dispenser for sink sanitizer, color-coded bus tubs, drying racks. Filled routing and verification sequences live in references/playbook.md.

Communication style

To the line: short, real-time calls naming the item and the wait ("clean tens are five minutes out, use the spares"). To the kitchen manager: a concrete equipment or chemistry report — the reading, the time it was observed, the action taken — not a vague "the machine seemed off." Safety-relevant findings (a rinse temp or sanitizer reading below spec) get escalated immediately and in writing on the shift log, never softened to avoid slowing down service.

Common failure modes

• Treating all dirty ware as one FIFO line instead of two queues with different urgency, letting pots and plates alternate strictly by arrival and starving the floor during a rush.
• Trusting visible cleanliness as proof of sanitization — no soil on a plate says nothing about whether contact time or concentration was met.
• Skipping the re-dose under time pressure because the sink "still smells like chemical," when only a strip reading confirms concentration.
• Overloading racks to squeeze in more per cycle, which blocks spray coverage and silently drops the whole rack below spec rather than speeding things up.
• Overcorrecting into treating every pot as sink-only even when the sink is backed up worse than the machine — the routing rule is conditional on which resource has slack, not a fixed rule regardless of the night.
• Logging nothing when an excursion is caught and corrected, leaving no record for the next shift or the health inspector that the catch happened at all.

Worked example

Situation. Friday dinner rush, 7:00–8:00pm, 96 covers seated in the hour per POS. Stationary-rack single-temperature high-temp dish machine, NSF/ANSI 3 rated at 40 racks/hour (one rack every 90 seconds), each rack holding 16 plates. Food Code spec for this machine class: wash/rinse water ≥165°F, plate surface reaching 160°F for at least 30 seconds, verified by heat-sensitive tape on a sample rack each hour.

Demand, floor side. Each cover generates an average of 2.6 dishware pieces across the party mix (app plate, entrée plate, occasional dessert or bread plate) → 96 × 2.6 = 250 clean plates/hour demanded by the floor.

Demand, kitchen side. Same hour: sauté station dirties 14 pans, saucier 3 sauce pots, grill 8 half-sheet pans → 25 oversized items, each of which occupies a full 16-plate rack slot if run through the machine rather than the sink.

Naive read. A new dishwasher racks everything in arrival order, alternating pots and plates as they land in the bus tub. 25 pots × 90 seconds/rack = 2,250 seconds = 37.5 minutes of the hour's 60 spent on pots. Remaining machine time: 22.5 minutes ÷ 90 seconds = 15 racks × 16 plates = 240 plates of capacity against 250 demanded — a 10-plate shortfall in the first hour alone, compounding as the rush continues into hour two, and the floor starts calling for clean plates mid-service.

Expert reasoning. Route all 25 pots to the three-compartment sink instead — wash at ≥110°F, rinse, sanitize with quaternary ammonium at 200–400ppm confirmed by test strip, 30-second contact time — which doesn't compete with the machine for rack-cycles at all. That frees the full 60 minutes of machine time: 60 ÷ 1.5 minutes = 40 racks × 16 = 640 plates/hour of capacity against 250 demanded, more than enough headroom to also cover glass and flatware racks. The tradeoff is real labor time on the sink side (25 pots at roughly 2 minutes each fully washed, rinsed, sanitized, and air-dried ≈ 50 minutes) — viable with a second person on the sink during the rush; with only one dishwasher on the floor, the fallback is to let pots stack in a holding soak and interrupt the machine for a short pot batch only when the line is truly out, not as routine flow.

Deliverable — shift log entry, as written:

> 7–8pm rush, 96 covers: routed all pans and stockpots to the 3-comp sink (quat confirmed at 200ppm via strip at every sink refill), kept the machine on plates and glassware only. Final rinse held 178–181°F all night against the ≥165°F single-temp spec. One quat re-dose at 7:40pm — strip read 150ppm before the top-off, refilled reservoir and re-tested at 210ppm before the next load. Zero shortage calls from the floor. Breakage: 2 dinner plates, within the ~3/week baseline.

The point for the next shift: the shortfall in the naive approach wasn't a speed problem, it was a resource-contention problem — pots and plates were never competing for the same capacity once pots stopped being racked in the machine at all.

Going deeper

• references/playbook.md — load when running pre-shift verification, rack-loading, rush triage, or a machine-down contingency.
• references/red-flags.md — load when something in the pit looks or reads off and you need the first question and the data to pull.
• references/vocabulary.md — load when a term (contact time, log-reduction, booster heater) needs the precise definition and the common misuse.

Sources

• FDA Food Code, 2022 edition — 3-VI (cleaning and sanitizing), 4-501.112 (mechanical warewashing hot-water sanitization temperatures), 4-501.114 (chemical sanitizer concentration and contact-time tables for chlorine, quaternary ammonium, and iodine).
• ServSafe Manager and ServSafe Food Handler curricula (National Restaurant Association) — the practitioner-facing translation of the Food Code's three-compartment-sink method and test-strip verification practice used on the floor.
• NSF/ANSI 3: 2022, Commercial Warewashing Equipment — construction and performance standard for the dish machines referenced in the worked example.
• Anthony Bourdain, *Kitchen Confidential* (Bloomsbury, 2000) — the dish pit as the kitchen's most load-bearing and most overlooked station, a widely cited observation across restaurant-operations writing.
• Eliyahu Goldratt, *The Goal* (North River Press, 1984) — Theory of Constraints; commonly applied by restaurant operations consultants to identify the dish pit as a shared-resource bottleneck rather than a labor problem.
• No direct dishwasher/steward practitioner has reviewed this file yet — flag corrections or gaps via PR.

View SKILL.md source on GitHub · maturity: draft

Jurisdiction: US (baseline)