Built 2026-08-30. Source: r6-cooling-rules-and-safety.md.

R6. Cooling targets and safety for the freezer-fan plus eutectic-goop retrofit

Written 2026-08-30. All prices and quotes were seen on 2026-08-30 unless noted. Web search budget ran out partway through, so a few product facts are marked "not confirmed" rather than guessed. The transcripts of both Tech Ingredients videos were read first (transcript2.txt = fan video, transcript.txt = eutectic video).

Scope: numbers to design against, plus the hazards that actually matter in a commissary. No compliance lecture. The one legal point that matters is at the top of section 1.

1. Cooling targets

1.1 No code requires a "blast chiller"

The FDA Food Code judges the product temperature curve, not the equipment. Section 3-501.15(A) lists "rapid cooling equipment" as one of seven allowed methods and ends with "(7) Other effective methods." A freezer with a fan in it is an "other effective method" as long as the food hits the times below. Same in NYC: Health Code 81.09(e) names "an ice bath, a rapid chill unit or adding ice" plus shallow pans, portioning, stirring, spacing and loose covers. Nobody asks for an NSF-listed blast chiller.

What the code does care about is the freezer's other job. FDA 3-501.11: "Stored frozen FOODS shall be maintained frozen." NYC 81.17(a) says refrigeration units "shall maintain potentially hazardous foods at or below 41 degrees Fahrenheit ... at all times." The fan video relies on warming the frozen stock "five or six degrees" from -20 to -25 C, which keeps it frozen. The eutectic bottles exist to take that heat instead of the food. Design so the frozen inventory never gets near 0 C, and you are inside both rules.

1.2 The numbers (FDA Food Code 2022, 3-501.14, verbatim)

"(A) Cooked time/temperature control for safety food shall be cooled: (1) Within 2 hours from 57 C (135 F) to 21 C (70 F); and (2) Within a total of 6 hours from 57 C (135 F) to 5 C (41 F) or less. (B) TIME/TEMPERATURE CONTROL FOR SAFETY FOOD shall be cooled within 4 hours to 5 C (41 F) or less if prepared from ingredients at ambient temperature, such as reconstituted FOODS and canned tuna."

Annex 3 adds the design intent: "An example of proper cooling might involve cooling from 135 F to 70 F in 1 hour, in which case 5 hours remain for cooling from 70 F to 41 F." And on why a plain freezer is a weak tool: "Commercial refrigeration equipment is designed to hold cold food temperatures, not cool large masses of food. Rapid chilling equipment is designed to cool the food to acceptable temperatures quickly by using very low temperatures and high rates of air circulation." That sentence is the whole justification for the fan.

Source: FDA Food Code 2022 PDF, https://www.fda.gov/media/164194/download (Chapter 3 pp. 25-27; Annex 3 pp. 113-117).

1.3 NYC and NYS differences

NYC Health Code 81.09(e)(1) (the one that applies to a NYC commissary): "rapidly cooled from 140 degrees Fahrenheit (60 degrees Celsius) to 70 degrees Fahrenheit (21.1 degrees Celsius) within 2 hours and from 70 degrees Fahrenheit (21.1 degrees Celsius) to 41 degrees Fahrenheit (5 degrees Celsius) within 4 hours after cooking or removal from hot holding." The clock starts at 140 F rather than 135 F, and the second stage is written as "4 hours" after the first rather than "6 hours total"; in practice the same 2 + 4 window. 81.09(e)(2): ambient-ingredient foods to 41 F within 4 hours. 81.09(e)(1)(A) puts a number on pan depth: "shallow pans or containers (less than 4 inches in height)". (E) says "not stacking or nesting". (F) says "loosely covered, or uncovered if protected from overhead contamination."

NYS 10 NYCRR 14-1.40 (outside NYC): "reduced from 120 degrees Fahrenheit (48.9 degrees Celsius) to 70 degrees Fahrenheit (21.1 degrees Celsius) within two hours and to 45 degrees Fahrenheit (7.2 degrees Celsius) or below within four additional hours." Looser endpoint (45 F), later start (120 F). NYC's 41 F governs in the city.

Sources: NYC Article 81 PDF, https://www.nyc.gov/assets/doh/downloads/pdf/about/healthcode/health-code-article81.pdf (81.09 at pp. 12-14, 81.17 at pp. 21-22); NYS 14-1.40 via Cornell LII, https://www.law.cornell.edu/regulations/new-york/10-NYCRR-14-1.40.

1.4 Design targets to hit (my translation)

Stage Code limit Design target Why
140/135 F to 70 F (60/57 C to 21 C) 2 h 60 min or less Annex example; leaves margin for the thickest pan
to 41 F (5 C) 6 h total (NYC: 4 h after stage 1) 3 to 4 h total Margin for a full load and door openings
Ambient-ingredient prep (salads, reconstituted, cooked-then-mixed) 4 h to 41 F 2 h Same
Frozen inventory during a cooling run must stay frozen never above -10 C (14 F) at the warmest bottle or pack Keeps 3-501.11 / 81.17(a) unarguable
Freezer air between runs no code number back to the thermostat setpoint before the next batch Compressor is the bottleneck; the goop only buys time

Benchmark from the fan video (800 mL water in a 1 kg stainless pot, residential freezer at -20 to -25 C): 50 C to 0 C in about 60 min with the fan at 9 V versus about 120 min without; fully frozen at 165 min versus 426 min. That load is 0.8 kg. A 4 in hotel pan of stock is 6 to 8 kg and a 2 in pan of cooked grain is 3 to 4 kg, so measure, do not extrapolate. Heat to remove for stage 1 is roughly 150 kJ per kg of watery food (57 C to 21 C, cp about 4.2 kJ/kg K); to 5 C it is about 220 kJ/kg. For scale, melting 1 kg of ice absorbs 333.5 kJ (Osborne, J. Res. NBS 23, 643, RP1260, 1939; PDF fetched but the NIST URL could not be confirmed), so 10 kg of hot stock needs the latent heat of about 6.5 kg of ice-equivalent to reach 5 C. The freezer's compressor pays it back later.

1.5 Which of your items are covered

Under the FDA definition (1-201.10), TCS food includes "an animal FOOD that is raw or heat-treated; a plant FOOD that is heat-treated". Stocks, sauces, braises, cooked grains and beans are all TCS, so the 2 h / 6 h clock applies to them. Cooked rice and large containers of rice are called out by name in the Annex as slow-cooling items and are the classic Bacillus cereus case.

Non-TCS (no cooling clock) only when pH or water activity is low enough per Table A/B of the definition: vinegar-forward sauces and pickling liquids at pH 4.6 or below, syrups and honey-based items at aw 0.85 or below (Table B), dry goods. If you want a sauce off the clock, measure pH; do not assume.

"Does not require a certified blast chiller" is true of every item. "Does not require timed cooling" is true only for the acid and low-aw items.

1.6 What inspectors flag (NYC scoring, from DOHMH's own mapping)

Code What NYC Health Code cite Points
02H cooked TCS food not cooled by an approved method within the time/temperature window 81.09(e) 7 to 10 depending on item count
02I ambient-ingredient food not at 41 F within 4 h 81.09(e)(2) 5 to 8
02G cold TCS food held above 41 F (this is what a hot pan warming a reach-in triggers) 81.09(a) 7 to 10, critical, 28 if not corrected on site
04J no accurate probe thermometer available 81.09(g) 8
04E / 08C toxic or hazardous chemical improperly stored, labeled or used 81.17(g) 7 to 10 / 2 to 5, critical 28 if not corrected
06C supplies or equipment stored near a source of contamination 81.07(i) 5 to 8

Source: https://raw.githubusercontent.com/nychealth/Food-Safety-Health-Code-Reference/main/Violation-Health-Code-Mapping.csv. Treat the point values as indicative; the CSV is DOHMH's reference file, not the inspection worksheet itself.

The practices behind those codes, in order of how often I have seen them written up: deep pots (over 4 in) cooling in a walk-in; pans covered with plastic wrap while hot; stacked hotel pans; a stockpot put straight into a reach-in that then reads 45 F on the inspector's probe; no probe thermometer in reach; unlabeled chemical containers. Cooling in a freezer is not itself a violation. Cooling in a freezer that thaws the frozen product, or that has unlabeled bottles of black goop on the shelf above the food, is.

2. Ammonium chloride goop: hazards that matter

2.1 Ammonium chloride (NH4Cl)

From the BASF food/pharma-grade SDS (rev. 2025-06-05, https://download.basf.com/p1/000000000030046480_SDS_GEN_US/en/Ammonium_chloride_S_foodpharma_grade_Safety_Data_Sheet_USA_English.pdf):

Food status: 21 CFR 184.1138 affirms ammonium chloride as GRAS as a direct food ingredient (dough strengthener, flavor enhancer, leavening agent, processing aid, cGMP limit only) (https://www.law.cornell.edu/cfr/text/21/184.1138). EU number E510; it is the salt in salty licorice. So a trace of the salt on a pot is a flavor problem, not a poisoning. The reason a leak still means "discard the food" is the other ingredients (carbon fiber, bentonite, technical-grade salt if you did not buy food grade) and the fact that you cannot tell how much got in.

Scale check: one bottle holds about 295 g of salt. For a 70 kg adult the rat LD50 scales to roughly 100 g, so a bottle is a few LD50-equivalents. Nobody is going to drink the goop, and this is a rodent number, but it settles the "it's just a food additive" argument: the bottle is a chemical container, label it like one.

2.2 The one reaction to actually worry about: bleach

Ammonium salts plus sodium hypochlorite make chloramines; industrial monochloramine is made exactly this way from ammonium chloride and hypochlorite. Washington DOH: "When bleach is mixed with ammonia, toxic gases called chloramines are produced" with coughing, shortness of breath, chest pain, and fluid in the lungs at high exposure (https://doh.wa.gov/community-and-environment/contaminants/bleach-mixing-dangers). A kitchen sanitizer bucket is chlorine. If a bottle bursts and someone mops it with the bleach bucket in a closed freezer box, that is the accident. Rule: goop spills get flushed with plain water, then a quat sanitizer. Keep the recipe sheet on the wall with that rule on it. Also keep the goop away from alkaline degreasers and lye (release of ammonia, per the SDS incompatibility with bases).

2.3 Leak on food or on the evaporator

2.4 The other ingredients

Bentonite (4 g per bottle): the mineral is not a listed carcinogen, but SDSs list 0 to 5 % crystalline silica (quartz), which IARC classes Group 1 for lung cancer when inhaled as respirable dust; OSHA PEL 0.1 mg/m3 respirable quartz (Western Clay SDS, https://www.westernclay.com/wp-content/uploads/2017/01/SAFETY-DATA-SHEET.pdf). 4 g scooped from a bag is a small exposure. Wear an N95 while scooping, wet-wipe the dust, do not dry sweep.

Xanthan gum (12 g): combustible dust in air, "Do not breathe dust", and the SDS's own capitalized warning: "Slippery, can cause falls if walked on" (CP Kelco Kelzan SDS, https://sds.chemtel.net/docs/Asbury%20Carbons%20Inc-0001931/Site%20SDS/CP%20Kelco%20U.S.%20Inc_Kelzan%20(GHS)_Unknown_06-27-2017_English.pdf). A wet xanthan spill on a quarry-tile floor is the most likely injury in this whole project. Mop it with lots of water at once.

Chopped carbon fiber (20 to 25 g): the SDS points that matter (HiTemp CF51 SDS, https://hitemp.ca/wp-content/uploads/2018/07/CF51_Carbon-Fiber.pdf.pdf): - "Carbon fibers are electrically conductive and may cause a short circuit in electrical equipment if filaments become airborne." Do not open the bag near the freezer's control board, the fan PSU, or the reach-in's evaporator fan and defrost heater terminals. - Mechanical skin and eye irritation from fragments; sizing can chemically irritate. Nitrile gloves, safety glasses. - No OSHA/ACGIH limit for carbon fiber itself; nuisance dust PEL 5 mg/m3 respirable, 15 mg/m3 total. Respirable fibers are "not expected" in normal handling; use a HEPA/N95 respirator if you see dust. Chopped tow sheds more than fabric. - Fibers embedded in skin: remove with tweezers. - Practical: the black fibers are a foreign-material contaminant in food. Do the mixing in a non-food area with a dedicated mixer bowl and whisk (or a cheap dedicated hand mixer); do not use the kitchen's KitchenAid bowl and then wash it for food.

2.5 Mixing safety

3. Containers

Option Cold behavior at -20 to -25 C Drop/leak Cleanability, labeling Verdict
PET sports-drink bottle (video) PET is glassy at all kitchen temperatures (Tg 69 to 85 C, https://ctherm.com/resources/newsroom/thermal-analysis-labs/determining-glass-transition-temperature-of-pet/); toughness falls further in the cold; the polypropylene cap is the weak point since "polypropylene and PVC are both plastics that are brittle at temperatures below freezing" (U.S. Plastic Corp., https://www.usplastic.com/knowledgebase/article.aspx?contentkey=954) Fine sitting on a shelf; a frozen bottle dropped on tile can split. Fill to 90 % for freeze expansion (video) Clear, so the black goop is visibly not food; smooth, washable; label sticks Acceptable for a trial, retire at first crack
HDPE wide-mouth jug or lab bottle "High, medium and low polyethylene ... remains impact resistant at low temperatures" (same U.S. Plastic source); HDPE is a container material named in the BASF SDS Survives drops; screw lids seal Opaque, so labeling is mandatory; wide mouth means no funnel or vacuum loader Best DIY choice. Price not confirmed
Stainless hotel pan with lid 304 pans are not on the SDS's material list (316Ti is); chloride pitting of 304 over many freeze-thaw cycles is plausible, no test data Lids are not sealed; thawed goop sloshes; looks exactly like a pan of food Cleanable but the confusion risk is the highest of the four Avoid, or dedicate an odd size (1/6 pan) with a riveted tag
Commercial PCM brick (Arctic Ice Tundra Series, 5 F / -15 C) Same phase-change point as the goop; rigid shell, "leak proof ultra sonic spin welded cap", labeled "non toxic" Made for coolers; drop-rated by design Nothing to explain to an inspector arctic-ice.com 2026-08-30: 2-pack Large 2.5 lb $42.99, X-Large 5 lb $64.99, XX-Large $85.99, all three showing sold out (https://arctic-ice.com/products/tundra-series). Walmart lists a single Large at $43.99 (search snippet, page blocked). Roughly $19/kg versus about $1.50/kg of DIY goop (video's 80 cents of salt plus about 45 cents of fiber per bottle)

Separation and labeling, both codes: - FDA 7-201.11: separate by spacing or partitioning, and "in an area that is not above FOOD, EQUIPMENT, UTENSILS". 7-102.11: working containers "clearly and individually identified with the common name of the material." NYC 81.17(g)(1)-(2): containers "prominently labeled", stored "in designated storage areas in which food, equipment and utensils are not stored." Read literally, 81.17(g)(2) does not want chemicals in a food freezer at all. The defensible position is that a sealed, labeled coolant pack is operating equipment (like a commercial ice pack), stored in its own bin. Expect the question; have the recipe sheet and the SDSs in the binder. - Practical layout: bottom shelf (upright) or floor (chest), in a lipped HDPE tote or milk crate that catches a full bottle's worth of leak, below all food, away from the evaporator. Fixed count written on the crate ("12 bottles"); count on the way in and out of every run. - Label text, freezer-rated tape or a laminated tag on a zip tie: "NOT FOOD. DO NOT OPEN. Ammonium chloride coolant. If leaking: water rinse, NO BLEACH. Discard any food it touched." Black marker on clear PET is not enough once frost forms. - Inspect every bottle at every thaw for cracks and cap seepage; PET goes first at the cap threads and the base petals.

4. Electrical safety of the 12 V fan

What is actually being installed (Amazon listings seen 2026-08-30): - Fan B0DHKQS3DC: "16 inch Slim Fan Push Pull Electric Radiator Cooling Fan, Black, 2 Packs", 12 V 120 W, CFM at or above 1000, 2000 rpm, 10 blades, 3.27 in thick. $59.99 for two. 120 W at 12 V is 10 A. The listing makes no sealed-motor or UL claim; the video's "sealed motor" statement is the video's, not the seller's. (https://www.amazon.com/dp/B0DHKQS3DC) - PSU B09YSJQWRG: Jesverty SPS-3010 0-30 V 0-10 A bench supply, $47.49 (list $49.99). The listing mentions short-circuit and overload protection, a grounding terminal and a "temperature-regulated fan", and no UL, ETL or other safety listing. It is an open-vented lab instrument. (https://www.amazon.com/dp/B09YSJQWRG)

Rules that make this safe in a kitchen:

  1. Low voltage side is not the shock hazard. 12 V DC is far under the 60 V DC Class 2 ceiling; NEC Article 725 / UL 1310 Class 2 sources are limited to 100 VA and their wiring "can be" run without conduit or raceway (PULS application note, https://www.pulspower.com/wp-content/uploads/2023/10/The-PULS-Advantage-Issue-11-NEC-Class-2-Power-Supplies-Aug-2025.pdf). The hazards on the DC side are heat from a short through a pinched lead, and a bench supply that can push 10 A into that short. Fuse the DC lead at the supply (inline blade fuse sized just above the running current you measure at 8 to 9 V).
  2. The 120 V side is the hazard, and it lives outside the box. Mount the PSU on a wall bracket or shelf above mop height, never on the floor. A fan-cooled bench supply will inhale grease and humidity; if this goes past a trial, swap it for a listed, enclosed supply. The Mean Well GST60A12 (12 V 5 A, 60 W) is UL 62368-1 and CSA C22.2 No. 62368-1 listed with a -30 to +70 C working range (Mean Well spec sheet, https://www.meanwell.com/Upload/PDF/GST60A/GST60A-SPEC.PDF); DigiKey shows it at $18.60 (search snippet; the page was blocked, so treat the price as unconfirmed). Note that 5 A will not run this 120 W fan at 12 V; it is enough only at the 8 to 9 V the video recommends, and only if you have measured the current there. A fixed 12 V supply needs a PWM speed controller or a 9 V supply to get the video's operating point.
  3. GFCI. NEC 210.8(B)(2) and (3) (2023) require GFCI on all 125 to 250 V single-phase receptacles up to 50 A in kitchens and "areas with sinks and permanent provisions for food preparation" in non-dwelling units (https://www.electricallicenserenewal.com/Electrical-Continuing-Education-Courses/NEC-Content.php?sectionID=1425). NYC's 2025 Electrical Code (Local Law 128 of 2024) is the 2020 NEC with amendments, in full effect 21 December 2025 (https://www.nyc.gov/site/buildings/codes/electrical-code.page), and the 2020 edition already required commercial-kitchen GFCI. So the receptacle you plug the PSU into should already be GFCI; if it is not, use a GFCI cord adapter. Leave the freezer's own plug where it is; freezers are often deliberately on non-GFCI circuits to avoid a nuisance trip spoiling inventory, and that is a separate decision.
  4. Cable through the box. One flat two-conductor 16 AWG lead through the door gasket at a lower corner, silicone-sealed, with a loop of slack so the door does not tension it. No splices inside the box; crimped, insulated spade or ring terminals at the fan tabs; heat-shrink over them. Strain relief (a zip-tie anchor) at the fan and a cord grip at the PSU. Do not tap the freezer's internal 120 V wiring (evaporator fan, defrost heater, light) for anything.
  5. Reach-in specifics. Keep the fan, cable and bottles clear of the OEM evaporator and its airflow. NYC 81.17(a)(2) wants air circulation "not obstructed" and refrigeration fans "clean, dust free, and in working condition"; an inspector will look at an aftermarket fan the same way. A 16 in radiator fan has open blades on both faces; add a wire finger guard on the intake side if the fan can run while someone is loading pans.
  6. Heat budget. At 8 to 9 V the video reports under half the fan's power for about 80 % of its airflow; call it 40 to 60 W of heat dumped into the box while the fan runs. The video's residential freezer compressor was 75 to 94 W; a commercial reach-in's is several times that, a chest freezer's about the same as the video's. Run the fan on a timer or switch only while a load is cooling. Measure with a plug-in watt meter before and after; the compressor duty cycle going up is your "freezer stress" indicator.
  7. Wet floor plus 12 V 5 A: not a shock issue, and a 60 W supply cannot push enough current to matter through a person. The realistic bad outcome is the 120 V adapter sitting in mop water. Keep it high, keep it plugged into GFCI, done.

5. Cooling log that takes five minutes

Required kit (NYC 81.09(g)): probe thermometer accurate to plus or minus 2 F, "made from food grade materials", "readily accessible" where cooling happens. FDA 4-302.12 says the same. ThermoWorks ThermoPop 2 is $31.20 (4.5 in probe) or $39.20 (8 in) at thermoworks.com on 2026-08-30 (https://www.thermoworks.com/products/thermopop-2). Ice-bath check (32 F) monthly, written on the log.

Wireless, since you already run YoLink (prices from shop.yosmart.com, 2026-08-30): - YS8004 Weatherproof Temperature Sensor with Probe, $35.49 (sale from $39.99): 4 ft probe, IP67 body, two AAA lithium (about 2 years), reports hourly and on any 0.5 C change lasting over a minute, hub required (https://shop.yosmart.com/products/ys8004 and https://shop.yosmart.com/pages/ys8004-support). YoLink's FAQ says to bag the body if it goes in a freezer; better to keep the body outside and pass only the probe through the gasket. Not sold as food-grade; use it in a sentinel pan of water at the same depth as the batch, or for freezer air, rather than in the food. The support page lists a probe range of "-80 to 200" for the current revision without units; confirm before relying on it below -20 C. - YS8007 Thermometer Hygrometer, $19.99 (sale from $27.99): -20 to 54 C, freezer-rated per the listing (https://shop.yosmart.com/products/ys8007). Its -20 C floor is marginal for a freezer that idles at -25 C; the YS8004 probe is the better air sensor.

Log columns, one row per batch: item, pan depth and mass, time in, product temp in, temp at +1 h, +2 h (must be at or under 70 F), +4 h, +6 h (must be at or under 41 F), freezer air low/high during the run, bottle count in/out, initials. If +2 h misses 70 F, reheat to 165 F (NYC 81.09(d)) and restart or discard; log which. Keep 90 days. The YoLink app exports CSV, and the change-triggered reporting gives a dense curve during the first hour, which is the data you want when tuning fan voltage and pan depth.

Sources (all seen 2026-08-30)