Cold chain · Frozen & ice cream

    Frozen food doesn't spoil. It quietly loses the quality you sold.

    The damage is recrystallization and freezer burn from brief partial thaws — a propped cross-dock door, a reefer cycling off — that never look spoiled at the dock and that a once-an-hour logger sails straight past. And −18 °C isn't even fully frozen. Navixy reads the reefer over CAN/J1939 and in-load probes, then alerts on rate-of-change and door-dwell — not a static line — so a thaw is caught while it's minutes old and packaged as proof.

    Get a deployment plan Cold-chain overview
    Reefer CAN / J1939 + in-load probesRate-of-change & door-dwell, not a static −18 °C lineDeep-frozen probes to −55 °CMKT & time-out-of-range proof
    Frost-covered ice-cream tubs and cartons stacked on a pallet inside a deep-freeze cold store
    Anatomy of a frozen quality loss
    Frozen corridor · ≤ −18 °C
    Peak product temp
    −11.4 °C
    Time above −18 °C
    17 min
    Disposition
    QA / FEFO

    A “green” logger still ships a ruined load

    Frozen quality loss is cumulative and invisible: it depends on how long and how warm, not on a single pass/fail reading. Navixy evaluates every packet, tells a defrost from a real thaw, and rolls the cumulative stress into a number a quality claim can't argue with.

    Catch the minutes-long thaw an hourly sample reports as clean

    A thermo-logger sampling every 15–60 minutes can land on either side of a 5–10-minute thaw and report a perfect trip. Navixy evaluates every packet against the band and the rate of change, then rolls up mean kinetic temperature and time-out-of-range — the cumulative damage the logger averaged away.

    Load FZ-118 · frozen · ≤ −18 °C
    per-packet
    Continuous trace vs. hourly samples7-min thaw at the cross-dock
    -25°-20°-15°-10°-18°C-26°C+6.2 °C
    Hourly logger verdict
    “In range” ✓
    What actually happened
    7 min at −11.8 °C
    Level of frozenness~96% @ −18 °C

    The last few % of water is still liquid — which is exactly what a brief thaw refreezes.

    −18 °C is a convention, not “fully frozen”

    At −18 °C most foods are only ~95–98% frozen, so a little liquid water remains for a brief warm-up to recrystallize. Navixy alerts on movement and product-specific bands, not a single line everyone hard-codes.

    Ice cream lives far below −18 °C

    The cited heat-shock threshold for ice cream is −28.9 °C — so a reefer “holding −18 °C” can still coarsen premium dessert. Deep-frozen probes and a colder band protect the products that need it.

    defrost rise — expected, muted
    real thaw — sustained, alerts

    Tell a defrost from a real thaw

    A scheduled defrost and a dock door both raise return-air briefly. Dwell timers and door correlation suppress those so the channel isn't muted — and the real, sustained thaw still pages someone.

    Frozen−22 °C
    Chilled+2 °C
    Ambient18 °C

    One rule per zone, not one fleet-wide line

    QSR and foodservice loads run frozen, chilled, and ambient on one vehicle. A single threshold both misses the frozen excursion and cries wolf on the chilled zone; Navixy gives each compartment its own.

    Anatomy of a frozen quality loss

    Two short thaws, one cumulative loss — caught while the load can still be saved

    No single reading ever screams “spoiled.” But a propped cross-dock door and a reefer that cycles off each lift the product above −18 °C just long enough to recrystallize it. Watch the rate-of-change alerts fire, the handler intervene, and the trip resolve into a record that quantifies the cumulative stress an hourly logger would have averaged away.

    Frozen corridor · ≤ −18 °Clive shipment telemetry
    -18 °C-26 °C
    Plant freezerLine-haulCross-dockRetail DC
    Condition timeline
    • Departs fully pulled down to −22 °C
    • Cross-dock: door propped open during transfer
    • Rate-of-change alert — product climbing fast
    • Partial thaw: −11.4 °C against the open door
    • Reefer cycles off in start/stop mode
    • Second thaw flagged; cumulative stress logged
    • Trip sealed — MKT and time-above-−18 on the record
    Proof of condition
    Peak product temp
    −11.4 °C
    Peak product temp
    Time above −18 °C
    17 min
    Time above −18 °C
    Disposition
    QA / FEFO
    Disposition
    How it's built

    From a reefer signal to a defensible record, in four moves

    The same composable platform that runs fleet and field operations, configured for frozen — buildable on hardware you already approve.

    1. 01

      Sense the unit and the load

      Read the reefer over CAN / J1939 — setpoint, supply and return air, mode, run-hours, defrost, fuel, door — and pair it with in-load 1-Wire probes rated to −55 °C, the range ice cream and IQF actually need.

    2. 02

      Decide on movement, not a line

      IoT Logic compares each reading to the previous — value('temperature',0,'valid') vs index 1 — to catch a rate-of-change, and uses door-open dwell and defrost-mode context so expected rises stay quiet.

    3. 03

      Act before recrystallization

      A sustained thaw or a propped door pages the handler nearest the load — close the door, restart the unit, re-cool — while the cargo is still minutes from quality loss, not at delivery.

    4. 04

      Prove the cumulative stress

      IoT Query rolls up MKT, time-above-band, excursion count, door-dwell, and freezing-time-to-core into a HACCP/FSMA-ready record, exported to your BI or quality system over the API.

    Hardware & integrations

    Reefer CAN, deep-frozen probes, and a rule for every zone

    Frozen is a range-and-mode problem: the right probe has to read below −30 °C, and the right alarm has to tell a defrost from a thaw. Navixy normalizes reefer J1939, 1-Wire and BLE probes, door, and GPS into one data model across 2,500+ device models, then pushes the proof into the systems your quality and claims teams already run.

    • Reefer J1939 for setpoint, supply/return air, mode, run-hours and defrost — so a normal defrost rise is never mistaken for a thaw
    • Deep-frozen 1-Wire probes (to ~−55 °C, ~1⁄16 °C resolution) for ice cream and IQF, where −20 °C-rated BLE tags fall short
    • One virtual sensor and one rule per compartment for multi-temperature QSR / foodservice loads — never a single fleet-wide line
    • On-board store-and-forward backfills the audit trail when a frozen trailer sits unhooked in a dead-zone yard
    • Push MKT, time-out-of-range, door-dwell and freezing-time records to your BI / QMS over an open API — and ship it white-label
    Multi-temperature trailer · one sensor & rule per zone
    Frozen
    ≤ −18 °C
    Chilled
    0–4 °C
    Ambient
    15–25 °C
    Each compartment becomes its own virtual sensor with independent thresholds — the frozen zone's thaw pages someone while the chilled zone stays quiet.
    Frost-covered ice-cream tubs and cartons stacked on a pallet inside a deep-freeze cold store
    Cross-dock thaw−11.4 °C
    4–15%
    of food is lost at the distribution stage (FAO)
    Invisible at the dock

    It passes inspection — then surfaces as a texture complaint weeks later

    Recrystallized ice cream and freezer-burned product look fine when the load is accepted, paid, and the responsible leg forgotten. That's why the highest-leverage frozen deliverable isn't a temperature read — it's a continuous, exportable record that pins the quality loss to a leg and a moment, so a chargeback is defended instead of absorbed.

    • Recrystallization (icy/grainy) and freezer burn are quality, not safety, failures — invisible to a receiver
    • A continuous record defeats an unfounded spoilage claim and pins a real one to the responsible party
    • The same data exposes reefer-off fuel fraud on contractor frozen routes — value beyond spoilage
    How real-time excursion alerts work
    FAQ

    What frozen & ice-cream teams ask

    What temperature should frozen food be kept at — and why isn't −18 °C “fully frozen”?
    −18 °C is the international standard, but most foods are only ~95–98% frozen there. The IIR sets −18 °C as the reference set-point; because a little unfrozen water remains, brief warm-ups still recrystallize it (Le-Bail et al., Int. J. Refrigeration, 2025). So a static −18 °C alarm is the wrong primitive — Navixy alerts on rate-of-change and product-specific bands.
    Why does my ice cream arrive icy or grainy if the truck “stayed cold”?
    Because texture is destroyed by fluctuation, not by an average. Short partial thaws cause recrystallization — small ice crystals melt and large ones grow and fuse, coarsening the texture (Lomolino et al., 2019). The cited ice-cream heat-shock threshold is −28.9 °C, far below −18 °C, so a reefer “holding −18 °C” can still ruin premium dessert. Per-packet rate-of-change monitoring catches the thaw the average hides.
    Isn't the reefer's own temperature reading enough for frozen?
    No — supply/return air is not product temperature, and it doesn't read defrost correctly. The unit's sensors sit in the airflow, not the load, and a normal defrost cycle shows an expected rise that fools naive alarms. Navixy reads the reefer over CAN/J1939 and in-load 1-Wire/BLE probes, and uses dwell and correlation logic so defrost and dock doors don't trigger false alarms.
    A logger said the trip was fine, so why the quality complaint?
    An hourly logger can sail past a 5–10-minute thaw. Quality loss is cumulative — it depends on time and temperature together (time-temperature-tolerance), so a brief excursion between samples does real damage while the log reads clean (FAO). Navixy evaluates every packet and rolls up mean kinetic temperature and time-out-of-range, quantifying the stress a sampling logger averages away.
    What's the difference between freezer burn and recrystallization?
    Freezer burn is surface dehydration; recrystallization is internal ice-crystal growth — both driven by fluctuation. Freezer burn is moisture subliming from the surface and oxidizing, leaving dry, discolored patches; it's a quality, not safety, issue (USDA; IFT). Both are worsened by temperature swings and door openings — which is exactly why door-dwell and rate-of-change rules matter for frozen.
    How do I prove frozen compliance for HACCP / FSMA / raw-fish freezing?
    With a continuous, exportable record tied to location and door events. FSMA Sanitary Transportation expects equipment that holds spec plus records; HACCP treats freezing and frozen storage as critical control points; the FDA Food Code requires documented freezing time/temperature for parasite destruction in raw-service fish. Navixy's IoT Query generates MKT, time-out-of-range, freezing-time-to-core, and chain-of-custody exports on a schedule.
    We run frozen + chilled on one truck. One threshold, right?
    No — use one sensor and one rule per zone. A single fleet-wide line both misses the frozen-zone excursion and cries wolf on the chilled zone. Navixy creates per-zone virtual sensors with independent thresholds and a unified view — built for QSR and foodservice multi-temperature distribution.
    Does monitoring frozen loads pay for itself?
    Usually on the first defended claim or prevented load. Industry puts distribution-stage food loss at 4–15% (FAO) and a single high-value rejected load at tens to hundreds of thousands of dollars. For frozen the bigger win is defeating unfounded spoilage claims with a record and exposing reefer-off fuel fraud — value well beyond spoilage itself.

    Protect the texture you engineered — all the way to the freezer aisle

    Tell us your products, lanes, and zones. We'll map the reefer integration, the deep-frozen probes, the rate-of-change and door-dwell IoT Logic rules, and the MKT / time-out-of-range records for your frozen cold chain — so an invisible thaw becomes an alert you can act on and a claim you can win.

    Get a deployment plan See how it's built