Frozen Fruit Incoming Inspection: Sampling and Defect Guide

    A frozen-fruit receiving check is reliable only when the sample represents the lot, the defect names are measurable and the release rule was agreed before the truck or container arrived. Looking into one open carton can reveal a serious problem, but it cannot prove that every pallet is uniform. A certificate can support inspection, but it cannot replace a connected check of identity, pack, frozen condition, cut distribution, defects, sensory quality and intended-use evidence.

    At XMSD, we recommend separating three decisions: whether the shipment can enter frozen storage, whether representative samples meet the purchase specification, and whether the fruit performs in the buyer's application. Each decision needs its own evidence and authority. That structure makes release faster when results are clear and makes a hold defensible when they are not.

Define the Lot Before You Choose the Sample

    A lot should be a defined quantity produced under conditions that make a common decision meaningful. Record supplier, product, specification version, production facility, production date or time range, supplier lot code, pack size, pallet count, container or vehicle, seal, purchase order and received quantity. If several production codes or facilities share one shipment, decide whether they are separate inspection lots.

    Do not merge two codes merely because the carton artwork and fruit name match. Variety, maturity, line setting, cutting time or pack condition can differ. Conversely, splitting one homogeneous production lot into many convenient mini-lots after a failure can distort the agreed acceptance plan. Write the rule during supplier approval and keep the commercial contract aligned with it.

    Use the XMSD frozen fruit range to define the product family, then make the inspection plan format-specific. Whole berries, dice, slices, chunks, pulp and puree require different sample units and preparation. A method designed for whole strawberries should not be copied blindly to mango dice or a frozen puree block. For example, an IQF frozen blueberry program may use diameter bands and whole-berry damage, while dice need three-dimensional cut and fines rules.

Build a Risk-Based Sampling Plan

    The plan should name the unit selected, how many units are selected, where they come from, the laboratory or inspection sample mass, the tests, the acceptance rule and what happens to opened packs. Random selection matters, but distribution across the shipment also matters operationally. Include front and rear, upper and lower pallet positions, different pallets and every declared production code where the agreed plan requires it.

    Do not invent an AQL after seeing the result. Codex commodity standards use defined sample units, defect tolerances and lot-acceptance language for the products within their scope. You may use a different contract plan for industrial fruit, but the plan must still state the statistical and commercial basis. Sample size and acceptance number belong together; copying only one creates false confidence.

    Before the first commercial shipment, run a method-alignment exercise with the supplier. Have both sides inspect split portions from the same composite sample, then compare sample preparation, category definitions, scale resolution, rounding and total-defect treatment. If the results differ, resolve the method before debating fruit quality. Save the agreed photo board and calculation sheet with the specification. This small exercise is especially valuable when fruit is fragile, categories overlap or inspectors work in different countries and languages.

Record Shipment Condition Before Opening Cartons

    Capture arrival time, seal, vehicle or container condition, odor, cleanliness, pallet movement, load shift, carton wetting, crushing, puncture, tape, labels and visible frost before the evidence changes. Match labels to the purchase order and specification. A wrong product code or undeclared ingredient is an identity issue, not a small visual defect.

    Measure temperature by the agreed instrument, location and stabilization method. Air, surface, between-pack and product temperatures answer different questions. Record the instrument ID, calibration status, exact location, time and whether a pack was penetrated. Do not label a lot thawed from a photograph of surface frost, and do not declare it sound from one cold reading near the door.

    Codex CXC 8 uses -18°C or colder as the quick-frozen cold-store and transport baseline and emphasizes controlled transfers and temperature records. Contract limits and competent-authority rules still govern the shipment. A deviation triggers an evidence review; it does not automatically prove safety, damage or fitness.

Use a Defect Dictionary That Inspectors Can Repeat

    Every defect entry should contain a name, observable definition, included and excluded conditions, severity class, measurement basis and limit. "Poor appearance," "too much damage" or "not premium" cannot support consistent sorting. If a visual boundary is difficult to describe, attach signed reference photographs and retain actual approved and rejected examples where practical.

    Codex CXS 52 for quick-frozen strawberries demonstrates the principle: it defines partially uncolored, completely uncolored, disintegrated, stalk, blemished and misshapen fruit, then connects those definitions to stated tolerances and a sample method. Use the applicable law, commodity standard and buyer contract for the actual product. The strawberry limits are not universal limits for every fruit or industrial use.

    Control the inspection environment as well as the wording. Use consistent lighting for color, a neutral tray, a verified scale and clean tools. Set a maximum time between thaw completion and sensory review so discoloration or softening during inspection does not become a supplier defect. If peel, core, stems or calyces are permitted at a limited level, collect each fraction separately and photograph it beside the sample code. We want another trained inspector to reach the same category and result from the record.

Separate Critical, Major and Minor Findings

    Severity should describe consequence, not how dramatic a defect looks. A small hard pit fragment may present a different consumer and equipment risk from a large amount of harmless broken fruit. Wrong identity, undeclared allergen or confirmed hazardous foreign material can require immediate hold and escalation. Broken pieces may be commercially major for a retail whole-berry pack but acceptable raw material for puree.

    Write which findings allow routine defect measurement and which stop the inspection for a controlled investigation. Preserve the item, its exact sample code, surrounding product, carton, pallet position, photographs and chain of custody. Do not wash, bend or discard suspected foreign material merely to identify it quickly. The investigation may need supplier, insurer, laboratory or regulatory evidence.

    For quality defects, keep categories mutually exclusive where the total is calculated. If one strawberry is both pale and crushed, state whether it is counted in both individual categories but once in total defects. Otherwise two inspectors can produce different totals from the same tray.

Inspect Frozen and Prepared Samples Under Signed Methods

    The frozen inspection captures free flow, clumps, loose ice, surface frost, pack adhesion, obvious color, foreign material and frozen cut. A controlled thaw or preparation step reveals odor, flavor, texture, internal discoloration, drip, drained yield and defects hidden by frost. Record sample mass, container, temperature, time, screen, drain period and any mixing.

    Do not force separate pieces apart with a method that damages them, then count the damage against the supplier. Conversely, do not ignore a block that cannot be separated under the signed free-flow test. The XMSD frozen strawberry dice reference illustrates why color, cut, pale tissue, fragments and clumping need separate observations.

Worked defect-percentage calculation

    Worked example: an inspector examines a 2,000 g prepared sample and separates 46 g of broken pieces, 18 g of blemished fruit and 8 g of defined extraneous vegetable material. If categories are mutually exclusive, total measured defects are 72 g. The result is 72 ÷ 2,000 × 100 = 3.6% by mass. Compare 3.6% only with a limit written on the same prepared-sample and mass basis. A count-based or surface-area limit cannot be substituted.

Measure Size as a Distribution, Not One Showcase Piece

    For dice, set length, width and thickness rules or a sieve method appropriate to the application. For berries, use diameter bands, count, mass or a calibrated screen. For slices, measure thickness and relevant length. State whether the unit is measured frozen, tempered or thawed and whether fragile broken pieces enter the size result or a separate defect category.

    Use enough pieces to describe normal variation. Report the percentage inside the target band, oversize, undersize and fines. One caliper photograph proves that measurement occurred on one piece; it does not prove the lot distribution. The XMSD frozen apple slice format is a useful example of why thickness and wedge geometry need more than a nominal size name.

Worked size-distribution result

    Worked example: 200 measured pieces produce 168 inside the agreed band, 12 oversize, 14 undersize and 6 too damaged to measure. In-range result is 168 ÷ 200 × 100 = 84%; oversize is 6%; undersize is 7%; damaged is 3%. Apply the signed rule to each category. If the application requires 90% in range, the lot does not pass merely because the average dimension looks correct.

Calculate Usable Yield Before Calling a Lot Economical

    Delivered mass can include fruit, released juice, loose ice, declared glaze, clumps, rejected defects and material that is unsuitable for the final application. Define which fractions are usable. A beverage may recover released juice; a retail whole-fruit pack may reject broken pieces; a bakery filling may accept fragments but reject peel or pits. The same lot can have different usable yield under two legitimate applications.

    When comparing the XMSD frozen mango dice format or another fruit, divide delivered cost by accepted yield and add sorting, handling, disposal and downtime. Keep trial costs separate from market promises and repeat the test across commercial lots.

Worked usable-cost comparison

    Worked example: a 1,000 kg lot at USD 1.70/kg costs USD 1,700 delivered. The signed inspection and application trial identify 930 kg as usable. Fruit cost per usable kilogram is 1,700 ÷ 930 = about USD 1.83/kg before sorting labor and disposal. A second offer at USD 1.76/kg with 980 kg usable costs about USD 1.80 per usable kilogram. The higher delivered price is lower in this trial only if both lots meet safety, sensory and process requirements.

Do Not Let Quality Inspection Replace Food-Safety Controls

    Visual inspection cannot prove that a fruit is free from pathogens, residues, undeclared allergens or microscopic contamination. Freezing is not a sterilization step. Supplier approval, preventive controls, intended-use assessment, validated processing, sanitation, laboratory programs and traceability must sit beside visual lot checks.

    Decide whether the fruit is ready to eat or will receive a validated lethal treatment. That decision changes microbiological specifications, environmental controls, sample handling and the meaning of a laboratory result. FDA's berry prevention strategy notes the history of hepatitis A and norovirus outbreaks associated with fresh and frozen berries and emphasizes preventive action across the chain rather than treating one finished-product test as complete protection.

    Keep laboratory chain of custody separate from a thawed sensory sample. Use clean tools, sealed containers, sample codes, required temperature control and documented transfer. A sample opened for sorting or tasting may no longer be suitable for every microbiological test.

Use a Controlled Hold and Disposition Process

    Physically and electronically identify stock on hold before it becomes available to production. Record quantity, pallet IDs, reason, evidence, date, responsible person and allowed movements. Preserve frozen conditions while the investigation proceeds. A hold is a controlled status, not an accusation and not permission to leave product in a warm dock.

    An alternate use is not a shortcut for a safety failure. Broken berries that miss a whole-fruit visual limit may be technically acceptable for a validated puree route, but only if the fruit still meets every safety, legal, composition and traceability requirement for that route. Record the quantity and prevent unintended release under the original label.

Build a Claim File That Another Person Can Reconstruct

    Keep the purchase order, specification version, supplier confirmation, shipping documents, seal, logger data, pallet map, carton and lot photos, sampling record, methods, raw results, calculations, separated defect photos, retained samples, laboratory reports, hold record and commercial communication. Use filenames and image boards that preserve the sample identity.

    Photograph scale displays with the sorted fraction and sample code in the same frame where practical. Include overview and close views, consistent lighting and a size reference. Avoid photographing only the worst handful without showing how it relates to the inspected mass. A credible claim explains the population, sample, method and result.

    Send the supplier a precise question: whether it accepts the result, needs a jointly agreed review, proposes a commercial remedy, requests evidence or identifies a production record for investigation. Keep urgent stock-control action separate from the long-term corrective-action review.

Frequently Asked Questions

Can you accept a frozen fruit lot from one carton?

    One carton can reveal identity, pack or severe defects, but it usually cannot establish lot conformity unless the signed sampling plan says so. Select distributed units under the agreed plan and preserve the location and code of every sample.

Should defects be counted by number or by weight?

    Use the basis written for that defect and product. Small hard fragments may be counted, while broken fruit may be measured by mass. State the sample preparation, denominator and units so the result matches the limit.

Does a COA replace incoming inspection?

    No. A COA can report selected results for an identified lot and method. Receiving inspection also checks identity, packaging, transport condition, temperature evidence, visual quality and the connection between documents and physical stock.

What should happen when one critical foreign object is found?

    Follow the signed critical-defect and food-safety procedure: stop routine handling, control the lot, preserve the object and sample identity, assess related stock, notify authorized technical personnel and investigate before disposition. Do not average it into ordinary visual defects.

Which information belongs in a frozen fruit inspection RFQ?

    State fruit, form, cut or size distribution, ingredient and treatment rules, defect dictionary, sample unit, sampling plan, measurement basis, intended use, pack, storage, safety documents, change control, release rule and claim evidence.

References

• Codex Alimentarius, Standard for Quick Frozen Strawberries, CXS 52-1981.
• Codex Alimentarius, Code of Practice for Quick Frozen Foods, CXC 8-1976.
• USDA Agricultural Marketing Service, Commodity Specification for Frozen Fruit.
• USDA, Commercial Item Description A-A-20379, Fruits, Frozen.
• U.S. FDA, Prevention Strategy for the Control of Enteric Viruses in Berries.