V5 Ultimate
Manufacturing · The complete guide

Checkweigher Reject

TL;DR

A checkweigher reject is the fail-safe ejection, segregation, and recording of out-of-tolerance units on high-speed lines. ISA‑95 frames it at the Level 2–3 boundary; Part 11/Annex 11 demand trustworthy electronic records and audit trails. V5 orchestrates the equipment interface, spec enforcement, event logging, and material control so rejects instantly reconcile in the batch/device history and quality system.

Reviewed · By V5 Ultimate compliance team· 3,500 words · ~16 min read

01What it is

A checkweigher reject is the automated, verifiable ejection and secure segregation of any unit whose measured mass is outside defined tolerance bands. In dynamic weighing, each package is weighed on-the-fly; the controller compares the net result to the target and tolerance limits (e.g., T1 for warning/SPC, T2 for hard reject). Out-of-tolerance units are positively diverted into a locked reject enclosure. The system confirms the rejection occurred; if not, the line interlocks to stop, preventing misdelivered product.

In cGMP and food preventive-controls contexts, checkweighers ensure net content compliance and detect upstream filling defects. The reject event, classification (under/over), and identifiers (time, lot/batch, equipment, unit serial where applicable) are captured as part of the batch/device history with appropriate review and disposition. Electronic records and signatures are governed by 21 CFR Part 11 and EU GMP Annex 11; integration to MES is framed by ISA‑95.

02Where it fits (ISA‑95) and architecture

A checkweigher is typically a Level 1/2 device (sensing/monitoring and control) integrated to Level 3 (MES) for specification management, event logging, and exception handling. It receives target weight and tolerance parameters from the MES (recipe or order) and publishes time-stamped measurements, pass/fail determinations, and reject events. The MES or historian consolidates these into the electronic batch/device history for review-by-exception and trending.

Layer (ISA‑95)Typical Role for Checkweigher
Level 0–1 (Process/Basic control)Load cell, weigh cell signal conditioning; conveyor speed control; reject actuator (air-jet, pusher, drop-gate)
Level 2 (Area supervisory)Local controller compares weight vs. limits; drives reject; local HMI shows alarms and verification prompts
Level 3 (MES)Maintains specs/recipes; dispatches tolerances; captures event logs; enforces holds and reviews; manages calibration checks
Level 4 (Business/ERP)Order parameters and labeling; reconciliation of yields; complaints/returns link to weight trends
  • Interface methods: discrete I/O for permissives/interlocks; OPC UA/DA or vendor APIs for real-time data.
  • Data set: unit weight (timestamped), pass/fail code, reject confirmation, lot/batch, order, equipment ID, verification results.
  • Security: unique logins/roles (RBAC), local time sync, and protected audit trails to satisfy Part 11/Annex 11.

03Tolerances, setup, and control logic

Tolerance bands are derived from product specifications and legal/quality requirements. In pharma/supplements, specifications for fill weight or net contents are established per 21 CFR 211 and 21 CFR 111; in food, hazard analysis under 21 CFR 117 identifies preventive controls to ensure correct fill and labeling. The MES downloads the target and tolerance sets at order start and on changeovers; any local edits must be controlled, limited by role, and audit-trailed.

Common reject strategies

  • Two-limit hard reject (underlimit < WL < overlimit): any out-of-band unit is rejected.
  • Tiered limits: inner limits (T2) for automatic reject; outer limits (T3) for line stop and immediate investigation.
  • SPC soft limits: warn operators of drift approaching reject, preventing waste via upstream filler adjustment.
ConditionActionRecord content
WL within T1/T2Accept, continueWeight, timestamp, unit ID (if used)
WL beyond T2 (under/over)Auto-reject to secure binWeight, reason code, reject confirmation, lot/batch, operator
Reject not confirmedHard stop, interlockAlarm, unit tracking flag, investigation required
Sustained drift toward T2Operator adjustment; SPC noteTrend, corrective action, filler set-point change

04Reject station design and fail-safes

Regulators expect positive rejection with physical confirmation: the unit must be demonstrably removed from the acceptable flow. Best practice includes a dedicated, lockable reject bin with access control, presence sensors to confirm arrival, and bin-full interlocks. If the confirmation sensor does not detect the rejected unit, the conveyor should stop and alarm to prevent release. Air-jet ejectors should be validated across the product mass range; heavier formats may require pushers or drop-gates.

  • Reject confirmation sensor located post-diverter; time-windowed correlation to the decision event.
  • Bin full/door open interlocks; line cannot run if reject integrity is compromised.
  • Serialized units (e.g., device kits) require unit ID capture and reconciliation of rejected identifiers.
  • Environmental controls: vibration isolation, belt tracking, and stable temperature to reduce scale noise.

05Data integrity, Part 11/Annex 11, and review

Electronic records of checkweigher measurements and rejects must be attributable, legible, contemporaneous, original, and accurate. 21 CFR Part 11 and EU GMP Annex 11 require secure user accounts, role-based access, audit trails for parameter changes and overrides, time synchronization, and record protection throughout retention. The audit trail must capture who changed tolerances, why (reason for change), and electronic signatures for critical overrides. MHRA’s GxP data integrity guidance underscores periodic audit trail review.

  • Audit trail scope: parameter downloads, local edits, bypasses, verification results, and recipe/version changes.
  • Review-by-exception: automated flagging of out-of-spec, missing confirmations, and repeated near-misses.
  • Link calibration/verification to lot: the eBMR should show the last successful challenge before and after the run.
  • Backups: validated backup/restore for raw event data and derived summaries, with checksums/hash integrity.

06Qualification, verification, and challenge testing

Annex 15 establishes the expectation for IQ/OQ/PQ of critical equipment. Checkweigher qualification covers installation (mechanical, electrical, environmental), operation (repeatability across speeds and formats), and performance (accuracy, precision, and reject efficacy). Use certified test masses or test packs that reflect product geometry to assess dynamic accuracy. 21 CFR 211.68 and 21 CFR 111 require calibration and routine checks; define challenge frequency (e.g., start-up, shift change, after maintenance, and at close) with documented acceptance criteria.

  1. Define accuracy targets under dynamic conditions (e.g., ±0.5 g at nominal 120 ppm).
  2. Establish challenge sets: under-limit, at-limit, and over-limit; include borderline cases to test decision logic.
  3. Verify reject confirmation and bin integrity; test bin-full and door interlocks.
  4. Document worst-case: highest and lowest tare, fastest speed, and tallest package; update PQ when ranges change.

07Risk management, CCP alignment, and SPC

Where net contents are a regulatory or safety-critical attribute, the checkweigher reject can function as a CCP under HACCP. Under 21 CFR 117, the hazard analysis should justify whether weight is managed as a preventive control and define corrective actions (reject, hold, filler adjustment, batch evaluation). For pharma and supplements, weight is often a critical quality attribute; link SPC to filler set-point control to reduce both consumer risk and waste.

  • Define decision trees: when T2 rejects exceed a rolling threshold, pause, investigate upstream fill variation, and document CAPA linkages.
  • Use SPC rules (e.g., Western Electric) to escalate before violations; tie to automated alerts.
  • Perform capability analysis (Cp/Cpk) on dynamic weights; re-center processes when capability erodes.
  • Trend by SKU, format, speed, and operator to reveal training or mechanical contributors.

08Exceptions, disposition, and records

All rejects must be accounted, reconciled, and dispositioned. The eBMR/eDHR should include counts by reason (under/over, unconfirmed reject, bin full), links to investigations for atypical patterns, and evidence of QA review. For pharma (21 CFR 211.188) and supplements (21 CFR 111.260), record actual yields vs. theoretical and the results of in-process controls. If any rejected units are reworked (e.g., reweighed after top-up), capture traceability to the original serial/lot and the re-inspection results.

  • Immediate actions: isolate reject bin under status control; verify seal and access logs.
  • Investigation triggers: repeated unconfirmed ejections, sudden mean shifts, or excessive underfills.
  • Documentation: reason codes, photos (optional), maintenance tickets, filler adjustments, and retest outcomes.
  • Release criteria: QA verifies completeness of challenges, audit trail reviews, and acceptable trend profiles.

09Integration, cybersecurity, and change control

Integrate via secure, authenticated protocols with least privilege. Apply defense-in-depth per NIST SP 800-82: network segmentation (cell/area zones), firewalling between Level 2 and Level 3, allow-listing of traffic, and hardening of vendor HMIs. Enforce RBAC for local parameter access; all changes should originate from approved MES recipes under change control and be audit-trailed. Validate interfaces (including time sync, loss-of-comms behavior, and store-and-forward) under GAMP 5 guidance.

  • Loss-of-comms: define whether the device runs last-known-good parameters or forces a controlled stop.
  • Clock and timezone: single source of time truth to ensure chronological integrity of unit-level events.
  • Software lifecycle: supplier assessments, version control, and regression testing of reject logic changes.
  • Disaster recovery: cold-standby configurations and validated restoration of device parameters and MES mappings.

10How V5 handles checkweigher rejects

Within V5, checkweigher parameters reside in the approved recipe/specification set, versioned and released through change control. At start, the order dispatch pushes targets and limits to the line; events stream back in real time to the execution record. Rejects automatically generate status-controlled material movements (WMS), optional quarantine holds, and quality events when thresholds are exceeded. SPC is embedded for drift detection and capability metrics, and audit trail reviews are workflow-driven for timely QA sign-off.

Frequently asked questions

Q.Is a checkweigher reject always a Critical Control Point (CCP)?+

Not always. In food, the hazard analysis under 21 CFR 117 determines whether net contents warrant a preventive control/CCP designation. In pharma/supplements, it is often a critical quality attribute governed by cGMP specifications rather than a food safety CCP.

Q.What must be recorded for each rejected unit?+

At minimum: timestamp, measured weight, pass/fail reason (under/over), lot/batch, equipment/line, reject confirmation status, and operator (if manual intervention). If unit-level IDs exist, record the serial. Include linkage to pre/post-run verification challenges.

Q.How frequently should checkweigher verification challenges be performed?+

Define in URS/SOPs based on risk and speed: typically at start-up, changeover, shift start, after maintenance, and at end-of-run. Evidence should show challenges across the allowable weight and speed ranges, with documented acceptance criteria and audit trails.

Q.Can operators locally adjust tolerance bands?+

Only under strict controls. Use RBAC to restrict edits, require electronic signatures and reasons for change, and ensure MES-driven specifications are authoritative. All changes must be audit-trailed per Part 11/Annex 11 and subject to change control and periodic review.

Q.What are common failure modes that require line stop?+

Reject not confirmed, reject bin full or opened, loss of communications/time sync, repeated near-threshold underfills indicating drift, or verification failures. SOPs should force a safe state and initiate investigation before resuming.

Q.How does this impact batch yield and reconciliation?+

Reject counts and reworks feed directly into actual yield, variance analyses, and reconciliation. The eBMR/eDHR must show theoretical vs. actual yields, the reasons for variance, and QA disposition of rejected or reprocessed units for release decisions.

Primary sources

Further reading

See Checkweigher Reject working on a real shop floor

V5 Ultimate ships with the Checkweigher Reject controls already wired in — audit trail, e-signatures, validation evidence. Free trial, no credit card, onboard in days, not months.