Fill Weight Checker

A fill weight checker is an integrated weighing control used during packaging or finishing to verify that each unit (e.g., bottle, vial, sachet, capsule, tray) contains the correct net mass. It may operate in-line (dynamic checkweigher with automatic eject) or at-line (static sample weighing), and typically implements both hard acceptance limits and statistical process control (SPC) to steer filler setpoints. In regulated industries, these measurements are in-process controls and must be defined in procedures/recipes, trended, recorded to the batch/device history, and used to disposition nonconforming units and lots.

From an MES perspective, the checker is a critical operation step: it pulls target limits and sampling plans from the approved recipe, captures raw/derived data (e.g., tare, net, mean, range, Cp/Cpk), drives automated actions (rejects, alarms, filler offsets), and generates evidence for release. Its reliability and data integrity are essential because its records substantiate conformance to established specifications and underpin release decisions.

Pharmaceutical manufacturers must establish and monitor in-process controls such as tablet weight or container fill per 21 CFR 211.110, and capture those results in the batch record per 21 CFR 211.188. For dietary supplements, 21 CFR 111.75 requires tests or examinations to verify that specifications (e.g., net contents reflecting strength) are met. Medical device manufacturers must control and monitor production processes per 21 CFR 820.70, which extends to device filling or dosing steps where applicable (e.g., prefilled syringes, reagent kits). EU GMP (EudraLex Volume 4) mirrors these expectations: define in-process controls, monitor them, investigate out-of-specification (OOS) results, and validate computerized systems supporting these activities.

GAMP 5 (2nd ed.) requires a risk-based validation approach for computerized systems like checkweigher integrations and MES functions, aligned to the intended use and data criticality. Data integrity guidance (e.g., MHRA) emphasizes ALCOA+ principles, including contemporaneous, attributable, complete, and enduring records, with secure audit trails. ISA‑95 provides the architectural reference for integrating Level 2 equipment controls with Level 3 MES, ensuring that recipe parameters, setpoints, and results flow in a controlled, traceable manner.

Dynamic checkweighers typically use strain gauge or electromagnetic force restoration load cells with high sample rates to infer net mass as a container crosses the weigh deck. At-line methods use calibrated balances with defined stabilization criteria. In both modes, tare handling (fixed tare tables by SKU, real-time container pre-weigh, or dual-scale gross-tare) is critical to assure net mass accuracy. Environmental controls—vibration, airflow, temperature—must be managed to the instrument’s specification. Metrological requirements (calibration frequency, linearity, eccentricity, and repeatability checks; standard weights; uncertainty budgets) must be defined in SOPs and linked to the LIMS/asset calibration program.

• Tare strategy: fixed nominal tare, measured tare per unit, or periodic tare verification by sample.
• Balance/checkweigher verification: daily checks with traceable weights across the operating range; linearity and eccentricity tests at qualification and periodic intervals.
• Sampling mode: 100% in-line inspection with auto-reject vs. statistically justified at-line sampling (defined plan, e.g., per container count or time).
• Environmental qualification: weigh platform isolation, deck warm-up, airflow shielding, and deck cleaning to avoid residue bias.
• Data capture: raw weights, timestamps, unit/container ID, machine speed, filler head, reject reason codes, and any auto-adjustments applied.

Acceptance criteria should be documented in the master recipe: nominal target, upper/lower action limits (reject thresholds), and where justified, tighter control limits for SPC. SPC (e.g., X̄‑R, EWMA, or CUSUM) helps detect small shifts and reduce giveaway while maintaining compliance. Capability indices (Cp, Cpk, Ppk) quantify how robustly the process meets its tolerance band. When supported by validated integration, the checkweigher can provide closed-loop feedback to the filler (e.g., servo-driven auger, piston, or pump) to nudge the mean back to target, with limits on adjustment magnitude and frequency to prevent oscillation.

• Define separate hard reject limits and SPC control limits.
• Use head/lanes-level charts to localize drift to specific filling nozzles.
• Apply statistically rational subgrouping (e.g., by time, by head) to improve signal detection.
• Record each auto-adjustment event with pre/post statistics in the batch record.
• Re-validate SPC parameters after major change control (e.g., new product viscosity).

A compliant setup ties each measurement to a discrete unit (or a defined subgroup) and its genealogy: container ID/lot, time, filler head, and line speed. Units breaching hard limits must be automatically segregated (mechanical reject, stop belt, or divert) with fail-safe detection (reject confirmation sensor, bin level interlocks). Rework/repackaging, if permitted by procedures, must be controlled and traceable; otherwise, scrap is recorded. The MES should post exceptions to QMS for investigation when trend alarms or reject rates indicate loss of control, and should place affected lots on hold pending evaluation.

For batch records, record: sampling plan execution, individual results or summary statistics (with raw data available), alarms and rejects by code, corrective actions (including filler offsets), and final disposition. This supports release decisions and, if needed, lookbacks to bound suspect intervals using time-stamped equipment states and SPC signals.

Integrating a fill weight checker involves deterministic acquisition of weight and event data, enforcement of recipe parameters, and secure exchange of commands (e.g., start, stop, download limits) with the machine layer. ISA‑95 partitions responsibilities: Level 2 (control) acquires signals and executes reject/feedback; Level 3 (MES) governs specification, execution context, electronic signatures, and record retention. Standard interfaces (e.g., OPC UA/DA, file drops, or device drivers) require qualification and transaction controls (store‑and‑forward, sequence numbers, acknowledgments) to avoid data loss and preserve order.

Design for robustness: time-synchronization (NTP), unique message IDs, idempotent retries, and local buffering for network outages. Security controls include least-privilege accounts, role-based access, and segregated networks; audit events (parameter changes, overrides) must be captured and attributable.

Apply a GAMP 5 risk-based lifecycle: define intended use (URS), assess data/process risks, qualify interfaces (IQ), verify functions (OQ) including limit enforcement, reject logic, SPC calculations, and feedback constraints, then demonstrate performance under normal and worst-case conditions (PQ) at target line speeds and product presentations. Calibrations and routine balance/checkweigher verifications should be integrated with LIMS/asset management, ensuring out-of-tolerance instruments trigger automatic MES interlocks and QMS deviations.

Ensure ALCOA+ data integrity: secure, time-stamped, attributable records; audit trails for parameter changes and overrides; controlled user access; and backup/restore tests. Electronic signatures for critical steps and exceptions should meet regulatory expectations, and audit trail review must be built into routine oversight.

Pharmaceutical (solid oral dose, sterile, radiopharma)

Solid dose often combines tablet press weight control with downstream checkweighers for bottles or blisters. In-process checks under 21 CFR 211.110 must be documented and justified; for sterile products, filling by mass may be preferred for viscous or foaming products, with heightened scrutiny on reject confirmation and line clearance. Radiopharma adds decay corrections and rapid cycle times; data capture must support time-adjusted quantities and very short holds.

Dietary supplements and cosmetics

21 CFR 111.75 requires verification that specifications are met; practical control often targets net content as a proxy for strength when APIs are homogeneously blended. Cosmetics rely on consistent net contents to meet labeling and quality promises; viscosity and temperature shifts often drive drift—SPC and filler temperature control are important.

Food and bakery

Variable-density and sticky products (e.g., doughs, fillings) challenge dynamic weighing. Hygienic design for clean-in-place and crumb build-up control prevents bias. Giveaway minimization is a business imperative: SPC and head-balancing can recapture margin while remaining fully compliant to labeled net quantity.

Key metrics include percent within spec, reject rate by cause, mean vs target, standard deviation, Cp/Cpk, and giveaway (mean minus target, translated to cost). Multi-head analyses often reveal long tails from a few problematic nozzles—head balancing and preventive maintenance can cut both rejects and giveaway. Measurement system capability matters: gage repeatability and reproducibility (R&R), linearity, and stability determine how tight your control limits can be without false rejects.

• Giveaway cost (per lot, per year) to prioritize improvement projects.
• Head/lanes Pareto by reject reason to target maintenance.
• EWMA or CUSUM sensitivity tuning to catch small drifts early.
• Correlation of ambient/product temperature to filler offsets to justify compensations.
• Continuous Process Verification (CPV): persist SPC/KPI trending post-validation.

For continuous verification, retain long-horizon control charts and periodic capability summaries. Integrate maintenance and calibration histories to explain capability changes over time. Use change control where SPC parameters or tolerance bands are modified based on CPV evidence.

V5 Ultimate models the fill weight checker as a controlled operation step within MES: recipes define tolerance bands, sampling frequencies, SPC parameters, and permissible feedback limits. The execution layer captures per-unit or per-sample weights with full context (line, head, container ID), renders live SPC, and enforces reject/hold logic. Records flow directly into eBMR/eDHR with audit trails and electronic signatures where required; deviations and CAPAs are opened automatically when trend rules or hard limits are breached. LIMS-driven calibration status gates execution; WMS quarantine and Maintenance work orders are raised from the same event stream, ensuring the compliance loop closes at execution.

• Uncontrolled tolerance changes at the machine HMI; fix by locking parameters and downloading from MES with audit trails.
• No reject confirmation; add sensors and interlocks to avoid leaking nonconforming units.
• Environmental bias (vibration/airflow) not qualified; implement deck isolation and shielding with periodic challenge tests.
• Poor tare management; ensure consistent container supply or measure tare at a justified frequency.
• SPC configured but not reviewed; automate rule-based alerts and integrate with deviation workflows.
• Calibration lapses; integrate metrology schedules with execution holds and prevent use when due/failed.

Finally, ensure robust exception handling. On communication loss, buffer locally with sequence numbers and reconcile on reconnect. If data are irretrievable, create an investigation record, define bounding assumptions (time windows, quantities), and document impact assessment and disposition in the batch/device record.

"If a tolerance, trend, or override isn’t recorded and attributable, it didn’t happen—at least not in a compliant sense."