When should I choose Gain-in-Weight over Loss-in-Weight?

Choose GIW for discrete batch additions into a weighed receiver; choose LIW for continuous or tightly metered flows insensitive to tank level or foam. Mixed lines often use both strategies where each fits.

How can I prevent LIW refills from breaking control?

Detect refill start/stop, suspend closed-loop control or use last-good feedforward, then re-enter closed-loop when the weight-loss slope stabilizes. Log refill events and verify behavior in OQ/PQ.

What accuracy is realistic for dosing?

With capable hardware, isolation, and correct filtering, ±0.5–1.0% of setpoint is common for macro doses; micro can be tighter with weigh-by-difference. Demonstrate with MSA and capability studies.

Our GIW doses overshoot. What should we adjust first?

Improve mechanical isolation and splash control, tune fast/dribble splits, add end-of-flow compensation, and review filter lag. Don’t put physics problems on operator training.

Can one feeder handle both batch and continuous modes?

Many LIW feeders can run batch setpoints, but design trade-offs exist. If cycle time and precision are critical, GIW may be better for batch. Let requirements drive the choice.

How do we prove dosing control to an auditor?

Provide IQ/OQ/PQ, calibration and MSA evidence, SPC on hit-to-target and flow stability, documented refill logic, and eBMR/audit-trail records. Calculations and events must be reconstructable in minutes.

Gain-in-Weight vs. Loss-in-Weight FeedingGlossary

Gain-in-Weight vs. Loss-in-Weight Feeding – Dosing Strategies

This topic is part of the SG Systems Global regulatory & operations glossary.

Updated October 2025 • Feeder Control, Batching & Continuous, Data Integrity • Process Engineering, Manufacturing, QA, Maintenance, Automation

Gain‑in‑Weight (GIW) and Loss‑in‑Weight (LIW) are the two dominant strategies for automated mass dosing. GIW weighs the receiver and measures material added by the receiver’s weight gain; LIW weighs the feeder/hopper and measures material removed as a weight loss over time. On paper both are “scale plus valve/screw.” In production, they behave very differently: GIW shines for discrete batch adds into a vessel; LIW dominates when you need a continuous, metered flow that ignores tank foaming, agitator forces, and level misreads. The unvarnished truth: poor feeder selection or sloppy setup will quietly bleed yield, destabilize quality, and corrupt genealogy. You don’t fix dosing physics with slogans—you fix it with good design, validation, and governed execution.

“If dosing ‘drifts’ every time you refill, you don’t have a people problem—you have a control strategy problem.”

TL;DR: Choose strategy by physics and requirement: GIW for batch increments into a weighed receiver; LIW for continuous or tightly metered flow from a feeder. Bind devices and math in a validated MES; supervise with SCADA; enforce windows and alert/action limits. Control minimum weight, filtering, refill holds, and feeder mechanics under IQ/OQ/PQ, Calibration Status, MSA, and TMV. Feed macro materials with governed Macro Dosing, micro actives with Micro‑Ingredient Dosing, and capture everything in the eBMR. No ungoverned calculators. No free‑pour trims.

1) What GIW/LIW Covers—and What It Does Not

Covers: feeder selection (screw, vibratory, belt), scale configuration, filtering, dribble/fast transitions, refill handling, minimum weight, dynamic response, and interlocks. It includes material handling (hopper geometry, flow aids), integration with MES/SCADA/HMI, recipe logic, and data integrity. It spans micro to macro dosing, batch vs. continuous, and downstream impacts on mass balance and yield variance.

Does not cover: wishful fixes for poor material flow, like telling operators to “tap the bin.” Feeder physics, calibration, and environment are not optional. Nor does dosing strategy replace quality governance—records still need audit trails, approvals, and Document Control.

2) Legal, System, and Data Integrity Anchors

Dosing is part of regulated evidence. Electronic records must meet 21 CFR Part 11/Annex 11 with validated software (CSV/GAMP 5), unique users (UAM), and immutable audit trails. Instruments must be qualified and in status (IQ/OQ/PQ, calibration), utilities verified (UQ), and methods capable (MSA, TMV). Dosing values, feed rates, and setpoint logic belong in the eBMR—not a sticky note on the HMI.

3) Definitions—How GIW and LIW Actually Work

GIW: The receiver (tank/vessel) sits on a scale. Open a feeder/valve; watch the receiver weight increase; stop at target (with fast → dribble transition). Pros: insensitive to feeder tare changes, easy to visualize, perfect for discrete batch additions. Cons: sensitive to external disturbances on the receiver (agitators, operators leaning, foam/entrained air), slower cycle times at large gains, and limited for continuous metering.

LIW: The feeder hopper sits on a scale. Run the feeder; watch the hopper weight decrease; control speed to maintain a mass‑flow rate (kg/h). Pros: excellent for continuous flow, immune to level errors in the downstream vessel, naturally self‑checks via slope; refills can be managed with holds or predictive compensation. Cons: requires reliable refill strategy; sensitive to mechanical disturbances on the feeder frame; demands careful filtering and minimum‑weight management to avoid “chasing noise.”

4) Accuracy & Resolution—Minimum Weight Rules the Game

Feeder accuracy is limited by scale resolution and stability. Define minimum weight per gravimetric best practice; operate well above it. For micro actives, use micro balances and weigh‑by‑difference (Micro‑Ingredient Dosing); for ton‑scale bulks, use robust load cells, isolation, and Macro Dosing feeders. Perform MSA to quantify repeatability and reproducibility; if R&R consumes your whole window, nothing else matters. Tune fast/dribble and end‑of‑flow compensation; stop trusting “feel”—trust data and capability (Cp/Cpk).

5) Dynamics & Filtering—See the Signal, Ignore the Noise

Agitation, splashing, and vibration inject noise into scales. Use mechanical isolation, proper mountings, and digital filtering that preserves step response. For GIW, filter enough to smooth splashes without delaying cutoff. For LIW, compute flow from the slope of weight vs. time; apply moving‑window regression rather than simple differencing; freeze slope during refills. Filtering is a validated parameter, not a weekend experiment—lock the algorithm and coefficients under Document Control and verify in OQ/PQ.

6) Batch vs. Continuous—Pick Strategy by Requirement

For discrete additions to a batch (e.g., “Add 125.00 kg salt to Mixer‑101”), GIW is simple and robust. For continuous processes (e.g., extrusion, coating, fluid bed granulation, inline blending), LIW feeders provide stable mass‑flow with feedback control. Hybrid lines often combine both: LIW for main flow and GIW for periodic minor adds. Don’t force GIW to act like a flow controller or LIW to behave like a one‑shot scale; mismatches produce drift, overruns, and long cycle times.

7) Refill Strategy—The #1 LIW Failure Mode

Refills break the slope. Good LIW systems detect refill start/stop, suspend flow control or switch to open‑loop (last‑good) feedforward, then re‑enter closed‑loop once weight loss stabilizes. Bad systems keep “controlling” during refill, chase a fake error, and overshoot. Define refill thresholds and hysteresis; coordinate with upstream valves; and verify with test runs in OQ/PQ. Capture refill events in the eBMR with timestamps for auditability and CPV trending.

8) Hopper, Screw, and Flow—Fix Physics Before Blaming Software

Bridging, ratholing, and flooding destabilize both strategies. Engineer hopper half‑angles, install mass‑flow liners, and specify the right feeder: single/twin screw for cohesive powders, vibratory for granules, belts for flakes. Add agitation with respect—over‑agitation aerates powders and fakes weight loss. Measure bulk density, angle of repose, and flow function during development; put the numbers into master data. “Operator taps bin” is not a control plan; it’s a symptom list for a future Deviation.

9) Environment & Utilities—Stop External Forces at the Door

Unstable floors, HVAC blasts, and compressed‑air hammer shake scales. Qualify utilities (UQ), map temperature (Temperature Mapping), and monitor vibration. Segregate forklift routes; shield feeder frames; ground for static. If the plant moves, the numbers move—your filters won’t save you. This is basic, and it’s where most dosing “mysteries” die when properly addressed.

10) Calibration, Checks & Status—Evidence or Excuses

Calibrate with traceable weights; verify linearity and corner load. For LIW, verify dynamic response by controlled discharges; for GIW, check end‑of‑flow compensation. Lock calibration intervals and status in Calibration Status; block dosing when expired. Record identities, timestamps, and results in the audit trail. If you “trust” scales without evidence, you are outsourcing quality to gravity and hope.

11) Rounding, UOM & Setpoint Windows—Decide Once

Compute setpoints and dribble thresholds in high precision; round once at presentation. Centralize conversions and significant figures in a governed service (UOM Conversion). Define windows by class (macro vs. micro). Enforce via MES/HMI; flag exceptions with reason codes and e‑signatures. “Close enough” is not close enough when micro actives drive CQAs or regulatory limits.

12) Integration—Make the System Do the Math

Recipes carry dosing logic under Recipe Management and flow through Recipe Versioning. The MES binds devices, pushes targets, supervises fast/dribble, and consumes live weights via SCADA. The WMS stages eligible lots; label scanning (Label Verification) prevents wrong‑material hookups. All events land in the eBMR. If any step depends on a person typing weights into a spreadsheet, your data integrity is already broken.

13) SPC/CPV—Trend What Drifts, Not What’s Comfortable

Chart hit‑to‑target error, LIW flow stability (kg/h %RSD), overshoot/undershoot, refill disturbance magnitude, and time‑to‑steady state. Use SPC rules and Cpk to quantify capability; roll into CPV. If Cpk is low, fix feeder mechanics, filtering, or windows before blaming people. SPC isn’t wall art; it’s a decision engine.

14) Mass Balance, Yield & Cost—Feeding Errors Aren’t Free

Every overrun is scrap or rework later. Tie GIW/LIW results to mass‑balance closure and yield variance. Quantify cost of poor dosing: extra rework solvent, wasted API, lost throughput. Put money on the graph—suddenly the urgency to fix feeder mechanics appears without another speech from QA.

15) Micro vs. Macro—Two Worlds, One Discipline

Micro feeders need high‑resolution balances, anti‑static controls, and weigh‑by‑difference (containers on scales, not cup‑and‑dump guesses). Macro feeders need rugged frames, isolation mounts, and belt/screw sizing that avoids starve or flood flow. Use the right windows, fast/dribble splits, and interlocks. See Micro‑Ingredient Dosing and Macro Dosing for implementation patterns that don’t collapse under audit.

16) Start‑Up, Warm‑Up & Drifts—Respect Device Physics

Load cells and drives need warm‑up; screws need to purge air; belts need to reach stable tracking. Encode start‑up checks in the recipe; block dosing until ready states pass. Track drift over long runs and temperature swings; implement periodic zero checks. Record all status transitions with user/device IDs in the audit trail. If the device is “almost in status,” it is out of status—stop rationalizing.

17) Alarms, Interlocks & Human Factors

Design HMIs for clarity: show live weight, target, fast/dribble state, and refill status. Alarms should be actionable (e.g., “Refill detected—holding control,” “Scale noise high—check isolation”). Require acknowledgments with e‑sign; capture notes in eBMR. Use Dual Verification for high‑risk connects. If operators need tribal knowledge to interpret the screen, the HMI failed—not the operator.

18) Change, Deviation & CAPA—Fix Causes, Not Symptoms

Parameter changes (filter coefficients, fast/dribble split, refill thresholds) must route through Change Control. Treat misses as signals, not blame: use RCA, implement effective CAPA, and close the loop. “Retrain operator” is not a fix for a feeder that bridges every third hour.

19) OEE & TPM—Reliability is a Quality Requirement

Feeder downtime hammers throughput and stability. Roll dosing stops into OEE loss trees; apply TPM to screws, drives, and scales. Standardize cleaning and changeover; validate cleaning to avoid residue that changes friction and flow. Reliability is not “ops’ problem”—it is quality risk in slow motion.

20) Common Pitfalls & How to Avoid Them

Controlling LIW through refill. Suspend or predict; don’t chase a fake error signal.
GIW on a shaky vessel. Isolate mechanics; filter judiciously; don’t fight agitator torque with prayers.
Starved screws/flooded belts. Size feeders and hoppers; add flow conditioners; stop blaming software.
Spreadsheets for setpoints. Centralize math in MES; record in eBMR with audit trails.
Ignored minimum weight. If you’re near the balance floor, you’re guessing—upgrade hardware or split doses.
Rounding drift. Round once at presentation under governed UOM rules.
No genealogy. Don’t refill from mystery bins. Tie lots with WMS and EPCIS.
“We’ll fix it at release.” You won’t. Fix physics, control, and governance at the point of dosing.

21) Metrics That Prove Control

Hit‑to‑target error per add (GIW) and flow stability %RSD (LIW).
Refill disturbance magnitude and re‑stabilization time.
Overshoot/undershoot distribution and Cpk against windows.
Scale noise index (σ of baseline weight) and min‑weight margin.
Deviation/trim frequency and reason‑code completeness.
Mass‑balance closure at batch close and yield variance vs. plan.
Calibration currency (% devices in status) and audit‑trail review health.

If these metrics never force a change in hardware, parameters, or suppliers, retire them. Reports that comfort leadership but don’t change behavior are waste.

22) What Belongs in the GIW/LIW Dossier

Process requirement (batch vs. continuous) and selected strategy; device specs (scale, feeder, frame); filtering and control algorithms; fast/dribble splits; refill logic; hopper design notes; minimum weight evidence; IQ/OQ/PQ results; calibration records; MSA/TMV summaries; integration mapping (MES/SCADA/HMI); WMS lot tie‑ins; SPC/CPV charts; deviations/CAPAs; and approvals under Document Control. Store with effective dates in the QMS.

23) How This Fits with V5 by SG Systems Global

Device‑tight execution. The V5 platform binds scales, screws, belts, and valves to recipe steps; pushes exact targets and windows; supervises fast/dribble and refill logic; and blocks dosing if identity or calibration status fails.

Governed math & records. Conversions, filtering coefficients, and end‑of‑flow compensation live centrally; the complete arithmetic chain lands in the eBMR with user/device IDs and full audit trails. SPC dashboards surface overshoot, flow stability, and refill impact in real time.

Supply & traceability. V5’s WMS stages eligible lots, enforces label verification, and publishes EPCIS/ASN events so genealogy and mass‑balance stay audit‑ready.

Bottom line: V5 turns GIW/LIW dosing into boring, repeatable operations—predictable for Finance, defensible for QA, and fast for Production.

24) FAQ

Q1. When should I choose GIW over LIW?
Use GIW for discrete batch additions into a weighed receiver and when external disturbances are manageable. Use LIW for continuous or tightly metered flow insensitive to tank level/foam. Mixed lines often use both.

Q2. How do I prevent LIW refills from wrecking control?
Detect refill start/stop, hold closed‑loop control or switch to last‑good feedforward, then re‑enter closed‑loop when slope stabilizes. Log events in the eBMR and verify behavior in OQ/PQ.

Q3. What accuracy can I expect?
It depends on minimum weight, mechanics, and filtering. With capable hardware and setup, ±0.5–1.0% of setpoint is realistic for many macro doses; micro can be tighter with weigh‑by‑difference. Prove with MSA and capability studies.

Q4. Our GIW overshoots—what now?
Fix mechanics (isolation, splash control), tune fast/dribble split, add end‑of‑flow compensation, and review filtering lag. Don’t push more training onto operators to compensate for physics.

Q5. Can one feeder do both batch and continuous?
Many LIW feeders support batch setpoints, but design choices differ. If batch precision and cycle time are critical, a dedicated GIW step may be better. Let requirements—not catalog promises—drive the decision.

Q6. How do I prove dosing is under control?
Show IQ/OQ/PQ, calibration and MSA results, SPC on hit‑to‑target/flow stability, refill behavior, and tight eBMR/audit‑trail evidence. If a junior auditor can’t reconstruct the numbers in minutes, your controls are weak.