Codes referenced
In-operation accuracy within ±0.5% of full scale; ±0.25% at installation; calibration every 3 months for load cells.
Batching tolerance ±2% for cement & mineral admixtures; ±3% for aggregate, chemical admixture & water; tare zero at least once each day.
Aggregate mass basis; relevant for batching by mass with allowance for free moisture content.
Accuracy class III reference for batching scales; complements IS 4926 limits.
Why weighing accuracy is the foundation
Concrete mix design assumes each constituent is dosed within a tight tolerance. The mix-design water-cement ratio, the SCM replacement percentage, the aggregate-cement ratio — all assume the cement content is what the design specified. A 5% over-batch of cement on every cubic metre for a 60 m³/h plant is 9 kg of cement per cubic metre wasted, or about 540 kg/h lost margin. A 5% under-batch on a 30 MPa design drops the target mean strength by roughly 2.5–3 MPa — enough to push some 28-day results below the IS 456 Table 11 individual gate of fck − 3 and trigger an acceptance investigation.
In a typical 1 m³ M30 batch, cement is about 360 kg, fine aggregate about 800 kg, coarse 20 mm about 660 kg, coarse 10 mm about 330 kg, water about 170 kg, and admixture about 3 L. The cement is the smallest batch component by mass — which is exactly why its weighing scale has the tightest tolerance and is the most sensitive to drift.
A cement scale that reads consistently 1.5% high is over-batching 5.4 kg of cement per cubic metre. On a 1000 m³/month volume at ₹8/kg cement cost, that is ₹43,200/month in cement cost that the plant cannot recover through billing — recovered only when the scale is re-calibrated.
The 5-point calibration test
A single test-point calibration (e.g. only the full-scale point) catches gain error but misses linearity error, hysteresis, and corner-loading error. IS 4926 Annex E's 5-point protocol covers all four:
| Test point | % of rated capacity | What it tests | Why it matters |
|---|---|---|---|
| 25% | Quarter-load | Low-end linearity | Catches offset error that single-point cal would miss |
| 50% | Half-load | Mid-range linearity | Where most production batches actually operate |
| 75% | Three-quarter | Higher-range linearity | Catches non-linear gain drift |
| 100% | Full capacity | Gain at full scale | The reference point for any single-test calibration |
| 110% | Overload check | Confirms the cell still reads correctly above nominal range | Required because batch controllers sometimes overshoot the setpoint |
Apply test loads in ascending order, then repeat in descending order — the average of ascending and descending at each point is the calibrated reading. Hysteresis (difference between ascending and descending) above 0.3% of full scale is an early indicator of cell fatigue.
The ±0.5% / ±0.25% distinction
The standard defines two accuracy bands and applies them to two different lifecycle moments:
- At installation / reconditioning — within ±0.25% of full-scale reading
- During operation — within ±0.50% of full-scale reading
- Calibration interval — at least once every 3 months for load cell systems
Source: IS 4926:2003 Annex E. Always quote the latest BIS revision.
Drift detection — when to suspect a load cell
A load cell almost never fails catastrophically. It drifts. The trick is to catch the drift before it crosses the ±0.5% line on the production floor. Five signals:
The 5-point test should be reproducible within ±0.1% point-to-point month to month. If the same point fails each month — say 75% reads high — that point is the cell's compromised range. Replace the cell.
If the plant's daily yield variance steadily trends over- or under-yielding for two weeks without an obvious moisture or water-meter cause, the cement scale is suspect. Check.
Tare the empty hopper. Display should read 0 ± 1 scale increment (typically 1 kg). If the zero is unstable, the cell has mechanical damage or cable degradation.
Apply 100 kg test weight. Rotate the hopper 90° in plan. Reading should be within 0.1%. A 0.5% or greater shift means the cell is taking a side-load — likely a mounting bolt has failed or the cell is bent.
A 50 kg cement bag dropped on the hopper, a stone block wedged in the gate, or a chute strike can shift the calibration. Always re-verify after any shock event before the next production batch.
Empty the hopper, tare, fill with the same nominal cement weight ten times. Standard deviation of displayed readings > 0.5% of full scale = cell is mechanically compromised.
Traceability — why NABL-accredited weights matter
A 5-point calibration is only as good as the weights used. IS 4926 Annex E clause (n) sets a minimum of 500 kg of stamped weights, except that for low-capacity scales an acceptable limit is 20% of scale capacity.
For the calibration to be acceptable to QCI RMC PCA auditors and to institutional clients (PWD, MES, Indian Railways, large private developers), the weights must be traceable to national standards through a NABL-accredited calibration laboratory. The chain is:
Working weights at the plant
Traceable to clause (n)
- · 500 kg of stamped weights — preferably in 20 kg and 50 kg pieces for stacking flexibility across the 5 points.
- · Recalibrated annually by a NABL-accredited lab or compared in-situ against reference weights.
Reference weights at the NABL lab
Annual certification
- · Maintained by a NABL-accredited calibration agency; calibrated against national prototype kilograms.
- · Provides a certificate traceable to National Physical Laboratory (NPL, New Delhi) standards.
Annual certification at the plant
External NABL or OEM service
- · External NABL-accredited lab or OEM service engineer performs the full 5-point calibration using its own traceable weights.
- · Issues a calibration certificate with measured error at each point and pass/fail against the ±0.5% in-operation limit.
- · This certificate is the document the auditor or client will ask for during a quality audit.
Why the cement scale is the most critical
Cement is the smallest mass component in the batch, which makes it the most sensitive to absolute weighing error. Worked example for an M30 batch:
| Parameter | Value |
|---|---|
| Cement scale full-scale capacity | 600 kg |
| Cement in 1 m³ M30 | 360 kg |
| ±0.5% of full-scale reading | ±3.0 kg per batch |
| Same absolute error as % of cement dose | ±3.0 / 360 = ±0.83% of cement mass |
| Aggregate scale full-scale capacity | 3,000 kg |
| Aggregate (sand + coarse) in 1 m³ M30 | 1,790 kg |
| ±0.5% of full-scale reading | ±15 kg per batch |
| Same absolute error as % of aggregate dose | ±15 / 1,790 = ±0.84% of aggregate mass |
| Interpretation | Both scales are calibrated to the same absolute error band, but the cement scale has the smallest margin in the mix design — a fixed 3 kg cement error shifts w/c and target strength more than a 15 kg aggregate error shifts aggregate-cement ratio. |
This is why IS 4926 Cl 9.1.2 imposes a separate, tighter batching tolerance on cement and mineral admixtures (±2%) than on aggregate and water (±3%). The cement scale deserves the most careful monthly attention.
The cost of uncalibrated load cells
Three direct costs and one indirect:
If cement is being over-batched, the customer is receiving more cement than billed for. Plant margin erodes silently. Often only caught during a yield reconciliation review.
If cement is under-batched, target mean strength is missed. Cube results drift toward the lower acceptance limit. Eventually a sample fails IS 456 Cl 16.1 — costly investigation, possible core testing, possible rejection of the pour.
A drifting cement scale produces a drifting batch-to-batch yield. The plant cannot reconcile batched vs. dispatched vs. returned concrete. Dispatch records become unreliable for invoicing.
An external audit finding on weighing accuracy can suspend QCI RMC PCA certification until the issue is resolved and re-verified. Loss of certification = loss of eligibility for institutional clients.
Worked example — M30 plant 5-point calibration
| Test point | Applied (kg) | Displayed (kg) | Error (kg) | % of full scale | Pass / fail (±0.5%) |
|---|---|---|---|---|---|
| 25% | 150 | 151 | +1 | +0.17% | PASS |
| 50% | 300 | 298 | −2 | −0.33% | PASS |
| 75% | 450 | 447 | −3 | −0.50% | BORDERLINE |
| 100% | 600 | 598 | −2 | −0.33% | PASS |
| 110% | 660 | 659 | −1 | −0.17% | PASS |
| Interpretation | 75% point is borderline; cell shows ~0.3% systematic under-read in the mid-range. Re-calibrate span. Repeat next month. If 75% point fails again, replace cell. | ||||
Note: "Error as % of full scale" is the convention of IS 4926 Annex E — not the more intuitive "% of applied weight" or "% of reading" used in some international standards (OIML R76 uses % of reading for accuracy classes). Track both conventions in your records; auditors will expect IS 4926's.
Common deviations in Indian RMC plants
Plant applies 100% capacity weight once a quarter, declares "calibrated". Annex E's 5-point requirement is missed; linearity error goes undetected.
A vendor's "calibration sticker" without NABL-traceable certificate. Auditor rejects the evidence; plant cannot prove the scale is within tolerance.
A failed cell is replaced but only the affected scale is re-tested. Per Annex E the whole weighing system — including adjacent cells in the same hopper assembly — should be re-verified for corner-loading.
Cl 9.1.2 requires scale zeroing at least once each day. A build-up of cement on the weigh hopper walls shifts the tare by 2–4 kg. Over a week this drift is invisible; over a month it changes the cement dose by 1%.
Field-checklist blocks
A short questionnaire, an inspector's checklist, and three worked examples that show how the 5-point reading, the accuracy band and the daily tare zero resolve into a cement dose you can defend to an auditor or a client.
Questionnaire — ask before the cement scale is released back into production
- Was the last full 5-point calibration done within the last 3 months, at 25 / 50 / 75 / 100 / 110% of rated capacity, as IS 4926 Annex E requires for load-cell weighing systems?
- Was it done with stamped weights traceable through a NABL-accredited laboratory, with at least 500 kg of weights (or 20% of scale capacity for a low-capacity scale), and is the weights certificate referenced on the calibration certificate?
- Does the certificate report the measured error at each of the five points and a pass/fail against ± 0.5% of full scale, in Annex E's convention, rather than a single "calibrated" stamp?
- Is the in-operation accuracy within ± 0.5% of the full-scale reading, and was the installation or reconditioning accuracy within ± 0.25%?
- Was the scale tared on an empty hopper before production started, and does it now read 0 ± 1 scale increment?
- Has anything hit this hopper since the last calibration (a cement bag, a stone wedged in the gate, a chute strike), and was it re-verified before the next batch?
- Is the readout increment within limits for cement, not more than 2 kg on a digital scale and not more than 5 kg on an analogue one?
- If a cell was replaced, was the whole weighing system re-verified, including the adjacent cells in the same hopper assembly, for corner loading?
Inspector's checklist — what to verify on site
- 5-point record: applied load, displayed reading, error in kg, error as a percentage of full scale, and pass/fail at each of the five points.
- Ascending and descending readings at each point, with the calibrated reading taken as the average and the hysteresis difference written down. Above 0.3% of full scale is an early indicator of cell fatigue.
- Calibration certificate from the NABL-accredited lab or OEM service engineer, showing measured error at each point, pass/fail against ± 0.5%, and the reference to the weights certificate.
- Daily tare-zero log: one entry per day on an empty hopper, with the reading taken before the first batch.
- Monthly repeatability test: ten fills at the same nominal weight, with the standard deviation of the displayed readings. Above 0.5% of full scale means the cell is mechanically compromised.
- Direction check: a 100 kg test weight applied, the hopper rotated 90° in plan, and the shift read. 0.1% is expected; a shift of 0.5% or more means side loading, so check the mounting bolts.
- Month-on-month comparison of the 5-point results, which should be reproducible within ± 0.1%. The same point failing repeatedly identifies the compromised range of the cell.
- Yield reconciliation trend, reviewed for a one-directional over- or under-yield over two weeks with no moisture or water-meter cause.
What happens if… three worked examples
What happens. On a 600 kg cement scale, the 75% point is 450 kg applied and the scale displays 447 kg, so the error is − 3 kg. As a fraction of full scale that is − 3 / 600 = − 0.50%, exactly on the ± 0.5% in-operation limit, so the point is borderline rather than a pass. Look at the three mid-range points together: − 2 kg, − 3 kg and − 2 kg, an average of − 2.33 kg = − 0.39% of full scale. The under-read is systematic, not scatter.
Likely outcome. Re-calibrate the span, repeat the test next month, and if the 75% point fails again, replace the cell. Meanwhile check where that error sits against the batching tolerance: − 3 kg on a 360 kg cement dose is − 3 / 360 = − 0.83%, comfortably inside the ± 2% cement tolerance of IS 4926 Cl 9.1.2. So the dose is still legal while the scale is sitting on its limit, which tells you the ± 0.5% accuracy band is the binding constraint and the batching tolerance is not.
Preventive correction. Report the error as a percentage of full scale in your records, because that is Annex E's convention. Some international standards use a percentage of applied weight, and an auditor will expect yours to follow the Indian code.
What happens. 1.5% of the 360 kg of cement in an M30 m³ is 5.4 kg of extra cement in every batch. On 1,000 m³ a month that is 5.4 × 1,000 = 5,400 kg, and at ₹8/kg that is ₹43,200 a month of cement the plant cannot recover through billing. Now put that against the two tolerances. 1.5% is inside the ± 2% cement batching tolerance of Cl 9.1.2, so the batching reconciliation looks acceptable. But 5.4 kg is 5.4 / 600 = 0.9% of full scale, which is outside the ± 0.5% in-operation band of IS 4926 Annex E.
Likely outcome. A scale can sit inside the batching tolerance and still fail the accuracy requirement, so a clean yield reconciliation is not evidence that the scale is calibrated. The plant finds this at the next 5-point test, not at the invoice, and the margin is only recovered once the span is corrected. That is the argument for the cement scale getting the most careful monthly attention of any scale on the plant.
Preventive correction. Reconcile the cement mass in the yield review against the scale's own calibration record. A yield trend that is consistently off in one direction is a scale problem until it is proved otherwise.
What happens. Cl 9.1.2 requires the scale to be zeroed at least once each day. A build-up of cement on the weigh hopper walls shifts the tare by 2 to 4 kg. On a 360 kg cement dose that is 2 / 360 = 0.56% to 4 / 360 = 1.11%, so roughly 1% of the cement in every batch, added or taken off the design quantity, every single day.
Likely outcome. Over a week the drift is invisible and over a month it has changed the cement dose by about 1%. The calibration certificate does not catch any of it, because the certificate is a snapshot taken on a tared hopper. The daily zero is the control that catches it, and the readout increment sets the resolution of the argument: not more than 2 kg for cement on a digital scale, so a 2 to 4 kg tare shift is one to two increments of the display. Nothing else in the plant sees it.
Preventive correction. Make the zero part of the start-of-shift routine and record it. It is the cheapest check on the list and it is the one that most often gets skipped.
Frequently Asked Questions
How often must load cells be calibrated at an RMC plant?
What is the acceptable tolerance for a cement load cell?
How do you detect a drifting load cell?
Related articles
Amit Haridas
Founder & Proprietor, ConcreteInfo. 25+ years experience in construction QA/QC, concrete technology, RMC plant operations, consulting and technical training across India. For load-cell calibration programme setup, 5-point verification procedures, or NABL-traceable certification, contact amit@concreteinfo.in.
Technical updates, not sales emails
One email when a new article goes out on concrete technology, QA/QC and construction quality. No promotions, and your address is never sold or passed to advertising networks. Subscribing is separate from making an enquiry.