Codes referenced
Abrams cone, three equal layers with 25 rodding strokes per layer, vertical subsidence in mm. The procedure every plant QA must run on every load.
Cube mould filling, rodding, compaction, initial curing at 27 ± 2 °C, and transfer to water curing at 27 ± 3 °C until test age. The spec that gives every cube a fair fight.
150 mm cube compression testing procedure — specimen preparation, capping (sulphur or neoprene), loading rate, failure-criteria reading.
Cl 15.2 sampling rate; Cl 16.1 / Table 11 acceptance (mean of 4 ≥ fck + 0.825σ or fck + 3, individual ≥ fck − 3). The downstream gate every cube must clear.
"Concrete mixes shall be randomly sampled and tested for workability, and where appropriate, plastic density, temperature and air content. Where significant variations from target values are detected, corrective action shall be taken."
C143 slump; C1064 concrete temperature; C172 sampling; C31/C192 making/curing cylinders; C39 compression. Use for cross-spec projects or NABL-audit cross-references.
Why production-level QA differs from site-level QA
A site lab tests concrete after it has arrived — to confirm what was delivered. A plant lab tests concrete before it ships — to catch the deviation while the mixer is still in front of you. The two labs do different work.
- Real-time adjustment. A plant slump of 60 mm against a 100 mm target means the controller recipe is wrong — adjust before the next batch. A site slump of 60 mm against a 100 mm target means the truck has already left the gate — re-tempering or rejection.
- Mix-design feedback loop. Plant QA feeds back into the recipe (target w/c, PCE dose, moisture correction). Site QA feeds back into the placement sequence (vibration, finishing, curing).
- Cost asymmetry. Catching a 2% over-watering at the plant saves a truck load. Catching the same 2% at the cube test wastes a pour.
- Frequency. Plant QA tests every load on the four core items (slump, temperature, density where required, occasional air). Site QA tests a representative sample for acceptance under IS 456 Cl 15.2.
The two are not interchangeable — they are complementary layers in the same quality system.
The production test matrix — frequencies and action limits
| Test | Standard | Frequency at plant | Action limit | ASTM parallel |
|---|---|---|---|---|
| Slump | IS 1199 (Pt 2):2018 | Every load, or 1 per 25 m³ | ±25 mm from mix-design target | ASTM C143 |
| Concrete temperature | Plant QA discipline (IS 7861 for hot weather) | Every load | Trigger cooling ≥ 32 °C; pause ≥ 38 °C | ASTM C1064 |
| Density (where required) | IS 1199 (Pt 3):2018 | 1 per shift per grade | ±2% of theoretical density | ASTM C138 |
| Air content (air-entrained mixes) | IS 1199 (Pt 4):2018 | Every 25 m³ or every load | ±1.5% of target | ASTM C231 |
| Cube sampling for strength | IS 456 Cl 15.2.2 / IS 516 Pt 5/Sec 1 | 1 sample (3 cubes) per 50 m³ or per grade per day | 7-day < 0.67 × target → investigate | ASTM C31/C192/C39 |
| Aggregate moisture | IS 2386 Pt 3 / IS 4926 §10.2.2 | 2× per shift per fraction | Probe-oven deviation > 1.0% → recalibrate | ASTM C566 |
| Admixture dose check | IS 9103 / pump calibration | Monthly linearity check | ±3% from linear fit; R² ≥ 0.998 | ASTM C494 |
Reference: IS 4926:2003 §10.2.2 (sampling & control); IS 456:2000 Cl 15.2.2 (sampling rate); IS 1199 series (fresh-concrete test methods); IS 516 Pt 5/Sec 1 (cube compression).
Slump — every load, every 7 m³ truck
IS 4926:2003 §10.2.2.1 requires concrete mixes to be randomly sampled and tested for workability, with corrective action where significant variations are detected. On a 60 m³/h plant with 7 m³ truck loads, that translates into a slump test on every truck. The test itself is two minutes: scoop, fill the cone in three equal layers, rod each layer 25 times with a 16 mm bullet-nosed rod, lift the cone vertically, read the subsidence.
The plant's action limit is ±25 mm from the mix-design target slump. Inside the band, log the value and dispatch. Outside, hold the load, check the controller recipe (moisture correction, admixture dose, batched masses), adjust, re-test, and only then release. Two consecutive out-of-band results is a process alarm — escalate to QA supervisor, check the load cells, check the admixture pump.
A shear slump (the mix shears sideways instead of subsiding) is an immediate reject — the result is not a valid slump number. Re-sample; if the second sample also shears, the mix design is the problem, not the load.
Concrete temperature — every load, with a 35 °C pause rule
Concrete temperature drives setting rate, slump loss, ultimate strength (slightly), and durability. IS 4926:2003 removed the concrete-temperature requirement from the 2003 revision of the standard, but practical concrete practice still treats 30–35 °C at placement as the comfort zone. ACI 305R and IS 7861 (Part 1) align:
- ≤ 30 °C placement — normal pours, no intervention.
- 32–35 °C placement — introduce chilled water, flake ice, or shaded aggregate stockpiles. Retarder dose up by 30–50%.
- 35–38 °C placement — pour only on structural members with hot-weather provisions; tighten curing discipline.
- ≥ 38 °C placement — pause pour. Cool the concrete, the mix, the trucks.
Measure at the discharge point of the truck or at the plant gate using a calibrated thermometer accurate to ±1 °C (ASTM C1064 parallel). Log the value on the delivery ticket — it ties the cube at 7 days back to the placement temperature that shaped it.
Cube sampling — IS 456 Cl 15.2.2 frequency
IS 456:2000 Cl 15.2.2 sets the sampling rate for cube compressive-strength acceptance, and it is a table of volume bands, not a single rate: 1 sample for 1–5 m³ placed, 2 for 6–15 m³, 3 for 16–30 m³, 4 for 31–50 m³, and above 50 m³ four samples plus one more for every further 50 m³ or part of it, with at least one sample from each shift. The one-sample-per-50-m³ rule that a plant actually runs on is a different one — it is IS 4926:2003 Cl 6.3.2, for ready-mixed production, together with one sample per grade per day, whichever gives the more frequent sampling.
Each sample is 3 cubes of 150 mm, cast in clean, dimensionally-checked moulds, compacted by rodding or vibration per IS 1199 (Pt 5):2018, and cured at 27 ± 2 °C until test age. The 7-day cube gives an early-strength trend; the 28-day cube gives the IS 456 Cl 16.1 / Table 11 acceptance number.
Sampling must be random — across the day, across the loads, not all from the morning's first three trucks. Composite sampling across two loads is permitted only with project-specific QA approval and is not the IS 456 default.
Air content — every 25 m³ when air-entrained
Most structural concrete in India is not air-entrained; freeze-thaw exposure is rare outside Himalayan belt pours. Where the mix is air-entrained (typically pavement, dam, or frost-exposed concrete), IS 1199 (Pt 4):2018 sets the pressure-method or volumetric method to measure the entrained air. Target is the mix-design air content (typically 4–6% for frost resistance) ±1.5%. Test every 25 m³ at the plant.
A drift in air content past the band is most commonly (a) a change in AEA dose, (b) a sand-fines change that affected AEA demand, or (c) over-mixing. Investigate before the next batch.
Worked example — daily QA schedule, 60 m³/h plant
Single 8-hour shift, 60 m³/h sustained, M30 as the dominant grade (~280 m³ total), one M40 order (40 m³), one SCC order (20 m³). Frequency vs. action limit on each line.
| Test | Frequency rule | Today's count | Action threshold |
|---|---|---|---|
| Slump (IS 1199 Pt 2) | Every load | ~50 tests | Target 100 ± 25 mm |
| Concrete temperature | Every load | ~50 tests | ≥ 32 °C → ice / chilled water; ≥ 38 °C → pause |
| Cube sampling (IS 4926 Cl 6.3.2) | 1 per 50 m³ M30 | 6 samples × 3 cubes = 18 cubes | 7-day < 0.67 × target → investigation |
| Cube sampling (M40) | 1 per 50 m³ | 1 sample × 3 cubes = 3 cubes | Same 7-day trigger |
| Cube sampling (SCC) | 1 per 50 m³ + extra flow/funnel tests | 1 sample + 2 flow + 2 V-funnel | Slump-flow 650 ± 50 mm; V-funnel 8–12 s |
| Aggregate moisture | 2× per shift × 3 fractions = 6 oven checks | 6 (sand + 20 mm + 10 mm × 2) | Probe-oven deviation > 1.0% → recalibrate |
| Density (M30) | 1 per shift per grade | 1 | ±2% of theoretical density |
| Yield reconciliation | End-of-shift | 1 | Yield factor 0.98–1.02 |
A QA team of one supervisor and two technicians covers this comfortably on an 8-hour shift. The supervisor owns the cube register and the moisture reconciliation; the technicians own the load-by-load slump, temperature, and density.
Reference: IS 456:2000 Cl 15.2.2 sampling rate; IS 1199 Pt 2 slump; IS 1199 Pt 3 density; IS 1199 Pt 5 making/curing specimens; IS 516 Pt 5/Sec 1 compression.
The integration loop — production QA → mix design
Plant QA is not a compliance chore; it is the plant's instrumentation. The data feeds back into the mix design on a weekly cadence at minimum:
- Cube trend — rolling 7-day mean and σ for each grade. If σ is creeping up, the aggregate grading or moisture correction is drifting.
- Slump trend — per-load and per-shift distribution. If the average is drifting away from target, the PCE dose or moisture correction needs tuning.
- Yield trend — batched mass vs. delivered mass. Persistent under-yield = moisture under-correction. Persistent over-yield = water over-dose.
- Returns rate — % of dispatched concrete returned. Above 2% is a dispatch or mix-design issue.
The same data, looked at monthly, drives the mix-design optimisation — cement content reductions where σ has tightened, admixture changes where retention is off, aggregate-source changes where grading variation is hurting the cube trend.
This loop is where a plant usually ends up on spreadsheets. CMDApp is our own platform for the concrete laboratory: its QA and QC module carries the frequencies and action limits above, the testing laboratory records each result against the standard and clause it was obtained to, and strength acceptance closes the loop back to the mix design on the same record.
Common deviations seen on Indian plants
Plant casts one 150 mm cube per "sample" because that's what they have moulds for. IS 456 requires 3 cubes per sample — test 1 at 7 d, 2 at 28 d. Single-cube sampling is non-compliant and any acceptance decision on it is unsafe.
Operator does the slump test, tosses the cone, walks back to the console, writes "110 mm" from memory. Actual reading was 70 mm. Cube result two weeks later is a surprise because the recorded slump said everything was fine.
Tank water not cooled, ambient temperature high. Cubes cure at 35–40 °C, develop higher apparent early strength, lower ultimate strength. The 7-day cube says "great"; the 28-day cube on the same batch, if cured properly, says otherwise.
Machine reads 5% high because the load-cell calibration drifted. Cubes pass acceptance at 32 MPa when the true strength is 30.4 MPa — both below fck. Wrong call on site acceptance.
Field-checklist blocks
A short questionnaire, an inspector's checklist, and three worked examples that show how the plant's four production tests resolve into a decision to dispatch, to hold the load, or to reject the batch.
Questionnaire — ask at the plant gate, before the truck is released
- Has slump been tested on this load, and is the reading inside ± 25 mm of the mix-design target?
- What is the concrete temperature at the discharge point, and are we already in the 32–35 °C cooling band or past 38 °C?
- How many cube samples has this grade taken today, and is the more frequent of 1 per 50 m³ and 1 per grade per day the one being met?
- Has the aggregate moisture been corrected for in the last shift, and what did the probe read against the oven?
- Is the admixture pump still inside its calibration, or is the monthly linearity check overdue?
- Is the sample being cast from a random point in the day's production, or from the first three trucks of the morning?
Inspector's checklist — what to verify at the plant
- Slump: tested every load, or 1 per 25 m³, target ± 25 mm, logged on the delivery ticket with the batch number, using the IS 1199 (Part 2):2018 procedure. A shear slump is a reject, not a number.
- Concrete temperature: measured on every load at the discharge point with a thermometer accurate to ± 1 °C, and the value written on the ticket. Cooling at ≥ 32 °C, pause at ≥ 38 °C.
- Cube sampling: 1 sample of 3 specimens of 150 mm per 50 m³ or per grade per day per IS 4926 Cl 6.3.2, sampled across the day and across the loads, cured at 27 ± 2 °C, with the sample identity tied to the batch ticket and the compression test run to IS 516 (Part 1/Sec 1):2018.
- Aggregate moisture: 2 checks per shift per fraction per IS 2386 Pt 3, probe against oven, and a deviation above 1.0% triggers recalibration before the next batch.
- Air content on air-entrained mixes: every 25 m³ per IS 1199 (Part 4):2018, within ± 1.5% of the mix-design target.
- Density: 1 per shift per grade per IS 1199 (Part 3):2018, within ± 2% of the theoretical density.
- Admixture pump linearity: monthly check, ± 3% from the linear fit and R² ≥ 0.998, with the certificate on file.
- End-of-shift reconciliation: batched mass against delivered mass with a yield factor of 0.98–1.02, the returns rate, and the slump, cube and yield trends written down for the weekly mix-design review.
What happens if… three worked examples
What happens. The 34 °C load is in the 32–35 °C band: chilled water, flake ice or shaded aggregate stockpiles come in, and the retarder dose goes up by 30–50%. The 39 °C load is past the 38 °C pause threshold by 1 °C, so the pour stops. The concrete, the mix and the trucks all get cooled before anything else moves.
Likely outcome. Both decisions are made at the gate, while the mixer is still in front of you, and that is the whole point. At the site end the same 39 °C concrete is already in the ground and the only options are re-tempering or rejection. One more thing follows from the same reading: if the curing tank is also running warm, at 38 °C the cubes develop higher apparent early strength and lower ultimate strength, so the 7-day cube says "great" and the 28-day cube on the same batch disagrees.
Preventive correction. Log the temperature on the delivery ticket. It is what ties the cube back to the placement conditions that shaped it.
What happens. The band runs 100 − 25 = 75 mm to 125 mm, so 60 mm is 15 mm below the lower limit. The load is held. You check the controller recipe (moisture correction, admixture dose, batched masses), adjust, re-test, and only then release it. A second consecutive out-of-band result turns it into a process alarm: escalate to the QA supervisor and check the load cells and the admixture pump.
Likely outcome. The truck does not leave until the reading is back in band, which is the difference between plant QA and site QA. At the site end, a 60 mm slump against a 100 mm target means the truck has already gone, and the options are re-tempering or rejection.
Preventive correction. Two consecutive out-of-band results is a recipe problem, not a bad batch of aggregate. Look at the moisture correction first; that is where the water went.
What happens. Divide the reading back out: 32 ÷ 1.05 = 30.5 MPa, so the true strength is in the region of 30.4 MPa. Both the recorded 32 MPa and the real 30.4 MPa sit below fck, yet the machine's output has carried the cube over the line. The lab's own record of a 32 MPa result is the false document, not the concrete.
Likely outcome. Every acceptance decision taken on that machine while the calibration had drifted is wrong, and they have to be traced back through the cube register and re-evaluated. The calibration certificate on the machine is the document that answers the question, and nobody checks it because the numbers have looked fine for months.
Preventive correction. Calibration currency is part of the plant QA checklist, not the lab's business. A 5% error does not show up as an obvious failure; it shows up as concrete that quietly does not meet its grade.
Frequently Asked Questions
How often must slump be tested at an RMC plant?
What is the cube sampling frequency per IS 456 at the plant gate?
At what concrete temperature must pouring be paused at the plant?
Related articles
IS 456 Cube Test Acceptance — Cl 16.1 / Table 11
The downstream gate every cube must clear — the two-mean and individual-result rule.
Acceptance Criteria — IS 456 Table 11 Explained
How the 0.825σ and fck − 3 N/mm² numbers are derived, and what to do when a sample fails.
Cube Testing — IS 516 Procedure
The casting, curing, capping and compression test method behind every cube result on the register.
Amit Haridas
Founder & Proprietor, ConcreteInfo. 25+ years experience in construction QA/QC, concrete technology, and RMC plant operations. NRMCA Certified Trainer (USA) and ISO Lead Auditor. For plant-QA/QC programme design, IS 1199 testing audits, IS 456 acceptance reviews, or operator training, 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.