Codes referenced
Sampling, sample reduction, sieve analysis, flakiness index, elongation index.
Aperture tolerances, frame dimensions, designation of the sieves used in IS 2386 Pt 1.
Names IS 2386 Pt 1 as the test method and sets the acceptance limits (Zone I-IV for fine, grading envelopes and FI/EI ≤ 25% for coarse).
Both read sieve-analysis output from IS 2386 Pt 1 to confirm aggregate suitability and to size the trial mix.
Sampling — the step that decides everything else
Every result downstream — sieve curve, FI, EI, AIV, silt content — is only as good as the sample it sits on. IS 2386 Pt 1 §2 is unusually direct about this: no result from a non-representative sample is valid. A grading curve on a 200 g grab from a stockpile is decorative; a curve on a riffle-reduced sample from the conveyor stream is auditable.
Sample size — Table IV of IS 2386 Pt 1
IS 2386 Pt 1 Table IV sets the minimum weight of sample for sieve analysis by the maximum size of aggregate present in substantial proportions. Sample size is matched to nominal size — not by a multiple rule, but by the explicit table below.
| Maximum size present (mm) | Minimum weight of sample for sieving (kg) |
|---|---|
| 63 | 50 |
| 50 | 35 |
| 40 or 31.5 | 15 |
| 25 | 5 |
| 20 or 16 | 2 |
| 12.5 | 1 |
| 10 | 0.5 |
| 6.3 | 0.2 |
| 4.75 | 0.2 |
| 2.36 | 0.1 |
Reference: IS 2386 (Part 1):1963 Table IV (Clauses 2.4.4, 2.5 and 2.5.2). The sample must be brought to air-dry condition before sieving; air-dry at room temperature or heat at 100–110 °C.
Reduction by riffle divider
From the bulk field sample (often 25-50 kg) down to the test portion requires progressive reduction. The riffle divider is the only method IS 2386 Pt 1 endorses for routine work; quartering by hand is permitted only when no riffle is available, and even then only for coarse aggregate.
Stream the sample into the divider
≥ 12 equal-width slots
Pour the bulk sample slowly along the full length of the riffle so each slot receives a representative increment. Do not dump from a height; the fall segregates fines.
Discard alternate pans
Cut sample in half each pass
Two pans sit under the slots. Keep one pan, discard the other. Each pass halves the mass. Repeat until the test portion reaches the required size in Table 1.
Record chain of custody
Source · lot · date · sampler · reduction method
A sieve-analysis sheet with no source lot ID and no sampler initials is not a sieve-analysis sheet. IS 2386 Pt 1 requires the metadata; IS 383 acceptance criteria apply to a sample whose origin you can trace.
Sieve analysis — the grading curve
IS 2386 Pt 1 §3 specifies sieve analysis on a stack of IS 460 test sieves, with the test portion sized per Table 1. The standard sieve set for concrete aggregate, top to bottom: 20 mm, 12.5 mm (or 10 mm), 4.75 mm, 2.36 mm, 1.18 mm, 600 µm, 300 µm, 150 µm, plus a bottom pan. The IS 383 grading envelope is read against the cumulative % passing each sieve.
Mechanical vs hand sieving
| Aspect | Mechanical (preferred) | Hand (acceptable) |
|---|---|---|
| Endpoint criterion | 10–15 min on shaker (varies by apparatus) | ≤ 1% by mass passes in 1 min continued sieving |
| Reproducibility | High — operator barely matters | Low — depends on shaker technique |
| Field use | Site lab with stable power | Quick site check or power outage |
| Equipment | Sieve shaker + full IS 460 set | Sieves + dust-tight lid + receiver pan |
| Record keeping | Time on shaker, frequency setting | Operator, duration, sieve count |
Reference: IS 2386 Pt 1 §3. End-point for hand sieving is "no more than 1% by mass of the residue passes in 1 minute of continued shaking." Mechanical endpoint is set by the apparatus manufacturer and validated against this criterion.
Calculation — cumulative % passing
For each sieve, weigh the mass retained. Cumulative % retained is the running sum of retained masses as a % of the total. Cumulative % passing = 100 − cumulative % retained.
Cumulative % passing = 100 × (1 − Σ Ri / Mtotal)
where Ri is the mass retained on sieve i (and every coarser sieve above it) and Mtotal is the sum of all retained masses plus the bottom-pan mass. Loss of more than 1% of the test portion during sieving invalidates the test.
Source: IS 2386 Pt 1 §3. Always plot grading against IS 383:2016 Table 2 (coarse) or Table 4 (fine) envelopes.
Flakiness index & elongation index
Two shape indices, two gauges, one shared calculation. IS 2386 Pt 1 §6 defines FI and §7 defines EI on the same sample split. IS 383:2016 limits each to 25% for concrete aggregate.
Definition: percentage by weight of particles whose least dimension (thickness) is less than 0.6 × mean of the two larger dimensions.
Apparatus: standard thickness gauge with parallel slots of width 0.6 × the mean dimension of each fraction (sizes 63, 50, 40, 31.5, 25, 20, 16, 12.5, 10, 6.3 mm).
Procedure: separate the sample into size fractions on the standard sieves. Pass each fraction by hand through the matching slot. Particles that pass are "flaky"; weigh them.
Calculation: FI = (Σ flaky mass / Σ total mass of all fractions) × 100.
Definition: percentage by weight of particles whose greatest dimension (length) is greater than 1.8 × mean of the two smaller dimensions.
Apparatus: standard length gauge — a plate with circular pegs at spacings of 1.8 × the mean dimension of each fraction.
Procedure: test each size fraction from the FI test (or a parallel split). Particles retained between the pegs that cannot pass through are "elongated"; weigh them.
Calculation: EI = (Σ elongated mass / Σ total mass of all fractions) × 100.
Field acceptance — what to actually do at site
Sample from the conveyor discharge or chute outlet — never from a stationary pile. Use a riffle divider for reduction. Label every test portion with source, lot, date and sampler. Read the sieve curve against the IS 383 grading envelope before accepting the delivery.
Quarter by hand on the floor. Use a stockpile grab as a "sample". Skip the riffle because the lab will redivide anyway. Accept a grading sheet with no chain-of-custody metadata. Reject a delivery on the truck driver's word that "this is Zone II sand."
Worked example — coarse aggregate grading + FI + EI
Scenario. 20 mm nominal crushed basalt, test portion 5 kg, sieved per IS 2386 Pt 1 §3.
| IS sieve | Mass retained (g) | Cumulative retained (g) | Cumulative % retained | Cumulative % passing | IS 383 20 mm band |
|---|---|---|---|---|---|
| 40 mm | 0 | 0 | 0 | 100 | 100 (pass) |
| 20 mm | 150 | 150 | 3 | 97 | 95–100 (pass) |
| 10 mm | 1850 | 2000 | 40 | 60 | 25–55 (pass) |
| 4.75 mm | 2700 | 4700 | 94 | 6 | 0–10 (pass) |
| 2.36 mm | 200 | 4900 | 98 | 2 | 0–5 (pass) |
| Pan | 100 | 5000 | 100 | — | — |
Grading — pass. All sieves within the IS 383 20 mm nominal envelope. Cumulative loss (pan) is 2% of the test portion, well within tolerance.
FI on the same sample split. Particles retained between 20 mm and 10 mm sieves = 2000 g. Slot gauge for this fraction: 12.5 mm slot. Particles passing the slot = 480 g. Apply the same procedure to each fraction (10-4.75 mm: slot 6.3 mm; 4.75-2.36 mm: slot 4 mm — often omitted).
| Fraction | Slot width | Fraction mass (g) | Flaky mass (g) |
|---|---|---|---|
| 20.0–10.0 mm | 12.5 mm | 1850 | 480 |
| 10.0–4.75 mm | 6.3 mm | 2700 | 620 |
| Total | 4550 | 1100 |
FI = (1100 / 4550) × 100 = 24.2%. Below the IS 383 limit of 25% — pass, but borderline. The same procedure applied to the EI rod gauge gives EI = 19.5% — comfortable pass.
Common deviations seen on Indian sites
Hand quartering segregates coarse particles outward and concentrates fines at the centre. The sample drawn is no longer representative. IS 2386 Pt 1 §3 prefers a riffle divider — quartering is permitted only for coarse aggregate when no riffle is available.
A 500 g sieve analysis on a 20 mm nominal aggregate is half the IS 2386 Pt 1 minimum. The result will be biased because the largest particles were not represented in proportion. Match the test portion to the nominal size — 2 kg minimum for 20 mm, and 15 kg for 40 mm.
Sieve results without source, lot, date and sampler initials are technically non-compliant under IS 2386 Pt 1 and difficult to defend in a dispute. Always label the test portion and the sheet.
Sieve shaker endpoint depends on amplitude, frequency and run time. Apparatus without a service log drifts and produces inconsistent results. Verify against a reference sample every quarter.
The thickness gauge has a slot for each fraction. Using the 12.5 mm slot on a 4.75-10 mm fraction under-counts flakes; using the 6.3 mm slot on a 10-20 mm fraction over-counts. Match slot to fraction.
A 600 µm sieve with a torn mesh passes more material than the standard aperture. Visual inspection of every sieve at the start of each test is cheap; replacing a damaged sieve is cheaper than accepting a wrong grading curve.
Field-checklist blocks
A short questionnaire, an inspector's checklist, and three worked examples that show how a sieve analysis and a pair of shape indices are reduced from a bag of aggregate to a single accept-or-reject decision that will survive a dispute.
Questionnaire — ask when the bulk sample reaches the laboratory
- What is the nominal maximum size of the aggregate, and does the test portion meet the IS 2386 Pt 1 Table IV minimum for that size?
- Was the bulk field sample taken from a moving stream, and how many increments went into it? A grab off a standing stockpile is not a sample.
- How was it reduced — a riffle divider with ≥ 12 equal slots, or quartering? Quartering is permitted only for coarse aggregate, and only when no riffle is available.
- How many riffle passes were made, and is the reduction method written on the sheet?
- Was the sample brought to air-dry condition before sieving, and how: room-temperature drying, or heating at 100–110 °C?
- Is the sieve set an undamaged IS 460 (Part 1):1985 set, inspected before the run, and does the shaker endpoint sit on a service log?
- Which fractions did the thickness gauge cover, and does each slot width match the mean dimension of that fraction rather than the fraction above it?
- Who took the sample, from which lot, on what date — and is all of that on the sheet?
Inspector's checklist — what to verify on site
- Chain of custody on the label and the sheet: source, lot, date, sampler, and the reduction method.
- Test portion mass weighed and measured against the Table IV minimum for the maximum size present, and the weighing recorded.
- Riffle divider with ≥ 12 equal-width slots, the sample poured slowly along the full length rather than dumped, alternate pans discarded, and the number of passes counted.
- Air-dry condition confirmed before the sample goes on the stack.
- Complete sieve set in order: 20 mm, 12.5 mm (or 10 mm), 4.75 mm, 2.36 mm, 1.18 mm, 600 µm, 300 µm, 150 µm, plus the bottom pan. Every mesh inspected for damage at the start of the run.
- Retained masses recorded for each sieve and the pan, and the total reconciled against the weighed test portion: a loss of more than 1% invalidates the test.
- Cumulative % passing computed as 100 × (1 − Σ Ri / M total) and every sieve plotted against the IS 383:2016 envelope, Table 2 for coarse aggregate or Table 4 for fine.
- FI and EI on the same sample split, slot and rod gauge matched to each fraction, both reported separately against the 25% limit.
What happens if… three worked examples
What happens. The 5 kg test portion is comfortably above the 2 kg Table IV minimum for 20 mm nominal aggregate, so the size is not the problem. The sheet shows 150 g retained on the 20 mm sieve, 1 850 g on the 10 mm, 2 700 g on the 4.75 mm, 200 g on the 2.36 mm and 100 g in the pan, and those five figures make up the 5 000 g portion. Recompute the 10 mm row from the weights: cumulative retained is 150 + 1 850 = 2 000 g, so 2 000 ÷ 5 000 × 100 = 40% retained and 100 − 40 = 60% passing. The IS 383:2016 Table 2 envelope for 20 mm nominal aggregate puts the 10 mm sieve at 25–55% passing, and 60 − 55 = 5 percentage points sits outside the band.
Likely outcome. One row is outside the band and four are inside, which is precisely how a non-compliant lot gets signed off: the curve looks clean and the one failing sieve is easy to miss because the other rows agree with it. Recompute the whole column from the retained masses, compare every sieve with the band, and if the weights cannot produce a compliant curve, rerun the test on a fresh portion rather than adjusting the sheet.
Preventive correction. Keep the sieve masses and the weighed test portion on the same sheet so the reconciliation is a line of arithmetic, not a memory.
What happens. IS 2386 Pt 1 §3 invalidates the test when more than 1% of the test portion is lost during sieving. Take a 5 000 g test portion where the recorded masses are 150 g on 20 mm, 1 850 g on 10 mm, 2 700 g on 4.75 mm, 200 g on 2.36 mm and 40 g in the pan. Those figures add to 4 940 g, so the shortfall against the weighed portion is 60 g, and 60 ÷ 5 000 × 100 = 1.2%.
Likely outcome. 1.2% is over the limit, so the sieve analysis never happened as far as the standard is concerned and the sheet cannot be used to accept or reject the lot. The error also moves the curve: dividing the cumulative retained at 2.36 mm (4 900 g) by the wrong total of 4 940 g gives 99.2% retained and 0.8% passing, against the 2% passing the correct 5 000 g total produces. Reconcile the weighed test portion against the sum of every sieve and the pan before you plot a single point, and repeat with a fresh portion if it does not balance.
Preventive correction. Write the weighed test portion mass at the top of the sheet and total the pans at the bottom, so the difference is visible to anyone who picks the sheet up.
What happens. The shape test totals 4 550 g of fraction mass across two fractions with 1 100 g of it flaky, so FI = 1 100 ÷ 4 550 × 100 = 24.2%. That clears the IS 383 limit of 25% by 0.8 of a point, and EI on the same split reads 19.5%, giving a combined off-shape share of 24.2 + 19.5 = 43.7%. But the 4.75–2.36 mm fraction, 200 g of it, was left out of the FI test, and that is the fraction flagged as often omitted.
Likely outcome. Put the omitted fraction back and the decision can swing either way. If none of those 200 g is flaky, FI = 1 100 ÷ 4 750 × 100 = 23.2% and the lot passes. If all of it is flaky, FI = 1 300 ÷ 4 750 × 100 = 27.4% and the lot fails the 25% limit. That 4.2 point spread is the entire pass-or-fail, and it closes by running every fraction the gauge has a slot for and reporting the total.
Preventive correction. Record the slot width and the fraction it was used on, so an omitted fraction is visible on the sheet instead of hiding in the arithmetic.
Frequently Asked Questions
What is the minimum mass of a test portion for sieve analysis under IS 2386 Pt 1?
How is the flakiness index defined and what is the IS 2386 Pt 1 limit?
Can a riffle divider be replaced by quartering for aggregate sampling?
What is the difference between mechanical and hand sieving?
Related articles
Aggregates for Concrete — IS 383:2016 Field Walkthrough
The acceptance limits. This post is the method behind the grading sheet.
IS 10262:2019 Mix Design Walkthrough
Proportioning concrete from aggregate properties, target strength and exposure conditions.
Concrete Materials — Field Handbook
Cement, water, admixtures and aggregates — what to check at the gate, what the lab catches.
Cube Testing Procedure — IS 516
Sampling, casting, curing and acceptance of concrete cubes against IS 456 Table 11.
Amit Haridas
Founder & Proprietor, ConcreteInfo. 25+ years experience in construction QA/QC, concrete technology, RMC plant operations, consulting and technical training across India. For a project-specific grading or aggregate-validation review, 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.