IS 3535 : 1986Methods of sampling hydraulic cement

IS 3535:1986 specifies the methods of sampling hydraulic cement — how a *representative* sample is drawn from bags, bulk silos, wagons or a consignment for testing. It is the always-first step before any cement acceptance: the IS 4031/IS 4032 test results only describe the consignment if the sample was drawn per IS 3535.

It is read with the cement-acceptance stack:

• IS 4031 Part 1 (parts) — physical tests for hydraulic cement (run on the IS 3535 sample) · IS 4032 — chemical analysis
• IS 269 / IS 455 / IS 1489 — OPC / PSC / PPC (the product specs the result maps to)
• IS 456 — concrete (the structure that depends on accepted cement) · IS 4082 — site storage of cement

Cement varies — bag-to-bag, by silo stratification, by age — so IS 3535 fixes:

• Lot definition — cement of one type/grade/source/dispatch grouped as a lot
• Number & points of sampling as a function of lot size — for bagged cement, a number of bags selected at random across the stack; for bulk (silo/wagon), sampling at intervals through the flow/depth so stratified or aerated cement is represented
• Gross → composite: combine the increments into a representative composite, mix, and reduce (quartering) to the test quantity, with retains
• Sample handling: airtight containers immediately (cement absorbs moisture and carbonates fast — a poorly-stored sample false-fails on setting/strength), labelled (type, grade, lot, date), tested promptly

The engineering point: cement is the one ingredient whose defect propagates into *every* pour, and the test only protects you if the sample is representative — a few bags off the top of one pallet certifies nothing.

Scenario: acceptance of an OPC 53 consignment for structural concrete.

Step 1 — define the lot: group by type/grade/source/dispatch; split large deliveries into lots per IS 3535.

Step 2 — draw increments: for bagged cement, select the IS 3535 number of bags at random across the whole stack (top/middle/bottom, multiple pallets) and draw cement from within (not just the bag surface); for bulk, sample through the flow/depth.

Step 3 — composite & reduce: combine increments, mix, quarter down to the test quantity + retains.

Step 4 — protect: seal airtight immediately, label (OPC 53, lot, date), and test promptly — a sample left exposed absorbs moisture and will falsely fail setting/soundness/strength.

Step 5 — test & map: run IS 4031 physical (fineness, setting, soundness, strength) + IS 4032 chemical against IS 269; accept/reject the lot on that valid result. Skip Step 2/4 and you either false-reject good cement or pass bad cement into every structural pour.

1. Convenience sampling (a few top bags of one pallet). Not representative of the lot — the result certifies those bags, not the consignment.

2. No proper lot definition. Treating a huge mixed-dispatch delivery as one 'lot' (or each pallet as a lot) breaks the statistics.

3. Exposed/poorly-stored sample. Cement absorbs moisture and carbonates within hours — an unsealed sample false-fails setting/soundness/strength; seal airtight immediately and test promptly.

4. Surface-only sampling of bulk. Silo/wagon cement stratifies and aerates — sample through the flow/depth, not just the top.

5. Sampling/acceptance applied in isolation. The plan exists *so that* the IS 269 acceptance criteria are valid — sample per IS 3535, test per IS 4031, accept per the product spec, together.

• IS 4031 Part 1 (parts) — physical tests for hydraulic cement · IS 4032 — chemical analysis
• IS 269 — OPC (33/43/53) · IS 455 / IS 1489 / IS 8112 / IS 12269 — PSC / PPC / OPC-43 / OPC-53
• IS 456 — plain & reinforced concrete (acceptance dependency)
• IS 4082 — site stacking & storage of cement · IS 1070 — reagent water (test water)
• IS 5454 / IS 2530 — sampling of bricks / paints (parallel sampling logic)

IS 3535 is old (1986) and reaffirmed; it is invisible on site but it is the statistical foundation under every cement acceptance — and cement is uniquely consequential because its defects (unsoundness, flash set, low strength) propagate into *every* concrete pour, not one element. Disputed low-strength results very often unwind not to the cement or the test but to the *sample*: convenience-drawn, an undefined lot, or moisture-degraded before testing.

The practitioner contract is procedural and cheap: define lots properly, draw the IS 3535 number of increments at random across the whole stack/through the bulk depth, composite and reduce, seal airtight immediately, test promptly, and apply the result through the IS 269/IS 4031 acceptance chain. Witness the sampling — if the engineer doesn't, the 'representative' sample is whatever passes. It's the same lesson as IS 5454 (bricks) and IS 2530 (paint): a test only certifies a lot if the sample was drawn properly from a properly-defined lot.