Concrete Quality: Slump, Cubes and Acceptance

The pour finished at 6 pm. At 6:30 someone asks the question that should have been asked at 9 am: "Cube liya kya?" No cubes were cast. Now 45 m3 of M25 sits in a slab with no strength evidence, the structural consultant is unhappy, and the client is entitled to hold the entire item — or demand core tests at your cost. Concrete quality control is not complicated, but it is unforgiving of afterthoughts. This lesson covers the three things a QS must command cold: slump (the right workability for the work), sampling (how many cubes, when), and acceptance (what makes a result pass under IS 456).

Slump: the right workability for the placing condition

The slump test (per IS 1199) measures workability — how easily concrete flows and compacts. More slump is not better concrete; extra water bought workability at the price of strength. IS 456 (clause 7.1) suggests workability by placing condition:

Two field realities to layer on top:

• Pumped RMC is commonly ordered at 100–120 mm slump on Indian sites — above the code's suggested band for pumping — achieved with plasticisers at a controlled water-cement ratio. The ordered slump comes from the approved mix design, not from the pump operator's preference. What is never acceptable is hitting that slump by adding water at site.
• Delivery tolerance: IS 4926 (ready-mixed concrete) allows the delivered slump to vary from the ordered value by ±25 mm or one-third of the specified value, whichever is less. Order 100 mm and 80 mm arrives: within tolerance. 140 mm arrives: you are entitled to reject the truck — and should, because that truck likely carries extra water.

Never allow water addition at site to revive a stiff load. Every added litre raises the water-cement ratio of concrete whose cubes will carry your pour card number. If a truck is unworkably stiff, send it back to the plant.

Sampling: how many cubes, from where

IS 456 (clause 15.2.2) fixes the minimum number of samples per grade of concrete per day:

The rules around the table matter as much as the table:

• One sample = three 150 mm cubes, tested at 28 days (per IS 516). The test result is the average of the three; if any individual cube deviates from that average by more than ±15 percent, the sample is invalid.
• At least one sample per shift where concreting is not continuous.
• Sampling must be random and spread across the pour — not three scoops from the friendliest truck.
• Cubes for 7-day testing, formwork striking, or curing decisions are cast in addition to the 28-day set. The common site practice of 6 cubes per sample = 3 for 7 days + 3 for 28 days.
• 7-day results are an early warning, not an acceptance criterion. OPC concrete typically shows on the order of two-thirds of its 28-day strength at 7 days, but acceptance under IS 456 rests on 28-day results — and blended cements (PPC/PSC, now the default in most Indian markets) gain strength slower early and catch up later. Do not condemn or accept concrete on 7-day numbers.

For our 45 m3 pour: the table row 31–50 requires 4 samples = 12 cubes for 28-day testing (plus 12 more if the site also wants 7-day indications).

Acceptance: what "pass" means under IS 456

Acceptance for compressive strength (IS 456 clause 16 with Table 11, as amended) has two simultaneous conditions:

1. Group mean: the mean of any group of four non-overlapping consecutive test results must be at least the greater of — fck + 0.825 x established standard deviation (rounded to the nearest 0.5 N/mm2), or fck + 3 N/mm2.
2. Individual: no single test result falls below fck − 3 N/mm2.

Until the site has enough results (30) to establish its own standard deviation, IS 456 Table 8 permits an assumed value — 4.0 N/mm2 for M20 and M25.

Small pours are stricter, not looser. Where the quantity is up to 30 m3 and fewer than four samples exist, the amended Table 11 notes require the mean of all samples to be at least fck + 4 N/mm2 with no individual result below fck − 2 N/mm2; with only one sample, that sample must reach fck + 4 N/mm2. So a 4 m3 chajja pour of M25 needs its single sample to hit 29 N/mm2 — a common surprise.

Worked example: the 45 m3 M25 slab

Four samples were cast; the 28-day test results (each already the average of 3 cubes) are 29.5, 31.0, 27.5 and 30.0 N/mm2.

Both conditions pass: the concrete is accepted. Note what would have happened with one result at 21.5: the individual check fails regardless of a healthy mean, and the lot is non-compliant.

When a result fails

A failed cube result is not an automatic demolition order — and not something to bury either:

1. Verify the evidence first. Check the ±15 percent within-sample validity, the cube register trace, curing of the cubes themselves (site-cured cubes in a dry corner routinely under-report strength).
2. The engineer-in-charge investigates — typically non-destructive testing (rebound hammer, ultrasonic pulse velocity) followed by core tests if needed, per the code's framework. Concrete may be accepted, accepted with strengthening, or rejected based on that investigation.
3. The QS's job is money and records: hold billing for the affected element, record the non-conformance, and capture the cost trail — core tests, consultant reviews, strengthening — against whoever's failure caused it. A missing cube record converts a winnable technical discussion into a lost commercial one.

The QS angle: strength results are billing documents

Cube results decide when money moves. Standard practice worth adopting: RA bills for concrete items are certified provisionally on pour cards and 7-day indications, with a defined holdback released on satisfactory 28-day results. The cube register — sample ID, element, casting date, testing date, three cube values, average, pass/fail — is therefore a financial ledger as much as a technical one. Keep it current, keep it cross-referenced to pour cards, and reconcile it monthly against the concrete billed.

What testing costs vs what it protects

Run the rupee comparison once and you will never skip a cube again (indicative mid-2026 figures; use local lab quotations):

A few thousand rupees of cubes is the insurance premium on lakhs of structure. This is also your dispute armour in both directions: for an owner, a passing cube register is what forces a reluctant thekedar's RMC supplier to own a bad batch; for a contractor, it is what stops a client from holding a lakh-scale RA bill on a vague "quality doubt." No register, and the party holding the money wins by default.

On site-mix projects without a nearby lab, cubes still get cast — most district headquarters have a polytechnic, engineering college or private lab that tests cubes for a small fee; build the transport into the routine. And on hand-mixed pours, watch the farma, not just the cube: a gauge box shaved 2 inches short leans every batch, and the cube you cast from the one honest batch will not save the ninety dishonest ones.

Common mistakes

• Casting cubes but not labelling them with sample ID and date while the concrete is green — anonymous cubes prove nothing.
• Testing all cubes at 7 days "to know early," leaving nothing for the 28-day acceptance test.
• Treating 7-day = 67 percent as a rule for PPC concrete and panicking at a healthy slow-gaining mix.
• Averaging away a failed individual result across pours — the individual criterion exists precisely to stop this.
• Accepting a wet truck because the pump was waiting; rejecting paperwork is cheaper than rejecting a slab.

What comes next

Cubes and slump control the concrete you meant to place. The next lesson is about the concrete, mortar and labour you never meant to spend at all — the quiet execution errors, measured in millimetres, that add up to lakhs.