M20 Slab Rejected on Cube Results

What the results say

• Sample result = average of the three cubes = (14.2 + 15.0 + 16.1) / 3 = 15.1 N/mm2 against a characteristic strength of 20 N/mm2. That is about 75 percent of requirement.
• For M20, the usual acceptance rules are: no individual sample result below fck minus 4 (i.e. below 16 N/mm2), and the running mean of four consecutive results at or above fck plus a margin based on the established standard deviation. With no established standard deviation on this site, the conservative check is simply that results must not sit below fck minus 4.
• Here the mean, 15.1, is below 16, and the lowest cube, 14.2, is well below. The sample fails both ways. It is a clear non-compliance, not a borderline case.
• The 7-day cubes at 11 to 12 N/mm2 (roughly 55 to 60 percent of 20) were already signalling a weak mix; normal OPC concrete should be around 65 to 75 percent at 7 days.

Why retesting the same cubes after another 28 days proves nothing

1. Cubes cured for only four days and then dried on the slab have lost the water needed for hydration. Their later gain is not the gain of the slab.
2. Concrete does not jump from 15 to 20 N/mm2 between 56 and 84 days; typical gain is a few percent.
3. The retest would answer a question nobody asked. The question is the strength of the slab in place, and cubes that were badly made and badly cured are not evidence either way.

Judgement: reject the retest request, but do not stop there. Accept the cube result as a warning about the slab and move to in-situ testing, which is the only evidence that can settle it.

Ranked causes

1. Water-cement ratio, uncontrolled. 'Until it looks right' with a hose is the single biggest strength killer. Every extra 0.1 w/c beyond about 0.5 costs roughly 15 to 20 percent strength. Confirmation: no slump records exist; start recording slump (for this work, 25 to 75 mm with 20 mm aggregate is a sensible target) and keep the mix consistent.
2. Cement partly hydrated in store. Bags stacked on the floor against a damp wall for seven weeks; hard lumps near the wall mean some bags have taken up moisture and lost activity. Confirmation: squeeze-test the bags, reject lumpy cement, check the store's stack height and dates.
3. Silt and dust in aggregate. Soapy-feeling sand and dusty 20 mm metal coat the particles so the paste cannot bond. Acceptable silt is low (a few percent); a field settling test in a jar should show a thin silt layer, not a thick one. Confirmation: jar settling test and sieve analysis on both sand and aggregate.
4. Volume batching with damp sand. Damp sand bulks by 20 to 30 percent, so a 'pan' of sand holds far less solid than the nominal 1.5 parts; the mix drifts fines-rich or cement-poor depending on who is shovelling. Confirmation: weigh a pan of sand dry and damp, and compare.
5. Inadequate curing. Four days in a tank then drying in the sun on the slab - hydration stops at the surface. This explains part of the cube loss and will cause surface dusting and shrinkage cracks in the slab.
6. Cube making and handling. No compaction standard, no density check. Low cube density (below about 2.3 g/cc) would suggest poor compaction rather than a weak mix.

Ordering logic: causes 1, 2 and 4 are visible in the record and likely to dominate; 3 and 5 amplify them. Do not blame the lab or the cubes first - the practice on site is the prime suspect.

Action on the slab - evidence, not argument

1. Reject the 'retest of cubes' route. Document the failure in the test register.
2. Non-destructive screening: rebound hammer and ultrasonic pulse velocity on a grid over the slab, to find weak zones. Neither gives strength by itself, but they locate the areas worth coring.
3. Core testing: take cores from the weak zones, at least three per 100 m2 or part, avoiding reinforcement, and test. Judge in-situ strength against the usual acceptance band - cores averaging at least 85 percent of fck (17 N/mm2) with no core below 75 percent (15 N/mm2) is the normal working rule; anything lower needs structural review.
4. If cores fail, refer the slab to the structural engineer for a load test or a design check; do not quietly accept it because the building is single-storey.
5. Meanwhile, do not load the slab, and record any cracking at the surface as it appears.

Changes before the next pour

• Cement store: stack on timber pallets at least 150 mm clear of the floor, 300 mm clear of walls, maximum 8 to 10 bags high, first-in first-out, and use within 90 days. Reject lumpy bags.
• Aggregate: wash the 20 mm metal and test the sand for silt before it enters the mixer.
• Batching: move to weigh batching, or at minimum use calibrated volume boxes and apply a bulking correction for damp sand.
• Water: fix the maximum free w/c (about 0.5 for M20) and control it with a measured can plus a slump cone; the hose is not a measuring instrument.
• Curing: wet cure for at least 7 days (10 for blended cement); keep cubes in water for 28 days.
• Cubes: cast six per sample, record cube density, store and transport them without damage.

The judgement to defend is this: strength is decided by batching, water and curing, and no argument about retesting changes what was actually built.