A ground-floor slab in Cavite, a 28 MPa mix approved at a water-cement ratio of 0.50, and 28-day cylinders back at 22 to 24 MPa. The investigation found nothing: same cement, same pit, tank-cured cylinders, a calibrated machine. Nobody had written down that it rained for two days beforehand and the sand stockpile was soaked.
The mix was 0.50 on paper and close to 0.58 in the truck. Every other line in the QA file was correct and the concrete still came up short, because the one quantity that governs strength was never delivered. This is the most common reason a compliant paper mix produces marginal cylinders, and the fix is arithmetic. One caveat on that opening result: the 22 to 24 MPa is attributed to the ratio drift, not derived from it. Nothing in this article converts a ratio into a strength, so read those cylinders as an assumed consequence of a known cause rather than a calculated one.
A trial mix is proportioned on aggregate in the saturated surface dry condition: pores full, outside of the grain dry. Such an aggregate neither gives water to the paste nor takes water from it, so the design water is exactly the water in the paste.
A stockpile is never in that condition. The sheet says 190 kg of water, the batchman adds 190 kg, and the mixer receives 190 kg plus whatever the sand carried, minus whatever the gravel drinks. Nobody deviated from the approved mix; the arithmetic step between design sheet and batch plant was simply never performed.
Absorption is water drawn into the grain's own pores. It belongs to the aggregate, not the weather, and it is not available to the paste. Free moisture, or surface moisture, is the film on the outside, and for batching it is mix water.
Free moisture = total moisture content - absorption. Keep the sign. Positive means the aggregate brings water in; negative means it is drier than saturated surface dry and will steal water from the paste.
Stock (as-delivered) weight = saturated surface dry weight × (1 + free moisture as a decimal).
Water to add = design water - the sum of the free water carried by every aggregate.
That is a mass balance, not an empirical correlation: no calibration constant, no range of validity. It is either done or not done.
Design per cubic metre, saturated surface dry basis: cement 380 kg, water 190 kg, fine aggregate 720 kg, coarse aggregate 1,050 kg, so the design ratio is 190 ÷ 380 = 0.50. With cement as the only cementitious material here, water-cement ratio and water-cementitious materials ratio are the same number; with fly ash or slag in the mix they are not, and it is the water-cementitious figure that a specification limits. Site condition: sand total moisture 6.5 percent, absorption 1.2 percent; gravel total moisture 0.4 percent, absorption 1.0 percent. The gravel is drier than saturated surface dry, which is normal in dry weather and exactly the case people ignore.
Sand: 6.5 - 1.2 = +5.3 percent. Gravel: 0.4 - 1.0 = -0.6 percent.
Sand: 720 × (1 + 0.053) = 720 × 1.053 = 758.16 kg.
Gravel: 1,050 × (1 - 0.006) = 1,050 × 0.994 = 1,043.70 kg.
Sand: 758.16 - 720 = +38.16 kg into the mix.
Gravel: 1,043.70 - 1,050 = -6.30 kg, absorbed out of the mix.
Net free water: 38.16 - 6.30 = 31.86 kg.
Water to add = 190 - 31.86 = 158.14 kg. Taking water at an assumed density of 1,000 kg per cubic metre, that is 158.1 litres. Batch the fuller figure rather than rounding down to 158, because rounding water down is precisely what leaves a mix stiff and puts a hose on the chute.
Check: 158.14 added plus 38.16 on the sand gives 196.30 kg in the drum; the gravel absorbs 6.30, leaving 196.30 - 6.30 = 190.00 kg of free water, and 190.00 ÷ 380 = 0.50.
Water in the paste = 190 + 38.16 - 6.30 = 221.86 kg. Actual ratio = 221.86 ÷ 380 = 0.584. That is 221.86 ÷ 190 = 1.168, roughly 17 percent more water than designed, and 0.084 above the approved ratio.
Be clear about which failure that models: the aggregate weights were corrected and batched at 758.16 and 1,043.70 kg, and only the water was left at the design 190 kg. The opening story is the other failure, where nothing is corrected at all. Batch 720 kg of wet sand and 1,050 kg of dry gravel and the sand delivers only about 683.8 kg of saturated surface dry solids, a shortfall near 36 kg per cubic metre; free water is 36.24 kg, absorption 6.34 kg, paste water 219.90 kg and the ratio 0.579. The conclusion is the same, the number is slightly different, and there is a second defect — short sand — that the yield check below does not cover.
The direction is certain even though the magnitude is not something to guess at. The empirical relationship usually credited to Abrams holds that for given materials, age and curing, compressive strength falls as the water-cement ratio rises. Its published limits matter: it applies to fully compacted concrete only, it breaks down at ratios too low to consolidate, and its constants must come from trial mixes on your own materials rather than a generic curve. So 0.084 of extra ratio will lower 28-day strength and raise permeability, by an amount your trial data can quantify and a textbook chart cannot. It breaches the approved mix design ratio outright, and it breaches any specified maximum water-cementitious materials ratio for durability that sits at or below 0.58. The first of those is a non-conformance whatever the cylinders read.
Corrected batch mass = 380 + 158.14 + 758.16 + 1,043.70 = 2,340.00 kg, against a design total of 380 + 190 + 720 + 1,050 = 2,340.00 kg. Identical, because a correct correction only moves water between the aggregate column and the water column. A unit mass near 2,340 kg per cubic metre is reasonable for normal-weight concrete.
By absolute volumes, using illustrative relative densities of 3.15 cement, 2.60 saturated surface dry sand, 2.65 saturated surface dry gravel: 380 ÷ 3,150 = 0.12063 m³; 190 ÷ 1,000 = 0.19000 m³; 720 ÷ 2,600 = 0.27692 m³; 1,050 ÷ 2,650 = 0.39623 m³. Sum = 0.98378 m³, leaving 1.00000 - 0.98378 = 0.01622 m³, or 1.62 percent entrapped air. Use your own measured densities.
Uncorrected, paste water becomes 221.86 ÷ 1,000 = 0.22186 m³ and the total is 0.12063 + 0.22186 + 0.27692 + 0.39623 = 1.01564 m³. Holding the same air volume, the batch makes about 1.0319 m³, an over-yield of about 3.2 percent, so cement content falls from 380 to 380 ÷ 1.032, about 368 kg per cubic metre. More concrete than you ordered, and every cubic metre of it weaker.
None of the following is measured. Each is an input the worked example takes for granted, and each would move a number if your job differed.
Strictly, free moisture referred to the saturated surface dry mass is (total moisture - absorption) ÷ (1 + absorption), because both percentages are reported on the oven-dry mass. Run rigorously, the example gives 757.71 kg of sand, 1,043.76 kg of gravel, net free water 31.47 kg and water to add 158.53 kg. The shortcut differs by 0.39 kg, shifting the ratio by 0.001 — noise beside the 0.084 error it prevents, but say it before someone says it to you.
Oven or hotplate dry-back. The reference method, one line of arithmetic. Weigh a representative wet sample, dry to constant mass, weigh again. Wet mass 1,065.0 g, constant dry mass 1,000.0 g, so total moisture = (1,065.0 - 1,000.0) ÷ 1,000.0 × 100 = 6.50 percent on the oven-dry basis. Constant mass means two successive weighings agree within a stated tolerance, not that the sand looks dry.
Calcium carbide gas-pressure tester, what most sites call a speedy moisture meter. Fast, and good for tracking drift between shifts, with two cautions. Its published standard test method is written for soils, so confirm your specification accepts it for aggregate. And its dial normally reads on the wet mass basis while the arithmetic needs the dry basis: 6.10 percent wet becomes 6.10 ÷ (100 - 6.10) × 100 = 6.10 ÷ 93.90 × 100 = 6.50 percent dry.
Microwave dry-back. The fastest field expedient: short bursts with a weighing between each, stopping when the mass stops changing. Recognised in some agency test suites and not others, and it can overheat and fracture porous aggregate, so correlate it against the oven on your own material and use it to track drift rather than replace the reference test.
One morning reading does not hold all day. A stockpile drains under gravity, so the base is wetter than the crown; by afternoon the exposed face has dried while the interior has not; a fresh delivery tipped on top resets everything; and the loader opening a new face changes the number without anyone touching the plant. Fine aggregate swings hardest, because its surface area per kilogram is the largest.
All of this collapses if the pump operator opens the hose at the chute because the mix looks stiff. One act can add tens of litres per truck, undoing a correction nobody outside the plant knows was made.
The ready-mixed concrete specification in general use, ASTM C94/C94M, handles this by making the addition visible. It requires the delivery ticket to record water added at the site and the initials of whoever directed it, permits at most a single addition of water at the purchaser's request, before discharge has substantially begun, and only when slump is below the specified value, forbids exceeding the specified maximum water-cementitious materials ratio or the specified slump, and requires the specified further mixing at mixing speed afterwards. ASTM C94/C94M itself also limits discharge to 1.5 hours or 300 drum revolutions after mixing water is introduced, whichever comes first — a limit the purchaser may expressly waive when the concrete is still placeable without added water. Your project documents may tighten it, so read both. In practice: a ticket line for litres added, who authorised it and the time, then the QA/QC engineer signing corrected batch weights and site water on the same sheet. Unrecorded water hurts twice, once in the cylinders and again in the argument afterwards.
Some plants react to high moisture by adding cement instead of subtracting water. Holding 0.50 against 221.86 kg of water needs 221.86 ÷ 0.50 = 443.72 kg of cement, an extra 443.72 - 380 = 63.72 kg per cubic metre. At an illustrative ₱280 per 40 kg bag, or ₱7.00 per kg, that is 63.72 × 7.00 = ₱446.04 per cubic metre; over a 60 m³ pour, 60 × 446.04 = ₱26,762.40, on top of 60 × ₱4,800 = ₱288,000 of concrete at an illustrative supply rate. Those peso figures are illustrative, not quotations. The extra cement still does not fix the yield, and it raises paste volume, heat of hydration and drying shrinkage. Subtracting 31.86 kg of water costs nothing.
If corrected batch weights carry through into a quantity and cost sheet, keeping the take-off, the rates and the yield in one place is what stops the volume you priced from drifting from the volume you batched. The RHCES Estimator is built for that kind of running sheet.
That 0.6 percent is 6.30 kg of water per cubic metre, which is 6.30 ÷ 190 = 3.32 percent of the design water and roughly 0.017 of ratio. Alone it may be tolerable; combined with a sand error in the same direction it is not, and it decides whether the concrete arrives workable or stiff enough to tempt somebody with a hose. It costs one subtraction.
No. Slump responds to water, aggregate grading and shape, temperature, admixture dosage and time since batching, so low slump is not by itself evidence of low water. Correct the mix water from measured moisture and treat slump as the verification. If slump is still low with the arithmetic done properly, the answer is a water-reducing admixture, retempering strictly within what the specification allows, or a look at the mix design.
Ask in writing for two things: whether it is total or free moisture, and whether it is on the oven-dry or the wet mass basis. Those questions have four possible answers and only one drops straight into the batch correction. A report that does not say is not a usable report.