Skip to the article

Adding Water to Concrete on Site: What Half a Pail Does to Strength

By 6 min readUpdated

Every engineer has seen it. The mix in the one-bag mixer looks stiff, the pour is slow, and someone tips another pail of water into the drum. The concrete flows nicely, the crew is happy, and the slab or column gets finished early.

I don't blame the worker. Nobody explained to him what that water does. So here is the explanation, with numbers, so you can show it to your crew.

Strength comes from the water-cement ratio

Concrete gets its strength from the cement paste. The paste is cement and water. The less water there is for each kilo of cement, the denser and stronger the hardened paste. Engineers call this the water-cement ratio, w/c:

w/c = weight of water ÷ weight of cement

Water weighs 1 kg per litre, so for one 40 kg bag of cement it is simply litres of water ÷ 40.

The link between w/c and 28-day strength is well known. ACI 211.1, the standard American guide for proportioning concrete mixes, gives a table of typical values (Table 6.3.4(a), for concrete without entrained air):

w/c Typical 28-day strength
0.4240 MPa
0.4735 MPa
0.5430 MPa
0.6125 MPa
0.6920 MPa
0.7915 MPa

These are typical values for average materials, not a guarantee. A trial mix with your own sand, gravel and cement is the real answer. But the trend is the same for every mix: more water, less strength.

Water-cement ratio versus 28-day strength (ACI 211.1 Table 6.3.4(a)) Line chart of the typical 28-day compressive strength of concrete without entrained air against water-cement ratio: w/c 0.42 gives 40 MPa, 0.47 gives 35, 0.54 gives 30, 0.61 gives 25, 0.69 gives 20 and 0.79 gives 15 MPa. A second scale shows litres of water per 40 kg bag (w/c times 40). A dashed line marks f'c = 21 MPa on the plans. Three site batches are marked: 22 L per bag (w/c 0.55) at about 29 MPa, OK; 30 L (w/c 0.75) at about 17 MPa, below 21; 34 L (w/c 0.85) past the end of the table, below 15 MPa. 0 5 10 15 20 25 30 35 40 45 0.40 16 0.45 18 0.50 20 0.55 22 0.60 24 0.65 26 0.70 28 0.75 30 0.80 32 0.85 34 0.90 36 w/c L/bag Water-cement ratio w/c, and litres of water per 40 kg bag (w/c × 40) 28-day compressive strength (MPa) Past the end of the table f'c = 21 MPa on the plans More water, less strength 40 35 30 25 20 15 22 L per bag, w/c 0.55 about 29 MPa: OK 30 L per bag, w/c 0.75 about 17 MPa: below 21 34 L, w/c 0.85 below 15 MPa ACI 211.1 Table 6.3.4(a), no entrained air Site batch, read between table points
Figure 1. Typical 28-day strength against w/c from ACI 211.1 Table 6.3.4(a), with the matching litres of water per 40 kg bag. The rings mark the 22 L and 30 L batches worked out below; 34 L is past the end of the table.

The numbers for one bag

Let's take a typical residential job with f'c = 21 MPa on the plans. A mix designed with about 22 litres of water per 40 kg bag gives w/c = 0.55. From the table, that is about 29 MPa at 28 days. That margin above 21 MPa is on purpose: real batches vary, and the design has to come out above f'c on average.

Now add water on site:

Water per bag w/c Approx. 28-day strength Against 21 MPa on the plans
22 L (as designed)0.55about 29 MPaOK, with the normal margin
30 L (+ half a pail, 8 L)0.75about 17 MPaBelow 21 MPa, about 19% short
34 L (+ 12 L)0.85below 15 MPa (past the end of the table)Far below 21 MPa

Half a pail per bag takes the concrete from about 29 MPa down to about 17 MPa. That is roughly 40% weaker, and below what the structural engineer designed for.

One 40 kg bag of cement with 22, 30 and 34 litres of water Bar chart and table. 22 L of water per 40 kg bag (as designed): w/c 0.55, about 29 MPa at 28 days, OK against f'c = 21 MPa with the normal margin. 30 L (22 L plus half a pail, 8 L): w/c 0.75, about 17 MPa, FAIL, about 19% short of 21 MPa. 34 L (22 L plus 12 L): w/c 0.85, below 15 MPa, past the end of the ACI table, FAIL, far below 21 MPa. Going from 22 L to 30 L makes the concrete roughly 40% weaker. 0 5 10 15 20 25 30 35 28-day strength (MPa) about 29 MPa about 17 MPa below 15 MPa past the end of the table f'c = 21 MPa on the plans roughly 40% weaker about 19% short As designed + half a pail (8 L) + 12 L Water per 40 kg bag w/c Against 21 MPa on the plans 22 L 30 L = 22 + 8 34 L = 22 + 12 0.55 = 22 ÷ 40 0.75 = 30 ÷ 40 0.85 = 34 ÷ 40 OK normal margin FAIL about 19% short FAIL far below 21 MPa
Figure 2. One 40 kg bag with 22, 30 and 34 litres of water, checked against the 21 MPa on the plans. Strengths are approximate values from ACI 211.1 Table 6.3.4(a).

The column, beam or slab was designed with 21 MPa in every calculation: shear, development length of the bars, punching around columns, deflection. All of those go down when the concrete does.

It is not only strength

Extra water also causes:

  • More shrinkage cracks. The extra water evaporates and leaves voids; the concrete shrinks more as it dries.
  • Bleeding and segregation. Water and fines rise to the top, gravel settles. The top of the slab becomes weak and dusty.
  • Lower durability. More voids means water and salts get in faster, and the bars rust sooner. This is why NSCP 2015 Table 419.3.2.1 puts a maximum w/c on concrete exposed to sulfates, freezing, wet conditions or chlorides. For those exposure classes, extra water can break a code limit outright, not just lower strength.
Slab section: designed mix compared with a watered mix Two slab cross-sections, not to scale. Top, the mix as designed with 22 L of water per bag: even, hard surface, gravel spread evenly through the depth, dense paste with few voids, bars well protected. Bottom, the same slab with extra water added on site: bleed water on top, a weak and dusty top layer where water and fines rose, shrinkage cracks, voids where the extra water evaporated, gravel settled to the bottom, and water and salts reaching the bars so they rust sooner. As designed: 22 L of water per 40 kg bag Even, hard surface Gravel spread evenly through the depth Dense paste, few voids Bars well protected With extra water added on site water, salts Bleed water on top Weak, dusty top layer (water and fines rose) Shrinkage cracks Voids where the extra water evaporated Gravel settled to the bottom Water and salts reach the bars sooner: rust Not to scale
Figure 3. What extra water does inside a slab: water and fines rise to the top, gravel settles, and the voids and shrinkage cracks let water and salts reach the bars. Not to scale.

"But it has always been done this way"

Many buildings with watered concrete are still standing. That is true, and it is because designs have safety factors. But the safety factor is there for the things we cannot control: material variation, loads larger than expected, an earthquake. When we spend the safety margin on easier pouring, there is less left for the earthquake.

What to do instead

If the mix is too stiff to place:

  1. Do a slump test first (ASTM C143: the cone, rod 25 times in 3 layers, lift, measure). Compare with the slump in the specifications. Often the mix is actually fine and only needs proper vibration.
  2. Ask the engineer to adjust the mix. If more workability is needed, the fix is to add cement and water together at the same w/c, or to use a water-reducing admixture (plasticiser). Both keep the strength.
  3. Use a vibrator, not water. Proper vibration fills the forms and removes voids.
  4. Pre-wet the forms and the dry aggregates, so they don't steal water from the mix.
  5. Cure properly. Keep the concrete wet for at least 7 days. Curing is the cheapest strength gain on any site.
  6. Take cylinder samples on important pours, so you know what you actually got.
Slump test in four steps (ASTM C143) Four panels. 1: fill the slump cone in 3 layers. 2: rod each layer 25 times with the tamping rod. 3: lift the cone straight up; the concrete slumps. 4: set the cone beside the concrete, lay the rod across the top of the cone and measure the drop to the top of the concrete: that is the slump. Compare it with the slump in the specifications. 1 2 3 1 Fill the cone in 3 layers × 25 2 Rod each layer 25 times 3 Lift the cone straight up slump 4 Measure the drop (slump) and compare with the specs
Figure 4. The slump test (ASTM C143): fill the cone in 3 layers, rod each layer 25 times, lift the cone straight up and measure the drop. Compare the result with the slump in the specifications.

For the engineer in charge

Explain it to the crew with the table above. Most workers add water to make their job easier, not to cheat. Once they see "half a pail = 40% weaker", most will stop.

On the design side, remember that every member was designed for the f'c in your general notes. If you suspect site water was added, take cores or check the cylinder results before you load the structure.

General information only. Always follow the approved plans and the instructions of the engineer of record.