Laboratory Concrete Testing

All the raw materials used to produce concrete are variable. The Readymix supplier must absorb these variances in his production process and produce concrete that will meet with the specification. This is inherently the risk that the Readymix Supplier will carry.

The first party that needs to know whether the Readymix supplier is producing concrete within the strength requirement, is the Supplier himself. To that end he will take samples normally at the plant from the back of the truck to have tested as process control samples. It is up to the client to take samples for cube making and testing for acceptance control for quality assurance purposes.

These two types of testing are very similar, yet there are fundamental differences between the two as well. The main difference in testing is the sampling that is done from the back of the truck at the plant for process control testing, whereas on site sampling is done from a moving stream after the first 15% and before the last 15% of concrete has been offloaded for acceptance control for quality assurance testing (SANS 5878 par 12.2). 

The criteria for assessing these results are then also different. Normally process control cubes are not (but may be in certain circumstances) accepted as acceptance control testing for quality assurance.

The following things can go wrong with the testing (of perfectly good concrete):

    1. Process control cubes are to high
    2. Process control cubes are to low
    3. Acceptance control cubes are to high
    4. Acceptance control cubes are to low

In number 1 above, this would give a false sense of security and not even knowingly substandard concrete may be produced that could result in a failure and huge claim.

In number 2 above that Readymix supplier will increase cement contents, because he is scared of a claim. The customer will be very happy with this concrete that will give higher strengths and even earlier strengths, but this is at the Readymix suppliers expense.

In number 3 above, it is great for everyone, until something goes wrong and cores are requested, that could then result in huge failures.

In number 4 above, the specification rules that the engineer may call for the process control test results and will then call for cores to be drilled, if the situation could not be resolved. Ultimately the concrete will be accepted, but with mistrust and unhappiness due to delays and penalties and the additional cost of the drilling and testing of cores. The Readymix supplier will probably increase cement contents to prevent possible failures.   

Concrete laboratory testing:

Due to our involvement in disputes and court cases, Go Consult have been asked on a number of occasions to investigate the state of testing at some laboratory. Eight of the eleven laboratories visited in a sample run were doing testing commercially at the time. Not one of the laboratories included in this article was accredited by SANAS (South African National Accreditation Systems, for concrete testing. SANAS accreditation would imply that they have a management system whereby they will ensure proper documentation of everything that they do and that they have systems in place to ensure a repeatable test that will reproduce a similar test result every time.

Incorrect testing:

At every laboratory there were signs of incorrect crushing of cubes with tensile cracks and the cubes only being crushed on one side, with the other side failing in tension. This error will definitely give a lower result and create the impression that the concrete is of substandard quality.

This of course would raise the following logical questions (as the advocate would of course ask in a court): how many of the cubes were tested like that? Which ones were they and what would the effect of that be? This should result in a rejection of all test results, as we know that some results are lower, but we do not know which and we do not know by how much, so we cannot use any of that set of results.

We thus might have to drill cores to determine the strength of the concrete. These are often incorrectly done and interpreted, which has lead to many instructions to break out perfectly fine concrete.

So what can we do as Readymix suppliers and even Contractors?

Visit the laboratory that is testing for you and check the following items to start off with. The fact that you are checking will make them alert and because you are now informed, you can then be sure (or at least hope) that they will jack up their systems etc etc.

Items to be checked:


    • Is every cube properly marked on the side (paper labels tend to come loose in the water).
    • Are the edges sharp (cube mould did close properly on the edges).
    • Are the cube moulds square (Measure two diagonals).
    • Are they oiled (not old car oil).
    • Is each cube tamped enough? (25 times for two layers on 100mm cubes and at least 45 times for three layers in 150mm cubes) or on a vibrating table.
    • Is there a paper label on each cube with at least some reference number on it.
    • Do the two halves of each cube mould correspond (see the stamped marks on each cube.
    • Do they have a slump rod or vibrating table to compact the cubes with.
    • Do they understand that no vibration is allowed after the cubes have been made? (even nearby trains).



    • Curing temperature in the curing tank (measure it with the loose thermometer) is often not right. It should be between 22 – 24 degrees C.
    • Is the curing tank heating system always on and plugged in (not only in the day when the generator is working)? The Specification states that for site cubes that have to stay on site for an extended period, should be covered and then placed in water preferably 22-24 Degrees C.
    • Is there a pump stirring the water constantly to prevent hot spots due to heat of hydration of the cement (this is not a requirement of the test method).
    • Are all of the cubes constantly submerged in the water.



    • Is there a calibration certificate and is it less than 18 months old? Was it done by a SANAS accredited organisation for press calibrations?
    • Are there locating pins on the platens to ensure that the cube is placed right centre between the platens. (Just marks (circles or lines) on the platens alone too often leave it to the operator to locate cubes and they do not do it correctly.)
    • Did they ensure that two faces formed by the cube mould were against the platens while crushing.
    • Is the top platen loose to swivel?
    • Is the machine clean?
    • Has the operator been trained to operate the machine?
    • Has the loading rate (N/s) been adjusted for the type of test (It should be different for different sizes of cubes and even cores).(0.3MPa/sec +- 0.1 MPa/sec SANS 5863 par 5.3).

Should a dispute arise, Go Consult can assist to settle this and put it in perspective. Go Consult is also available to do an audit at the laboratory to ensure proper test results.

Jacques Smith Pr Eng (Civil)
Reg 880202
Concrete Technologist
Cell 082 309 1884

Figure 1:

Unsatisfactory failures in cubes as per SANS 5863:2006

Figure 4:

Cubes not submerged in water

Figure 2:

Cube placed eccentrically in the press, crushing one side and leaving a tensile crack on the other. (Fixed patens could also cause like failures).

Figure 5:

Cube placed eccentrically in the press (note the circles)

Figure 3:

Cubes not cured in water

Figure 6:

Incorrect failures found at one laboratory