Concrete Cube Failure and Acceptance Criteria

Dnyan Deshmukh

Honorary Member
Staff member
Sep 7, 2017
Concrete Cube Failure and Acceptance Criteria
More panic situation in any project comes after failure in cube strength and most of the Engineers or not aware about acceptance criteria of it.

I asked many about what is acceptance criteria, most of the people mixed the things and not able to tell the correct way.

So decided to write article on it.


Before going to acceptance criteria directly, we will go step by step towards it, so that we can also able to identify the root causes and does concrete really failed in compressive strength test.

First we should be well aware about how to sampling and cube casting activity, if this goes wrong your result might indicate data which can’t be concluded.

You can learn this from below thread link.

How to do concrete sampling and cube casting

Now you did correct sampling and cube casting activity, now let’s move on to testing part.

You know we did numbering on the cube and cube number 1,2 &3 represents sample number 1, cube number 4,5&6 represents sample no 2 and so on.

Why we numbered and we should take cube for testing according to sample.

In statistical analysis and finding root causes of failure it helps a lot.

As per IS 456, the individual cube strength should not have variation more than + / - 15% of average of that three cubes.

What if you get more variation in any test result, how you will find the root cause of it.

Say you are testing 3 cubes and the numbers written on it is in serial and represents a sample.

Say you have cube numbered 4,5 & 6 – by numbers I can say it’s a sample no 2 and taken after x cum of concrete.

If you have cube numbers like 3,5&7 taken for testing – it won’t represent a single sample, because it’s a mixture of multiple sample.

Always do verify cubes taken for testing at any age are from same sample.

Now you have cubes which represent single sample and same workmen who filled it.

Suppose you got results like 15, 20 and 25 mpa

What it will indicate – it will clearly indicate that, there is issue in cube casting and you can discard those results easily.

So this will give you a power to have better Quality control because you know the reason of failure and where to take action.

One single sample of 3 cubes casted by a single person and numbered correctly, will never show variation in results and you will always get results with lesser variation than +/- 15%.

When it’s more than this, it indicates cube casting done improperly or you took sample cubes incorrectly for testing.

Hope you will never see a fancy results like above example after implementing it and can caught root causes in failure easily by eliminating possibility of cube sampling and casting issue.

Will write another article on other causes of failure in cubes sometime in future when I get time.

So move on to acceptance criteria

IS 456, specifies the acceptance criteria for compressive as well as flexure strength.

As flexure is 0.7 of square root of compressive strength, it is important to get passed in compressive strength results.

When we do cube testing, we write its results in cube test register with serial numbers, say 1 is xyz date concrete from abc location, 2 is def date concrete from ghi location and so on.

When we compare acceptance criteria, IS say average of 4 non overlapping consecutive test results should not be less than fck + 0.85 x standard deviation or fck + 4mpa (whichever is higher) for grades of concrete M20 and above.

What is mean by 4 non overlapping consecutive test – it means when you average out 4 test results for acceptance criteria, they should not overlap in calculations.

Acceptable averages


5,6,7&8 and so on serial number wise results should be average out.

If you do like averaging below


If you have low results at 6 and 8 and if you try to make averaging like 3,4,5&6 – then its overlapped to make you benefit by tweaking, this is not allowed and meaning of 4 non overlapping consecutive test results.

If you have good quality control at concrete, then your standard deviation will be less than 2 or slightly more.

In those cases, or in maximum cases applicable acceptance criteria will be always like fck + 4 mpa.

Sometimes there is a possibility that, in one of the test sample, you might get lower results and which makes that group of 4 non overlapping consecutive test result’s to not able to meet acceptance criteria.

In that situation, IS 456 also specifies, Individual test sample results should not be less than

fck – 4 mpa for M20 and above grade concretes.

So you should also check these criteria when single sample have lower results.

Each test results should meet fck + 4mpa criteria in order to ensure no failure and detect the issues at earliest.

Also it is important to observe failure patterns to get clues on what might be wrong or OK.