Resources

Why are my concrete cube results low? A lab checklist

Work through the specimen, the machine and the paper trail before you blame the mix. Most low results are found in that order.

6 min read

First, agree what “low” actually means

A cube result only means something against a target. Before anyone starts investigating, settle which target you are judging it against: the specified characteristic strength, the target mean the mix was designed to, or the acceptance rules of the project specification. A single cube below the specified strength does not automatically fail a set, and a set that looks low against the characteristic strength may still sit comfortably above the target mean at that age.

Check the age as well. A seven-day result is an indicator, not a verdict. Results are routinely compared at 28 days, or at whatever age the specification calls for, and a mix working at a slower early gain can look alarming at seven days and perfectly normal later.

Check the specimen before you blame the mix

Most low results that turn out to be laboratory-side sit on the cube itself. Confirm the dimensions were measured rather than assumed, that the faces are within the flatness and squareness tolerance of the test method in use (EN 12390, ASTM C39 and SANS 5863 all set one), and that the cube sits squarely on the platen without rocking.

Look at how the load faces were prepared. An uneven cap, a cap that is too soft or too thick, or a face that was ground without being checked for squareness will all produce a low result and an odd failure pattern. Record the cap type, the cap thickness and who did it — a run of low results in one batch of cubes often traces back to one afternoon of capping.

Then look at density and compaction. A cube that is noticeably lighter than its neighbours, or one with visible voids at a broken face, was not compacted properly and will read low no matter how good the concrete was when it left the truck.

Check the machine and the way the load was applied

Confirm the compression machine is within calibration and that the calibration certificate covers the range you are using. A machine that is accurate at 1,000 kN is not automatically accurate at 150 kN, and small cubes routinely fail in that lower band.

Check the seating. The upper platen should be free to align itself on the cube; a spherical seat that has been locked, is stiff with dirt, or is sitting off centre will load one edge first and pull the result down. Examine the platen faces for wear, pitting or embedded grit as well.

Finally, the loading rate. The three common methods each specify a controlled rate of stress increase, typically expressed as a band in megapascals per second. Testing too fast usually reads high, testing too slowly reads low, and a manually operated machine drifting outside the band is one of the easiest things to rule out and one of the easiest to miss.

Walk the paper trail behind the cube

If the specimen and the machine are clean, go back through the day. When was the concrete sampled, and how long after batching? Was it remixed before casting? How long did the cubes sit in the mould before demoulding, and where did they sit? A cube that spent its first night next to a warm wall, or was demoulded late and dry, is not a fair sample of the concrete.

Then check curing. The tank temperature, the time out of water before testing, and whether the specimen was actually the one the record says it is. Mix-ups between two sets cast on the same day are far more common than most labs like to admit, and a QR label on every cube is cheap insurance against them.

When the result really is the concrete

Sometimes the checklist clears the laboratory and the concrete is genuinely low: a batching error, an aggregate moisture correction that was not applied, a cement delivery that differed from the design, an admixture dispensed at the wrong dose, or simply water added on site to make a load place more easily.

That is a conversation to have with evidence rather than one result. Plot the set against its history, look at the trend across the same mix, same plant and same ages, and use cumulative sum or a rolling mean to separate ordinary scatter from a real shift. One low cube is noise; a run of them is a finding.

The short version

  • Confirm the age and the target you are judging against.
  • Measure the cube: dimensions, flatness, squareness, density.
  • Inspect the cap and the failure pattern.
  • Check machine calibration, seat freedom, platen condition and loading rate.
  • Reconstruct sampling, transport, casting, demoulding and curing times.
  • Confirm the specimen identity against the record.
  • Only then assess the concrete, using trend and set statistics rather than a single cube.

Try Simplecrete free for 1 month

The full version, on your own specimens and presses: QR labels, automatic kN capture, calculated MPa and density, pass/fail against the mix target, and a client-ready report.

Full version for 1 month. No card needed.

Your next cube could report itself.Start it today.

Sign up and use the whole platform on your own specimens and presses: QR labels, automatic kN capture, calculated MPa and density, pass/fail against the mix target, and a client-ready report.

Book a demo
Full version for 1 month. No card needed.
Or talk to us