Concrete Cube Testing

Definition

Concrete cube testing is the routine practice of sampling fresh concrete at the point of pour, casting standard specimens, curing them under controlled conditions and crushing them at defined ages to verify that the mix has achieved its specified characteristic strength. It is the single most consequential quality test on a building site — the results directly determine whether the concrete stays in the structure, gets remediated, or gets broken out.

The Standard Regime

• Sample taken at the point of discharge, in accordance with EN 12350 / ASTM C172.
• Six 150 mm cubes cast per sample (in metric practice) or six standard cylinders (in ASTM practice).
• Slump and temperature recorded at the time of sampling.
• Cubes cured in water at 20 ±2 °C after initial 24-hour set.
• Two cubes tested at 7 days, three at 28 days, one held in reserve.
• Sampling frequency — typically one set per 20 m³ or one set per pour, whichever is more frequent, with a minimum of one set per structural element.

Why It Matters

Concrete is the only significant structural material on a building site that is manufactured on the day it is used, from ingredients that vary batch by batch, mixed by a driver you have never met, and placed by a crew who did not blend it. The cube result is the only independent evidence that the material you paid for is the material you got. When a cube fails — and cubes do fail — it triggers a chain of decisions that can affect the whole structure. When the sampling regime itself is unreliable, every one of those decisions is made on sand.

Real-World Example

On a bridge abutment in the West Midlands, three consecutive 28-day cube results came in at 28 MPa against a C35 design. The site engineer flagged them immediately. The batching plant's records showed compliant mixes. The supplier disputed the results, arguing that the site had cured the cubes badly. The client's independent testing house cored the abutment itself at three locations. The cores returned 41 MPa. The abutment was fine; the cubes had been left standing in a hot yard for three hours before curing, and their strength had suffered. The lesson cost nobody anything except a week of arguing. Had the cores returned 25 MPa, the abutment would have been demolished. The cube result is a proxy — a very important proxy — but only as reliable as the discipline behind it.

How to Run It Properly

1. Sample at the discharge point, every pour — not from the batching plant, not from a stockpile. The concrete on test must be the concrete in the structure.
2. Cure in a bath, not a bucket. Temperature-controlled, indoors, undisturbed. Cube results are more sensitive to curing than to the mix itself.
3. Label each cube — pour ID, element, date cast, age at test. Ghost cubes with no traceability are the source of most disputes.
4. Test at an accredited laboratory — UKAS, NATA, A2LA — never in-house on an uncalibrated press.
5. Trend the results — mean, standard deviation, characteristic strength — pour by pour, mix by mix. Downward drift is a warning even when individual results still pass.
6. Have a written response procedure for low results — additional testing, coring, load testing, structural assessment.

Practical Lessons Learned

• Curing is the number one cause of "failed" cubes. A cube that spent its first 24 hours in the sun will never recover the lost strength.
• Batching-plant records lie occasionally. A cube taken at the truck cannot lie.
• Pump-truck concrete tested at the hopper gives different results than concrete tested at the boom end. Standardise the sampling point.
• Cold weather concreting distorts everything. Ambient temperature at the time of cast belongs on the record.
• The reserve cube exists for a reason. Do not discard it early — it is the only defence against a disputed result.

Expert Tips

• Pair every cube result with the delivery ticket in the QA record. Mix design, water added on site, time on truck — all part of the same story.
• Adopt the 7-day rule — a 7-day result below 65% of the design strength triggers a design engineer review, not a wait-and-see.
• Two low results are a pattern, not an anomaly. Escalate on the second, not the third.
• Core rather than argue. If cube results are disputed, cut cores from the structure and settle the debate on physical evidence.
• Keep the curing tank photographed daily. If a result is challenged eighteen months later, the photo timeline is your defence.

Common Mistakes

• Sampling from the wrong point — plant instead of pour, hopper instead of boom.
• Cubes cured in ambient conditions.
• Poor labelling — pour ID guessed later.
• Discarding the reserve cube because "the others passed."
• Ignoring downward trend because the individual results still meet spec.
• No written escalation procedure — every low result triggers panic instead of process.

Key Takeaways

• The cube is a proxy for the concrete in the structure — only as reliable as the sampling, labelling and curing behind it.
• Sample at the discharge point, every pour, minimum one set per element.
• Cure in a controlled bath, not a site bucket.
• Trend the results — drift matters more than a single number.
• Have a written response procedure for low results, and follow it every time.

Related Concepts

Interlocks with Concrete Pour Cards, Quality Management, NCRs, and Method Statements. Sample logs at PMMilestone.org.