Concrete grades: what M25 actually means

Concrete is specified by grade. M20, M25, M30, M40 and upward. Everybody on a site says these numbers all day, and a surprising number of the people saying them are not certain what the number refers to.

It is a strength. Specifically, it is the characteristic compressive strength in newtons per square millimetre, measured on standard cubes at twenty-eight days. The M stands for mix.

So M25 concrete is concrete specified to reach a characteristic cube strength of 25 N/mm² at 28 days. That is the whole definition, and each part of it is doing work.

"Characteristic" is not "average"#

This is the part that gets skipped, and it is the most important.

Characteristic strength is not the mean of your test results. It is a value below which only a small proportion of results is expected to fall — the code defines it as the strength below which not more than five per cent of test results are expected to lie.

Concrete is a variable material made in the open air by people with buckets and lorries. Test a hundred cubes from a well-controlled batching plant and you get a spread, not a number. Specifying the average at 25 would mean that something close to half of your concrete is weaker than the design assumed, which is not a position any structural engineer will accept.

So a mix designed for M25 is targeted well above 25, by a margin that depends on how consistent that particular plant and that particular supply chain actually are. A plant with tight control needs a smaller margin. A plant with variable aggregate needs a bigger one. The margin is a statement about the process, not about the concrete.

Two practical consequences follow.

A single cube result below 25 is not automatically a failure. The specification is about a distribution, and the code sets out how results are grouped and judged. A single low cube is a signal to look, not a verdict.

A single result above 25 does not prove compliance either. People misuse this constantly in both directions.

"At twenty-eight days"#

Concrete continues to gain strength for a long time. Twenty-eight days is the conventional age at which it is judged, and it was chosen as a practical compromise between waiting for the strength to develop and needing an answer before the building is finished.

It has one very uncomfortable property: by the time you have the result, the concrete has been in the structure for four weeks and quite possibly has three floors on top of it.

That is why seven-day results are taken as well. They are an early indication — a bad seven-day result is a reason to stop and think, not a certification of anything. The relationship between the seven-day and twenty-eight-day figure varies with the cement, the mix and the weather, so treating it as a fixed percentage is a mistake that has led people to accept bad concrete and reject good concrete in roughly equal numbers.

The full mechanics of the test, and what actually goes wrong with it, are in the piece on the concrete cube test.

Grade is not a mix ratio#

An older way of specifying concrete is by nominal proportions — one part cement, so many parts sand, so many parts aggregate. You will still hear these on site.

Nominal mixes are permitted only for the lower grades. Above that the mix has to be designed: proportions worked out for the specific materials, tested by trial, and adjusted. This is not bureaucracy. The strength of concrete depends much more on the water-cement ratio than on the cement content alone, and a nominal proportion says nothing at all about water. Two batches at the same nominal ratio, one made with a wet mix because it was easier to place, can differ enormously in strength.

Which is why the argument about water on site is not really an argument about workability. It is an argument about strength, conducted in the language of convenience, and it has its own piece: the slump test.

What a grade does not tell you#

Grade is compressive strength and nothing else. A specification often carries several other requirements alongside it, and they are not implied by the M number:

• Exposure conditions and durability. Minimum cement content, maximum water-cement ratio and minimum grade are set by how aggressive the environment is. A structure in a coastal or chemically aggressive setting has requirements that a strong-but-porous mix will not satisfy.
• Maximum aggregate size, which is governed by the spacing of the reinforcement and the thickness of the member.
• Workability, expressed as a slump range appropriate to how the concrete will be placed.
• Cement type, which affects heat of hydration and rate of gain.

A load that meets the strength and misses the durability requirements has passed the test everybody watches and failed the one that decides how long the structure lasts.

The record that outlives the pour#

The thing nobody thinks about on the day is that the grade is a claim you may have to defend for decades.

For each pour, the useful record is: the date, the location in the structure, the grade specified, the quantity, the supplier or the mix used, the challan or batch reference, the slump measured on arrival, who received it, and the seven- and twenty-eight-day cube results with their identification marks.

Most of that is captured at the barrier, on a load the gate should already have been expecting — why a lorry surprises the gate.

That is not a long list, and almost nobody has it complete. We brought eight months of concrete history into our own system from a spreadsheet and found the value was in exactly those fields — the receiving officer, the challan number, the rate, the slump, and the cube results where they had been taken. What made the import trustworthy was that it reconciled: the number of rows, the total volume, and the total value all had to reproduce the figures the register claimed before a single row was written. Two rows had to be hunted down first because a serial number was blank on one and mistyped on the other.

A parse that cannot reproduce the total it claims is not a parse anybody should import. The same is true of a concrete register. If the loads on the page do not add up to the volume in the structure, one of the two is wrong, and finding out which one after the fact is very expensive.

The short version#

M25 means: designed so that, at twenty-eight days, no more than a small proportion of standard cube results fall below 25 N/mm².

It is a statistical statement about a process, not a label on a lorry.