Knowledge · Construction risk
The Eurocode 2-3 Problem: Why Designing Watertight Concrete Strictly to Tightness Class Can Make It Less Waterproof
EN 1992-3 tightness classes drive crack widths down, reinforcement up and concrete quality down - while the waterproofing admixture's certificate points back to EN 1992-3. A circle structural engineers and waterproofing designers have to break together.
By Ben Hickman - contributing member, BS 8102:2022 committee · Last updated 15 September 2026
Direct answer
There is a circle at the centre of watertight concrete design that most specifications never acknowledge. EN 1992-3 tightness class 1 - the class usually reached for on a Type B basement - limits crack widths to 0.2 mm and below, tightening further as the water head rises. Achieving that in a thick retaining wall means heavy reinforcement at close spacing, which congests the section, which makes the concrete hard to place and compact, which produces the voids and honeycombing through which basements actually leak. Meanwhile the waterproofing admixture specified to help is certified for crack widths up to about 0.3 mm, and its certificate refers the structural designer back to EN 1992-3. Strict compliance with the tightness class can therefore produce concrete that is less waterproof than a looser limit, properly placed, would have been. This article states the problem. My colleague Johnny Etkind will set out the resolution in the third IStructE webinar of the compliant-basements series in November 2026.
Full explanation
Two standards, two questions
BS 8102:2022 asks what internal environment the space needs. Its grades - Grade 1B permitting damp areas but no water penetration, Grade 2 no penetration and controlled dampness, Grade 3 dry - describe the outcome. It then offers three routes: Type A barriers, Type B structurally integral protection and Type C drained cavities, alone or in combination.
EN 1992-3 asks a different question: how much leakage through the concrete itself is acceptable? Its tightness classes run from 0, where some leakage is acceptable, to 3, where none is permitted. Class 1 permits a small amount of leakage with some surface staining or damp patches, and it controls that by limiting the design crack width through the section - about 0.2 mm where the ratio of water head to wall thickness is low, falling to 0.05 mm where the ratio is high.
The two are routinely equated: Type B concrete, tightness class 1, Grade 1B. That equivalence is loose. A tightness class describes the concrete; a grade describes the room. Treating them as interchangeable is where the trouble starts, because the engineer specifies the class and assumes the grade follows.
The circle
Take a large warehouse basement I reviewed, designed by a competent structural engineer who had - as engineers routinely do - self-determined the waterproofing grade and the Type A plus Type B solution without a waterproofing specialist appointed. The specification referenced EN 1992-3 without stating a tightness class. To be safe, the reinforcement was designed to a tight crack-width limit. The result was bar densities and coupler congestion that the contractor could not properly place concrete through. The walls came out with voids you could put your arm into. The remediation was resin injection, wall by wall, for months.
Now follow the reasoning that produced it. Tighter class means smaller design crack width. Smaller crack width means more steel. More steel means congestion. Congestion means poor compaction. Poor compaction means voids and defective joints. Voids and defective joints are how basements leak - not through a 0.25 mm flexural crack that a crystalline admixture would have sealed in a week, but through a cavity the size of a fist that no admixture can touch.
And the admixture is the other half of the circle. Integral waterproofing products are certified - typically by the BBA - on the basis that they self-seal or resist water at crack widths up to around 0.3 mm. That is a wider crack than tightness class 1 permits, so on paper the admixture is doing nothing the structure has not already done. Its certificate then directs the designer to EN 1992-3 for the structural design. The waterproofing product defers to the structural standard; the structural standard drives the reinforcement that defeats the waterproofing. Nobody in the loop is wrong, and the wall leaks.
Where compliance and dryness part company
The audience at the IStructE webinar in September 2026 asked, in effect, whether there is a point at which strict compliance with EN 1992-3 makes the concrete less waterproof. Yes, there is, and it is reached whenever the reinforcement needed for the tightness class exceeds what the section can be concreted through. Past that point the designer has a choice the standards do not make for them: thicken the section, relax the class and rely on combined protection, change the concrete mix and placement method, or accept the congestion and the remediation programme that follows it. The first three are design decisions. The fourth is the default when no one has been appointed to make them - which is why BS 8102 expects a waterproofing specialist on the team before the structure is designed, not after.
What to do now
Until the November session, three things any project can do.
State the tightness class. A specification that references EN 1992-3 without one has not specified the concrete; it has left the contractor and the reinforcement detailer to guess, usually conservatively and usually wrongly.
Decide whether Type B is being relied on alone. If the grade demands dryness the concrete cannot guarantee, the answer is combined protection - and a looser, buildable tightness class behind a Type A membrane or a Type C cavity drain will outperform a tight one that cannot be compacted.
Put the structural engineer and the waterproofing designer in the same room before the reinforcement is detailed. Crack width, section thickness, mix design, pour sequence and joint detailing are one decision, not two disciplines’ worth of assumptions about each other. The structural engineer’s scope, and its limits, sets out where the handover sits.
The resolution - how to set a tightness class the wall can actually achieve, and when to stop chasing crack width and change the strategy - is Johnny’s subject for November. If you have a Type B basement at Stage 3 or 4 now and are unsure which side of the line it sits, put the section, head and grade to the Waterproofing Wisdom agent.
Frequently asked questions
What is a tightness class in Eurocode 2-3?
EN 1992-3, the Eurocode for liquid-retaining and containment structures, classifies a concrete element by how much leakage is acceptable: tightness class 0 (some leakage acceptable) through to class 3 (no leakage permitted). Tightness class 1 limits leakage to a small amount with some surface staining or damp patches acceptable, and controls it through a crack-width limit that tightens as the ratio of water head to section thickness increases - around 0.2 mm at low heads, down to 0.05 mm at high ones.
How does the Eurocode 2-3 tightness class relate to BS 8102 grades?
They measure different things. BS 8102:2022 grades describe the internal environment required - Grade 1B, for example, permits damp areas but no water penetration. EN 1992-3 tightness classes describe leakage through the concrete itself. Tightness class 1 is often treated as broadly equivalent to BS 8102 Grade 1B for a Type B structure, but the equivalence is loose, and it is where much of the confusion in specification starts.
Why does a tighter crack-width limit make watertight concrete harder to build?
Controlling cracks to 0.2 mm or less in a thick retaining wall means much more reinforcement at closer spacing, often with couplers. That congestion makes the concrete harder to place and compact, which is how voids, honeycombing and leaking construction joints occur. The measure intended to make the structure tighter can produce the defects that make it leak.
What crack width do waterproofing admixtures and additives assume?
Most integral waterproofing products - crystalline and hydrophobic admixtures - are certified, typically by the BBA, to self-seal or resist water at crack widths up to about 0.3 mm. Their certificates then refer the designer back to EN 1992-3 for the structural design. A specification that demands tightness class 1 crack widths and an admixture certified for 0.3 mm is asking for two different things and getting neither cleanly.
How should a watertight concrete basement be specified?
As a waterproofing design first and a structural design second, with the two co-ordinated. The waterproofing specialist sets the BS 8102 grade and whether Type B concrete is being relied on alone or in combination with Type A or C; the structural engineer then chooses the tightness class and crack-width limit that the concrete can actually achieve on site. Specifying EN 1992-3 by reference, without a tightness class or with one the reinforcement cannot deliver, is the most common failure.
Related guidance
- 'We've Got the Contractor's Guarantee' - What It Covers, and When the Belief Costs Developers Millions
A genuine contractor's guarantee is worth having: it covers workmanship and, where the contractor designed the system, usually some design l
- Why the Defect Liability Period Expiring Is When Your Real Waterproofing Risk Begins, Not Ends
The defect liability period is a contractual convenience, not a description of how water behaves. On a major commercial basement it may run
- Flood Risk and Life Safety in Basement Design: The Risk Waterproofing Cannot Fix
A membrane keeps ground water out of a wall. It does nothing about water arriving through the lightwell window, the sewer surcharging into t
Working on a live scheme?
Put our AI agent to work. It'll reason through your specifics from BS 8102:2022 and land on a defensible recommendation.