Three generations of pipe, one house
An interwar semi was plumbed once, then re-plumbed partially at least twice. What survives is a composite.
| Era | Typical material | How it fails |
|---|
| Original, 1930s | Lead supply, galvanised steel or lead internal distribution, cast iron waste | Internal corrosion and scaling, pinholes at thread roots, soft-metal creep |
| Post-war to 1980s | Copper on soldered or compression joints, first central heating installation | Pinholing from flux residue and dezincification, weeping compression joints, notched joists |
| 1990s onward | Plastic push-fit and barrier pipe, plastic waste, MDPE supply | Improperly seated push-fit, missing insert sleeves, expansion at rigid clips |
Any one of those systems on its own is reasonably reliable. It is the boundaries that fail, and a house like this has a lot of boundaries.
Galvanised steel, and what it does before it goes
Galvanised steel distribution pipework was standard in this era and a surprising amount survives, particularly on cold feeds in lofts and on runs to outside taps. It fails from the inside, which is why it gives so little warning.
The zinc coating protects the steel until it is consumed, and then corrosion products build up inside the bore. The sequence is predictable: flow drops at the furthest outlet first, water runs slightly brown after the house has been empty, and the pipe eventually perforates, most often at a threaded joint where the thread cutting removed the galvanising.
Two practical points. Reduced flow at one tap in an older house is worth investigating as corrosion rather than dismissing as low pressure. And galvanised steel that has started to fail at one thread will fail at others, so a repair that replaces one fitting is a stopgap rather than a solution.
Where copper was joined directly to galvanised steel during a partial re-plumb, the two metals in contact with water form a galvanic couple and the steel corrodes preferentially and faster than it otherwise would. It is not instant. It is a decade or two, and it puts the failure exactly at the joint where somebody once made an improvement.
If you can see a copper-to-steel junction with no dielectric fitting between them, that is a place to expect trouble, and it is worth photographing and noting even if it is dry today.
The solid ground floor
Here is the part specific to this stock. Many 1930s semis were built with suspended timber ground floors, and a great many of those were replaced with solid concrete at some point between the 1960s and the 1990s, either because of rot, because of a kitchen extension, or as part of a general modernisation.
When that happened, the pipework that had been clipped under the joists went into or under the new slab. Sometimes in a duct. Usually not. Frequently with no record.
What that means when a leak starts:
- There is no void to inspect and no board to lift. The only route is through the floor.
- A heating pipe in a slab produces a warm patch on the floor, which is the most useful symptom available and worth checking with a bare foot early in the morning.
- A cold water pipe in a slab produces nothing at all until the damage is significant, because there is no thermal contrast and no visible route.
- Water tracks along the slab and emerges at the perimeter, so the damp skirting is rarely above the leak.
Where the heating is losing pressure and nothing is visible, the pattern of the loss is the first diagnostic and it costs nothing: our boiler pressure pages set out what a weekly top-up means as against a daily one.
The buried run to the garage or extension
Almost every one of these houses has had something added: a garage, a rear extension, a conservatory, an outside tap, occasionally a garden room with its own supply. Each of those involved a buried run across the garden or the side passage, laid by whoever was doing that job at the time, at whatever depth was convenient.
These are the most common source of an unexplained meter reading in this stock, for three reasons. They are shallow, so they suffer frost damage. They are frequently in blue MDPE joined to whatever was there with a compression fitting in an unmarked location. And they are often isolated by a valve that nobody has turned since installation, so the first step of any diagnosis fails.
If the meter is moving with the house isolated, that is where to look before anything else. Water meter turning with everything off covers the test.
Supply pipe: worth checking the age
On a house of this date the incoming supply may still be lead. Thames Water's lead replacement scheme is aimed at properties built before 1970, which covers this stock comfortably, and the ownership line is the same everywhere: they are responsible for the mains and for the communication pipe linking your supply pipe to those mains, and you are responsible for the supply pipe running from the boundary of your property into your home, plus all internal pipes and fittings. Where they do replace, their standard is to put in 25mm where the existing pipe is 12-15mm, with like-for-like sizing on anything larger.
Supply pipe work should be carried out by someone competent in the Water Fittings Regulations. WaterSafe approved businesses are trained in those regulations, carry public liability cover and are audited, and a certificate of compliance from an approved plumber provides a defence for a property owner challenged by a water supplier enforcing the regulations. Ask what paperwork you will get.
Repair the pinhole, or replace the run
The honest answer in a house of this age depends on the material.
- Sound copper with one bad joint: repair the joint. There is nothing wrong with the pipe.
- Galvanised steel with a perforated thread: replace the run. The rest of it is in the same condition.
- Lead supply that has failed: replace to the boundary, for the reasons set out on our burst pipe repair page.
- Anything buried in a slab: where it is practical, re-route above floor or through a new duct rather than burying a repair. Opening the floor twice costs far more than the extra pipe.
That last one is the recommendation we make most often in this housing stock and the one most often resisted, because a surface run looks worse than a buried one. It also means that the next failure, in twenty years, is a twenty-minute job rather than a floor.
