This one really is a London problem
Most of what you read about local water is marketing. This is not. Thames Water takes about 65% of its supply from local rivers and about 35% from underground reservoirs, and states plainly that because the water passes repeatedly through chalky limestone, all the water in its region is hard. That is geology, not a service level, and the company says outright it cannot control it.
Thames Water publishes the bands it uses, measured as calcium carbonate:
| Category | Hardness (mg/l CaCO3) |
|---|
| Soft | Up to 100 |
| Slightly hard | 100 to 150 |
| Moderately hard | 150 to 200 |
| Hard | 200 to 300 |
| Very hard | More than 300 |
Your own postcode figure is on their site. We will not quote a number for your street, because it varies with the source your area is fed from. The point stands regardless: if you are in London, the water arriving at your cylinder is carrying dissolved calcium, and some of it is going to come out of solution inside the tank.
Where the scale actually lands
Calcium carbonate behaves backwards compared with most things you dissolve in water. Sugar dissolves better when it is hot. Calcium carbonate does the opposite — the hotter the water, the less it will hold, so it drops out onto whatever surface is hottest.
Inside a hot water cylinder, the hottest surfaces are in a short list:
- The immersion element sheath. Three kilowatts through a metal tube a few hundred millimetres long. Nothing else in the tank is close.
- The heating coil, if the cylinder is heated indirectly from the boiler. Less aggressive than an element, because the coil surface is cooler, but it still furs.
- The base of the tank, where anything that flakes off ends up as a grey-white sludge you will never see until someone drains it.
Scale is also a decent thermal insulator, which is the whole problem. A layer on an element does not just sit there — it stops heat leaving the element into the water. The element keeps producing the same 3kW and now has nowhere to put it, so the sheath temperature climbs. That is how a scaled element dies: not corroded away, cooked from the inside because its own coating stopped it shedding heat.
Running it hotter is the thing you control, and the thing people get wrong
Since scale precipitates faster at higher temperatures, a cylinder stored at 70 to 75°C will scale noticeably harder than the same cylinder at 60°C. Thames Water's own advice on reducing scale is to bring hot water down to 60 degrees.
Sixty is a floor, though, not a target you can keep sliding under. The Health and Safety Executive's temperature-control guidance for hot and cold water systems is that hot water should be stored at least at 60°C and distributed so that it reaches 50°C at the outlets within a minute. That guidance is written for duty holders in workplaces and rented and shared buildings rather than for someone heating their own house, but the biology behind it does not check your tenure. Stored water sitting lukewarm is a legionella question, and it is not one to answer by guesswork.
So the honest version is narrow: if your cylinder stat is set to 70 or 75 because someone turned it up years ago to get more usable hot water out of a small tank, bringing it back to around 60 will slow scaling and cost you nothing. Going below 60 to chase scale is trading one problem for a worse one. If 60 does not give you enough hot water, the tank is undersized or the stat is lying — that is a different fix.
The boss that will not come apart
This is where scale stops being an efficiency story and turns into a bill.
An immersion element screws into a threaded boss welded or brazed into the cylinder shell. On a vented copper cylinder that is a brass or copper boss in a soft copper tank. Years of hard water leave scale packed into the thread and, on some tanks, corrosion products bonding element to boss. Meanwhile the fibre washer underneath has gone hard and lost any give it had.
When the element eventually fails and somebody puts an immersion spanner on it, three things can happen. It comes out. It rounds off. Or it holds, the spanner wins, and the boss distorts or tears out of the shell. The third outcome is not rare on an old, heavily scaled copper cylinder, and it turns a straightforward element swap into a cylinder replacement, on a day nobody planned for it.
Unvented stainless steel cylinders behave differently. The shell is far stronger, so tearing the boss is much less likely, but the element itself can still seize solid in the thread and the seal arrangement is usually more particular about being reassembled correctly. Either way, nobody can promise the element will come out cleanly before they try.
If your cylinder is already weeping around the boss, do not start by trying to remove the element. Read the hot water section first, and isolate the immersion circuit at the consumer unit before anyone touches it.
Where scale is not the villain
Blaming every cylinder failure on hard water is lazy, and it lets the actual fault go unfound. Scale does not generally perforate a tank. When a cylinder genuinely leaks, it is usually one of these:
- Corrosion at a weld or seam, or a pinhole in an old copper shell.
- The boss, the element washer, or the immersion gasket.
- A fitting, an isolating valve or a compression joint on the pipework above the cylinder, where the water then runs down the tank and looks like the tank.
- On unvented systems, the expansion vessel, pressure relief valve or tundish discharging as designed — which is not a leak at all.
Scale's contribution is indirect and real: it creates local hot spots and thermal stress, it holds sludge against the metal, and it makes elements fail more often, which means the boss gets disturbed more often. That is a shorter life, not a hole.
If there is water on the airing cupboard floor, the useful question is which of those it is, and that is answered by looking, not by theorising about the water supply. Our water leak repair visits start by isolating and testing rather than by assuming.
Softeners and inhibitors, without the sales pitch
We do not sell water treatment, fit it, or take a commission from anyone who does, so take this as it is meant.
Ion-exchange softeners
These work. They swap the calcium and magnesium for sodium, so water leaving the unit genuinely will not form scale downstream. That is chemistry, not a claim. What they do not do:
- They do not remove the scale already sitting in your cylinder, on your element or in your coil.
- They do not reverse corrosion, fix a weeping boss, or extend the life of a tank that is already failing.
- They are not free to run. They need salt, they regenerate to drain, and they need servicing.
Thames Water's own guidance is that softened water is not recommended for drinking or cooking, that a separate unsoftened tap should be installed if you fit one, and that the plumbing and the unit should be inspected regularly by an approved plumber. Worth knowing before you commit, not after. Check your cylinder and boiler manufacturer's installation instructions too — some are specific about water treatment and about what invalidates a warranty.
Scale inhibitors and electronic or magnetic conditioners
These are a different category and they are frequently sold as if they were the same thing. They do not remove hardness minerals; the calcium is still in the water when it reaches your tap. The claim is that they alter how the crystals form so less of it sticks. Evidence for that varies by device and by installation, and results in the field are inconsistent enough that we will not tell you yours will work. They are cheap and passive, which is their appeal. Treat any promise of a specific percentage reduction with suspicion.
The one intervention nobody sells you: turn the stat down to 60 if it is above it, and do not leave a cylinder running hotter than it needs to be.
What we do and do not do about it
We are a leak firm. A cylinder that reheats slowly, rumbles, or gives you less hot water than it used to is a scaled or failing cylinder, and that is a cylinder engineer's job, not ours. We will say so rather than attend and charge you for the privilege.
What is ours: water where it should not be. A damp airing cupboard floor, a stain spreading on the ceiling below the cylinder, a warm wet patch you cannot trace, a meter that turns with everything off. We work at £150 an hour agreed before we attend, with no call-out fee, and if we do not find the leak you do not pay for the time spent looking. Where the leak has already damaged something, the written report from the visit is prepared so it supports a trace and access claim — what your policy then covers is the insurer's decision, not ours — and our own handymen and builders put floors, ceilings and tiling back afterwards.
