What Happens in an Annual Unvented Cylinder Service

Why an unvented cylinder needs servicing every year

A vented copper cylinder fed from a loft tank has nothing to check. It is open to the atmosphere through a vent pipe, so it cannot build pressure, and when a part fails the worst outcome is cold water or a weeping joint. An unvented cylinder is sealed and connected straight to the incoming main, and water at storage temperature inside a closed steel vessel holds real stored energy. The design answers that by giving the water somewhere to expand into and by stacking three independent devices behind the heat source, so no single failure can take the contents towards boiling.

Every one of those protections is silent when it is working and silent when it is not. An expansion vessel that lost its air charge two years ago gives no sign until a relief valve starts weeping. A thermal cut out that has gone open circuit changes nothing you can feel until the day the thermostat sticks closed and there is nothing behind it. The annual visit is when the silent parts have to prove they still work. Manufacturers also expect a yearly check to keep their own cover intact, which is the paperwork reason and not the real one.

The work is restricted, so the person doing it has to be a G3 qualified plumber and should be able to produce the registration without being asked twice.

How often it needs doing, and what a missed year costs

Once a year is the standard interval and it is what manufacturers ask for to keep their own cover in place. Some situations shorten it. A cylinder standing in hard water with scale already visible on the element, a cylinder in a let property where nobody is watching the tundish from one month to the next, and a cylinder that has needed a relief valve or an expansion vessel in the last twelve months are all worth looking at sooner than the anniversary.

What a missed year actually costs is predictable rather than dramatic. In the first missed year, usually nothing. By the second, the air cushion has gone and the expansion relief valve is passing water on every heating cycle, which almost nobody notices because the discharge runs outside and dries before anyone sees it. By the third, the valve seat has scaled and the valve either weeps constantly or no longer lifts at all. At that point the visit is a repair with parts in it rather than a service, and the cylinder has spent two winters with one of its three protections out of action.

If you have just taken over a property and nobody can tell you when the cylinder was last looked at, assume it was not. The benchmark book or the record card left near the cylinder is the only proof there is, and an unvented cylinder with no history behind it is the one most likely to be sitting on a flat vessel and a seized valve.

• Once a year is the standard interval.
• Sooner after any relief valve or expansion vessel work.
• Sooner where the water is hard and the element is already scaling.
• No record at all means treat the cylinder as overdue.

Checking the expansion arrangement, the check that matters most

Water expands by roughly four per cent going from cold to storage temperature, and in a sealed vessel it has nowhere to go unless something compressible is waiting for it. That is either an external expansion vessel, a steel drum with a rubber diaphragm and a charge of air behind it, or on bubble top designs a pocket of air trapped in the crown of the cylinder itself.

Air migrates. On a vessel it passes the diaphragm slowly or leaks at the schrader valve. On a bubble top design it dissolves into the water over months. Either way the cushion shrinks, and once it is gone the cylinder is heating a fixed volume of water inside a fixed volume of steel. Pressure climbs hard on every heating cycle, the expansion relief valve opens to protect the cylinder, and water passes through the tundish once or twice a day. That is the most common fault on an unvented cylinder in service, and it is a maintenance failure rather than a broken part.

The check is physical. The engineer takes the water pressure off the vessel, puts a gauge on the schrader valve, reads the air charge cold and compares it against the figure on the data badge. If it has dropped it is recharged with a pump and rechecked. If it takes air and loses it again within minutes the diaphragm has failed and the vessel is replaced rather than topped up. A bubble top cylinder has no vessel and no gauge, so the air pocket is re established by a sequence of isolating, draining to a specific point and refilling so air is trapped back in the crown. Done properly it takes twenty minutes and the tundish stops.

Where the expansion arrangement was wrong from the start, this is where it surfaces, which is why the quality of the original unvented cylinder installation tends to show up years later at a service rather than on the day it was commissioned.

• Charge read cold with the water pressure off, not guessed by tapping the vessel.
• Charge compared to the figure on the badge, not to a general number.
• A vessel that will not hold air gets changed, not pumped up and left.
• Bubble top cylinders get the air pocket re established by sequence, not topped up.

If water passes through the tundish daily and the cylinder was serviced recently without anyone taking a reading off the expansion vessel, the service missed the fault you were paying to have found.

The inlet control set and the strainer nobody cleans

Everything on the cold feed into the cylinder is bolted into one brass assembly at the base, and each job in it is checked separately.

The pressure reducing valve brings the incoming main down to the working pressure the cylinder was built for. It is checked static and again with a tap running, because a valve that has scaled up holds a good static reading and then collapses under flow. A reducing valve drifting upwards is one of the two reasons a relief valve weeps, the other being the expansion cushion. The single check valve stops heated water pushing back into the cold supply, and a passing one shows itself as warmth you can feel by hand on the cold pipe a foot from the control set.

The line strainer is a fine mesh inside the control set body and it is the part most often skipped. In London it fills with grit off old iron mains, scale flakes and fragments of jointing compound, and as it blocks it quietly starves the cylinder of flow. People report a weak shower and blame the cylinder or the main. The engineer isolates, unscrews the cap, lifts the mesh, flushes it, refits it and confirms the flow has come back. Ten minutes of work that settles a complaint which has been running for a year.

Exercising the relief valves, and why they seize

There are two relief valves and they answer different emergencies. The expansion relief valve sits on the inlet group, opens on pressure alone, and protects the cylinder from a failed expansion arrangement or a failed reducing valve. The temperature and pressure relief valve is screwed into the body of the cylinder with its probe in the stored water, and it opens if the contents ever approach boiling whatever the pressure is doing.

Both are exercised by hand. The cap is lifted or twisted so the valve comes off its seat, water is seen running through the tundish, the valve is released and it has to seat again cleanly with no drip behind it. That is the whole test and it is the only way to know the valve is not seized. A valve left untouched for five years in hard water grows limescale on the seat and the spindle, and by the time it is genuinely needed it may not lift, or it lifts and will not close.

The honest downside is that a valve sitting on years of scale sometimes refuses to reseat once it has been lifted, and then it is replaced on the visit. That is not the engineer breaking it. It is the valve telling you it would not have worked, found with the isolation valve within reach rather than at two in the morning.

A pressure relief valve leaking after a service is almost always a valve that had already failed and was being held shut by scale until somebody operated it.

Never cap, plug or tighten a relief valve to stop a drip. The valve is the last thing between a sealed vessel of near boiling water and the room it stands in. A valve that keeps passing is reporting a fault upstream of it.

Temperature control, the thermostat and the cut out behind it

Three devices control heat in order, and the service confirms all three rather than just the one you can feel working.

The cylinder thermostat is the everyday control, and the stored temperature is measured rather than assumed. Around sixty degrees is the target. Much lower and water is being stored in the range where bacteria are comfortable. Much higher and it is scalding at the tap and scaling the cylinder faster than it needs to. Behind the thermostat sits the thermal cut out, a manual reset device whose only job is to kill the heat source if the thermostat fails closed. It is checked for continuity and for whether it has already tripped, and a tripped cut out is a finding rather than a reset and move on job, because whatever drove the temperature that high is still there.

On indirect cylinders the motorised valve and its microswitch are checked, since a valve that fails to open is the reason for the common complaint of heating running normally while the hot water stays cold. On electric cylinders the immersion element is checked for insulation resistance and continuity, and the head is opened to see whether the element is furring.

The tundish and the discharge route

The tundish is the small clear funnel in the pipework beside the cylinder. It exists so a discharge is visible and so there is an air break between the cylinder and the drain. The check is short: clear of scale, visible rather than boxed in behind a panel, and nothing passing through it while the system sits idle and hot.

The pipe leaving it is followed by eye as far as it can be followed. What the engineer wants to see is a continuous fall, a material rated for near boiling water, nothing added to the run, and a termination where the discharge can be seen and cannot scald anyone. In converted London flats the usual finding is a discharge routed into a hopper or a waste pipe during a bathroom refit, which is exactly what it must not do. Staining or scale trails in the tundish are recorded even when nothing runs on the day, because they are evidence of intermittent discharge and they point back at the expansion cushion or the reducing valve. If your tundish is dripping now rather than at the next service, that symptom has its own short list of causes and is worth reading before you book anything.

What else the engineer checks on the cylinder itself

The rest of the visit is inspection rather than testing, and it is where most replacement decisions come from. The cylinder body and every connection is checked dry by hand and by eye, with attention to the immersion boss and the seams, because a stainless cylinder that has begun to weep does not get better. Isolation valves are operated, since a valve that will not turn is a problem you want found now and not during a leak. Where the design carries a sacrificial anode it is inspected, because a consumed anode stops protecting the vessel and leaves the steel to do the work.

Scale is judged against what London water does. Most of the capital sits on hard water and scale collects first on the hottest surfaces, which are the immersion element and the primary coil. A cylinder that takes noticeably longer to reheat than it used to, with the boiler firing normally, is usually reporting a coated coil. What London hard water does to a cylinder over its working life, and what slows it down, is covered separately.

At the end the findings should be written against the cylinder: the expansion charge read and what it was set back to, which valves were exercised and how they behaved, the stored temperature measured, and anything left outstanding. Ask for that record. It makes the next service faster and it turns a slow decline into something you can see year on year.

What a service will not do, and what it usually finds

A service is not a repair and it is not a descale. If the expansion vessel has failed, the reducing valve has gone, a relief valve will not reseat or an immersion element is dead, those are parts and they are dealt with separately once you know about them. A good engineer tells you on the day which is worth doing now and which can wait. It will not rescue a cylinder already leaking from the body either, because stainless cylinders fail at welds and bosses, and once the vessel itself weeps there is nothing serviceable about it. The signs that a cylinder is near the end rather than worth another set of parts are worth knowing before you agree to either.

The three findings that come up most often are a flat expansion vessel or a lost air bubble, a blocked strainer in the inlet group, and a relief valve scaled onto its seat. All three are cheap and quick while they are findings. All three become an emergency later.

A service is a check rather than a repair, so where the engineer finds a failed expansion vessel or a valve that will no longer seat, the work that follows is an unvented cylinder repair, quoted separately and agreed before anything is fitted.

It will also not defer a hot water cylinder replacement that the vessel has already earned, because a service records condition rather than restoring it.

On cost, the labour rate is fixed before we attend and the total depends on how long the job takes and whether parts are needed. Nobody should be quoting a figure for parts before seeing which valve set and which vessel you have.