How the two differ in kind
Two complaints get described the same way and they have opposite remedies, so the diagnosis is the whole job.
WAITING is a distribution fault. The water in the pipe between the heater and the tap has cooled to room temperature since it was last used, and when you open the tap you have to push all of it out before hot water arrives. How long that takes is simply the VOLUME of the pipe run divided by the flow rate — which is why a long run of large-bore pipe to a distant bathroom can waste a startling quantity of water before anything warm appears, and why the same house delivers hot water instantly at the basin next to the cylinder. Nothing about the heater is implicated: the hot water was there the whole time.
RUNNING OUT is a capacity fault. The heater cannot deliver hot water as fast as it is being drawn for as long as it is being drawn — either the stored volume is exhausted or, on an instantaneous heater, the flow rate exceeds what it can raise to temperature. Recovery rate and storage volume are the levers.
The remedies do not overlap, and one of them works slightly against the other. A RECIRCULATION loop keeps hot water moving through the distribution pipework so it is already hot when a tap opens, which eliminates the wait and the water wasted down the drain — and it is a continuous standing heat loss, because the loop is a radiator running through the building whenever it circulates. That makes the water heater work harder, which very slightly worsens a capacity problem. Increasing recovery or storage does nothing at all about the wait.
So the question to answer first is whether the water arrives slowly or stops.
The factors that actually differ
Show
| Recirculation loop | More recovery or storage | |
|---|---|---|
| The fault it fixes | Waiting — hot water taking too long to reach the tap. | Running out — hot water ceasing part way through the demand. |
| The physical cause | The volume of cold water sitting in the pipe run between heater and tap. | Storage volume, recovery rate, or the flow rate an instantaneous heater can raise to temperature. |
| Effect on the other problem | Slightly worsens capacity, because the loop is a continuous standing heat loss the heater has to make up. | None at all on the wait. A bigger cylinder full of hot water is still at the far end of the same cold pipe. |
| Water wasted | Eliminates it — the litres normally run down the drain while waiting. | Unchanged. |
| Energy | A continuous cost. Controls — a timer, a demand button, a thermostat, or all three — are what make it defensible. | Proportional to use, plus standing loss from a larger cylinder. |
| Pipe insulation | Essential rather than optional. An uninsulated loop is an expensive way to heat the ceiling void. | Worth doing, and far less critical. |
| Retrofit difficulty | A dedicated return needs a pipe run back to the heater. Where that is impractical, a crossover valve at the far fixture uses the cold line as the return. | Straightforward — a larger cylinder, a higher-output heater, or a second unit. |
| Side effect to watch | A crossover arrangement warms the cold water line at that fixture, which some people find unacceptable. | A larger cylinder means a longer initial reheat and more standing loss. |
| Cheapest first move | Insulate the run and check the pipe size — oversized pipe is a large part of the wait. | Check the thermostat setting and the heater's actual recovery before assuming it is undersized. |
| Legionella and safety | A loop must keep the return temperature above the range where bacteria proliferate; this is a design requirement, not an option. | Storage temperature is set for the same reason, with a blending valve to deliver safe temperature at the tap. |
Which one, and when
Choose recirculation loop when…
- The complaint is the wait, and the run to the fixture is long.
- Water is metered or scarce, where the litres run to waste have a direct cost.
- It is a large building or a long horizontal distribution where the wait is measured in tens of seconds.
- There is a return path available, or a crossover arrangement at the far fixture is acceptable.
Choose more recovery or storage when…
- The complaint is running out — the shower goes cold, or two draws cannot happen at once.
- Demand has changed: a new bathroom, more occupants, a higher-flow shower.
- The heater is genuinely undersized for the peak simultaneous demand rather than merely old.
- An instantaneous heater cannot raise the flow being drawn, which is a flow-rate limit rather than a volume one.
Now run your own numbers
This page holds no prices on purpose — a national average is wrong for almost every real project. Quantify both options with your dimensions and your local quotes.
Frequently asked questions
- How do I tell which problem I have?
- Time it and watch what happens. If you open the tap, wait, and hot water eventually arrives and then stays hot indefinitely, it is a distribution problem — the delay is the pipe run emptying, and no amount of extra heater will change it. If hot water arrives promptly and then goes cold part way through a shower, or the second shower is cold, it is a capacity problem and recirculation will not help. A useful second measurement is to catch the water run off while waiting: the volume you collect is the volume of the pipe run, and comparing it against the pipe's length and bore usually explains the wait completely. Both faults can coexist in one house — a distant bathroom that waits and a cylinder that runs out — in which case they are two separate jobs.
- Why does a bigger pipe make the wait worse?
- Because the wait is the volume of the run divided by the flow rate, and volume scales with the square of the diameter. Doubling the bore of a pipe quadruples the water sitting in it, so a generously sized run to a distant fixture holds far more cold water to push out than a correctly sized one and takes correspondingly longer to deliver. Oversized distribution pipework is extremely common, partly from caution and partly because larger pipe was assumed to guarantee flow. The consequence is a delay and a volume of water run to waste at every single draw, for the life of the building. Where pipework is being replaced, sizing the branches properly — and running smaller dedicated branches to individual fixtures rather than one large main — is the cheapest possible improvement to the wait.
- What are the ways to do recirculation?
- A dedicated return is the proper arrangement: a separate pipe from the far end of the distribution back to the heater, with a pump circulating through it, so hot water is always present in the flow pipe. It works well and needs a pipe route, which a retrofit often does not have. The alternative is a crossover valve installed under the furthest fixture, which opens a path from the hot line into the COLD line so the cold pipe acts as the return — no new pipework, but it warms the cold water at that fixture, which some people dislike and which is unacceptable where cold water must stay cold. The third arrangement is a demand system: a button or sensor runs the pump only when somebody is about to use the water, which removes almost all of the standing loss at the cost of a short wait.
- How much energy does a recirculation loop cost?
- More than people expect if it runs continuously and is not insulated, and very little if it is controlled and lagged. The loop is a pipe full of hot water running through the building, losing heat along its whole length whenever it circulates, and the heater makes that loss up around the clock. Three things cut it down. Insulating the loop properly — all of it, including the return and the fittings — is the single biggest factor and is not optional on a circulated line. Controlling it so it runs only when hot water is plausibly wanted: a timer for the household's pattern, a thermostat that stops the pump once the return is hot, or a demand control that runs it only when asked. And keeping the loop short, since the loss is proportional to its length.
- What actually limits an instantaneous heater?
- Flow rate at a given temperature rise, not volume — which is why an instantaneous unit never runs out and can still be inadequate. Its output is a fixed power, and that power divided by the temperature rise required gives the maximum flow it can deliver. Two things change that: the incoming water temperature, which is much colder in winter, so the same unit delivers meaningfully less flow in the season you want it most; and simultaneous demand, since two fixtures drawing at once share the same capacity. This is the reason a unit that seems fine in summer disappoints in winter, and the reason sizing is done against the winter incoming temperature and the peak simultaneous draw rather than against an annual average.
- Is turning the thermostat up a valid fix?
- It increases effective capacity and it introduces a scalding risk, so it is done in combination with a blending valve rather than on its own. Storing hotter means each litre of stored water blends with more cold to reach the temperature actually used, so the usable volume goes up without a bigger cylinder — a genuine effect. The safety problem is that water delivered at storage temperature can scald quickly, which is why a thermostatic mixing or tempering valve on the outlet is the standard arrangement: store hot, deliver safe. There is also a lower limit that runs the other way: storage temperature must stay high enough to control bacterial growth, which is why turning a cylinder down to save energy is bounded. Both constraints together define a fairly narrow band, and the valve is what makes it usable.
- Does a recirculation loop create a legionella risk?
- It creates a system that must be designed and maintained with that in mind, which is a requirement rather than a caution. Bacteria proliferate in a temperature band between roughly lukewarm and hot, so the design intent is to keep stored and circulated water above that band and delivered water below it through a blending valve. For a loop, that means the RETURN temperature — the coolest point in the circulation — has to stay above the threshold, which is a design calculation about loop length, insulation and flow rate rather than something to hope for. Dead legs are the other hazard: a branch that is not circulated and is rarely used sits at room temperature indefinitely. In domestic settings this is usually handled by following good practice; in healthcare and larger buildings it is a regulated regime with monitoring.
- What should I try before either?
- Insulation and measurement, both of which are cheap and both of which change the answer surprisingly often. Insulating the accessible hot water pipework reduces the standing loss and keeps the run warm for longer between draws, which shortens the wait on a second use meaningfully even with no pump. Measuring is the other half: time the wait at the worst fixture, catch and measure the water run off while waiting, check the cylinder's thermostat setting against what it is actually delivering, and note whether the complaint is about arrival or about running out. That handful of observations distinguishes the two faults definitively and frequently identifies a third that is neither — a partly closed valve, a scaled heat exchanger, a failed diverter, or a thermostat set low by somebody trying to save money.