A thermometer held against a tap at half past eleven on a Tuesday tells you one thing: that outlet was hot at half past eleven on a Tuesday. It tells you nothing about the previous night, the bank holiday weekend just gone, or the three weeks the room sat empty before someone finally turned it on. A great deal of temperature risk in a building hides in exactly those gaps, and a single reading is, by its nature, blind to all of them.

Data loggers exist to close that blindness. They clip onto a pipe or sit in a tank and record temperature on a schedule, every few minutes, around the clock, so you can see the shape of how a system behaves rather than one frozen frame of it. The useful question is not whether a logger is more accurate than a handheld probe. It is what the continuous record shows that a monthly sentinel check, by design, never can.

A care home that kept passing its checks

What follows is a composite, illustrative scenario, drawn from the same trap that catches many sites, not a real named case.

Picture a small residential care home that took its monitoring seriously. A contractor visited on the first Monday of every month and recorded sentinel temperatures: hot outlets comfortably above target, cold outlets sitting below 20 degrees C, every figure entered in the logbook. Month after month it was a clean sheet. On paper the water system was under control.

After a near-miss at another site in the group, the risk assessment recommended trialling continuous temperature monitoring on a few points. Two loggers went onto cold-water sentinel outlets, one into the cold-water storage cistern up in the roof void, and one on the hot return near the calorifier. The team left them recording for three weeks and then pulled the data.

What three weeks of logging showed

The hot side was reassuringly dull. The cold side was not.

The cistern in the roof void was below 20 degrees C every single morning the contractor had ever visited, which is precisely why the monthly check had never flagged it. But the logger told a different story across the day. On warm afternoons the loft heated up, and the stored water drifted with it, peaking in the low-to-mid 20s before cooling back overnight. Every monthly reading had landed in the cool morning window and missed the daily climb entirely.

One of the cold sentinel outlets showed a quieter version of the same problem: a brief warm spike each lunchtime, when heavy draw-off on a poorly separated hot pipe alongside it bled heat across during the busiest hour.

None of this was a catastrophe on any given day. The pattern was the point. The system was spending hours each day inside the temperature band where Legionella multiplies, and the monitoring regime had been sampling the one time of day it happened to look fine. That is the whole argument for logging: you are no longer judging the system on its best morning, you are watching how it actually holds temperature when nobody is standing in front of it with a probe.

The decisions that actually changed the risk

The fix was not “buy more loggers”. The continuous record simply made three decisions obvious that a clean monthly sheet had hidden:

  • Deal with the root cause, not the reading. The cistern problem was the roof void, not the water. Insulating and managing that storage, or relocating it out of the heat path on the next planned upgrade, addresses the cause rather than chasing the symptom.
  • Separate hot and cold properly. The lunchtime spike on the cold sentinel pointed at a length of cold pipe running too close to a busy hot line. Re-routing or insulating that section removes the daily heat gain.
  • Watch the worst point continuously. Rather than move the monthly visit to a less convenient hour, the team left a logger on the point most likely to fail, so the daily peak is always captured and the trend is reviewed, not just the snapshot.

That last decision is the one that transfers everywhere. Spot checks and continuous monitoring answer different questions, and a sensible scheme uses both.

Reading loggers well, not just installing them

A logger is only as good as where you put it and how you read it. A few things separate a useful deployment from an expensive way to generate ignored graphs.

Put them where control is actually decided. The points that earn a logger first are cold-water storage sitting in warm spaces, the hot return where circulation either holds or sags, and whichever sentinel outlet your assessment treats as the hardest to keep in range. The temperature targets those points are judged against are the familiar HSE benchmarks for hot and cold water systems: cold water kept below 20 degrees C where possible, hot water stored at 60 degrees C, and hot water reaching around 50 degrees C at the outlet within a minute, or 55 degrees C in healthcare premises [1]. Treat those as the common guidance values; your risk assessment sets the figures and the monitoring frequency for your specific system [2].

Read trends, not single points. The value of a logger is the line over time. An outlet that touches target once an hour but spends the rest of the time tepid is telling you something a pass/fail tick cannot.

Mind calibration. A logger reading two degrees low is worse than no logger at all, because you will trust it. Cross-check each unit against a calibrated reference probe periodically so a slow drift cannot quietly turn good data into false comfort.

Logging is the backbone of temperature control as a Legionella control measure, but it sits inside the wider scheme rather than replacing it [3]. The numbers prove what happened; people still have to decide what it means and act on it. For how that feeds into deciding what to fix first, see Acting on your risk assessment: setting priorities.

Before you trust the graphs

Continuous data is persuasive, which is exactly why it deserves a caution of its own. A logger records a temperature at a sensor; it does not confirm the sensor is well placed, in calibration, or measuring the water rather than the warm pipe clip it is strapped to. Nor does a tidy three weeks of in-range readings discharge any duty: the targets, the points that matter, and the response to an out-of-range trend all come from a competent, site-specific assessment, not from the device or from the benchmark figures quoted here. Use the record to inform that assessment, never to stand in for it.

FAQ

Do data loggers replace manual sentinel temperature checks?

No. They answer a different question. Sentinel checks confirm a point is in range at the moment of measurement and keep a human looking at the system; loggers reveal how that point behaves between visits. Most schemes keep manual checks and add logging where the risk assessment says continuous evidence is worth having.

If I can only fit a few loggers, where do they go first?

Start where temperature is most likely to drift unseen: cold-water storage in roof voids or warm plant rooms, the hot return where poor circulation shows up first, and the single sentinel outlet your assessment already treats as the hardest to keep in range. One logger on a genuine problem point beats a dozen on outlets that pass anyway.

How would I know a logger had drifted out of calibration?

You would not, unless you checked. Compare each unit against a calibrated reference thermometer at sensible intervals, and be suspicious of a sensor that reads suspiciously steady or has quietly diverged from a neighbouring point that should track it. Build that cross-check into the routine so a slow drift cannot turn into months of false reassurance.

Sources

[1] HSE, “Hot and cold water systems”. https://www.hse.gov.uk/legionnaires/hot-and-cold.htm [2] HSE, “Legionnaires’ disease: Technical guidance (HSG274)”. https://www.hse.gov.uk/pubns/books/hsg274.htm [3] HSE, “Legionnaires’ disease. The control of legionella bacteria in water systems - Approved Code of Practice and guidance (L8)”. https://www.hse.gov.uk/pubns/books/l8.htm