On this page
Why One Probe in the Middle Is Not Enough
Most Kenyan distributors monitor a cold room with a single probe, usually mounted on a wall near the door because that is where the cabling was easy. The problem is that a cold room is not one temperature. Air stratifies, the compressor cools unevenly, and the volume near the door warms every time a forklift breaks the seal during goods-in. The spot you happen to monitor may be the calmest part of the room, while product stacked in a far corner sits warmer than your records ever show.
Temperature mapping is how you find out. You place a grid of calibrated sensors throughout the room, run them for a defined period, and prove where the warm and cold extremes actually are. PPB Good Distribution Practice treats mapping and qualification of cold storage as a standard expectation, not an optional extra: an inspector wants to see that you identified the worst-case location and placed your routine monitoring probe there, so the number on your daily log represents the hardest condition any product faces.
The Cold Room Mapping Analyzer reads the multi-sensor file from that study, computes per-sensor statistics, identifies the hotspot and coldspot, calls the study pass or fail against your approved limits, and tells you where the routine probe belongs. This guide walks one realistic study end to end and shows how to write it up. It is a companion to your qualification SOP, not a replacement for it.
Worked Example: A 9-Sensor Study of a 40 Square Metre Cold Room
Take a distributor running a 40 square metre 2 to 8C cold room. The team places 9 calibrated sensors on a three by three grid covering low, mid, and high shelf levels and all four corners, then logs at 5-minute intervals for 72 hours under normal loading, including two routine goods-in shifts. That 72-hour window is the minimum that captures a full day-night cycle plus door-opening traffic, so the study reflects how the room really behaves rather than a quiet afternoon.
When the file goes into the analyzer, eight of the nine sensors cluster tightly:
- Eight sensors near the centre and back: means between 4.2C and 5.0C, sitting around an overall 4.6C, with no readings above 8C across the full 72 hours. Comfortable, stable, well inside the 2 to 8C window.
- One corner sensor near the door: a mean of 7.8C, about 1.5C warmer than the room average, peaking above 8C during every goods-in shift and spending roughly 6 percent of the 72 hours above the 8.0C limit.
That single sensor changes the verdict. The analyzer surfaces the door corner as the hotspot, flags the study as a fail against the 8C action limit (because a real storage location breaches the limit 6 percent of the time), and identifies that corner as the worst-case position. The corrective actions follow directly: relocate the routine monitoring probe to the door corner so daily logs reflect the worst case, adjust loading and airflow so product is not packed tight against that corner where it blocks circulation, and mark the hotspot as a no-store zone or add an air deflector and re-map to confirm the fix. You did not need a thermal engineer to find the weak point. The grid found it, and the analyzer named it.
What Data You Need and Why
Three things drive a defensible mapping result: a multi-sensor data file with one timestamp column and one column per sensor, clear sensor IDs and their physical positions (which grid point, which shelf level, which corner), and your approved storage limits (the 2 to 8C from your product labels and SOP, not a guess). Without the position map, the analyzer can tell you sensor 7 is the hotspot, but only your layout note tells you sensor 7 is the door corner you need to act on.
The minimum run is 72 hours at your normal logging interval, covering at least one full day-night cycle and the routine door-opening traffic of a working shift. Shorter windows miss the goods-in spikes and the warm part of the daily cycle, which is exactly where a hotspot reveals itself. Export raw interval readings from your logger platform (LogTag, Testo, ELPRO, DicksonOne and similar all export CSV), not chart images or daily averages. Averages flatten the very peaks that decide pass or fail.
One control matters most: every sensor must be calibrated before the study and the positions recorded as you place them, ideally with a labelled floor sketch. If a sensor drifts mid-study your hotspot may be an artefact, so check your loggers against a reference first. The Sensor Drift and Calibration Alert Tool helps you confirm none of the mapping sensors are reading false, and if an export comes through messy, the Data Logger QA Cleaner tidies it before upload.
Reading the Outputs
The analyzer returns several outputs. Read them top down:
- Per-sensor statistics: mean, minimum, maximum, and time-above-limit for each sensor. This is where the clustered eight and the lone warm corner separate out at a glance.
- Hotspot and coldspot identification: the warmest and coldest sensors by mean and by peak. The hotspot is the location your routine probe should occupy; the coldspot warns you where product nearest the cooling unit might risk freezing below 2C.
- Pass or fail against limits: the overall call. If any real storage location breaches the approved limit beyond your acceptance criterion, the study fails and triggers corrective action and a re-map.
- Recommended probe location: the worst-case position your daily monitoring should track, so the single number on your log represents the hardest condition in the room.
Never read these in isolation. Pair the numbers with operational context: which corner is the hotspot, how often that corner sees the door open, the loading pattern around it, and defrost timing. A hotspot driven by door traffic is a loading and airflow fix; a hotspot that persists with the door shut points at the cooling unit or an under-sized evaporator. If you later see frequent alarms from the relocated probe, that is expected (it now sits in the worst case), and the Alarm Threshold Optimizer helps you tune setpoints so transient door spikes do not bury a real failure in noise.
Documenting the Study for PPB and GDP Audits
A mapping study is a qualification record. Capture it the way an inspector expects to see it:
- Attach the raw multi-sensor export, the cleaned file used for analysis, and a labelled floor sketch showing every sensor ID and its position.
- Record the study window (72 hours minimum), the logging interval, the loading state, and the calibration certificates for each sensor.
- State the per-sensor statistics, the identified hotspot and coldspot, the pass or fail call against your approved limits, and your acceptance criterion (for example, no real storage location above 8C).
- Document the corrective actions, the relocated routine probe position, and the date of any re-map that confirmed the fix.
- Set a requalification frequency (annually, and on any change to the room, racking, cooling unit, or loading pattern), and route changes through change control with QA sign-off so the study stays defensible.
If you are still building these records, our GMP readiness checklist covers the qualification and monitoring documentation an inspector will ask for.
Frequently Asked Questions
How many sensors do I need to map a cold room?
Enough to cover the corners, the centre, and high and low shelf levels. A 40 square metre room is well served by a 9-sensor three by three grid; larger or irregular rooms need more. The aim is to reach every location product could sit, including the door corner and the air nearest the cooling unit.
How long should a mapping study run?
At least 72 hours at your normal logging interval, covering a full day-night cycle and the routine door-opening traffic of a working shift. Shorter runs miss the goods-in spikes and the warm part of the daily cycle, which is exactly where a hotspot reveals itself.
My study failed because one corner runs warm. What do I do?
Treat the warm corner as the worst-case location. Relocate your routine monitoring probe there, adjust loading and airflow so product does not block circulation in that corner, consider marking it a no-store zone or fitting an air deflector, then re-map to confirm the fix before closing the corrective action.
Where should my permanent monitoring probe go after mapping?
At the hotspot the study identifies, not the most convenient wall. Placing the routine probe at the worst-case location means the single number on your daily log represents the hardest condition any product in the room actually faces, which is what PPB GDP expects.
How often does PPB expect a cold room to be re-mapped?
Re-qualify at least annually, and additionally after any change to the room, racking, cooling unit, or loading pattern. Route the requalification through change control with QA sign-off so the study and the monitoring location stay audit-defensible.
Run the Tool with Your Own Data
Upload your multi-sensor mapping file and get per-sensor statistics, hotspot and coldspot identification, a pass/fail call against your limits, and a recommended probe location you can take straight into your qualification record.
Open Cold Room Mapping Analyzer