Blog · ColdWatch
Before you buy new sensors, use the ones you already have.
Many monitoring projects begin with a purchase order for new loggers. We think the first question should be what the building already measures.
When a site decides it needs continuous temperature monitoring, the conversation often jumps straight to hardware: which loggers, how many, wireless or wired, which gateway. It feels like the natural start, because a monitoring system needs readings. But most pharma warehouses, labs, blood banks and food cold stores already have probes in their freezers and cold rooms, connected to a building-management or refrigeration controller. The readings exist. They are just not being watched, escalated or kept as evidence.
The gap is rarely the sensor
Look at the problems people actually describe. A freezer fails at 2 am and the 8 am round finds it. An alarm goes to one person who is asleep. An inspector asks for six months of evidence and it is scattered across paper, exports and email. None of these is caused by a missing probe. They are caused by what happens after the reading: nobody is told, nobody is escalated to, nothing is recorded in a form that survives an audit.
Buying new sensors does not fix any of that on its own. It adds a second set of readings, often in a second system, that still has to be watched.
The building already speaks a common language
Building-management and refrigeration equipment commonly exposes readings over BACnet/IP or Modbus TCP. ColdWatch polls freezers, cold rooms, product probes and door contacts from that existing field equipment. In the delivery plan, connecting the existing sensors and configuring limits and alarm rules is week one, not a procurement cycle.
| Starting point | What tends to happen |
|---|---|
| New loggers everywhere | A hardware project first; existing probes ignored; two sources of truth |
| Existing sensors connected first | Monitoring, escalation and audit trail on what is already installed; gaps filled where they actually exist |
When you genuinely need new hardware
We are not arguing that existing sensors are always enough. Some sites will find gaps once everything is on one screen:
- A unit with no probe at all, or one probe in a large cold room where product sits far from it.
- A sensor that cannot be calibrated or traced, which is a problem for evidence however good the software.
- Door contacts missing where door-open events would explain excursions.
- Equipment that does not expose readings over a supported protocol.
The difference is that these become specific, justified purchases found by looking at real data, rather than a blanket replacement made before anyone knew where the gaps were.
Software is where evidence is made
For a regulated cold chain, the readings matter less than what is done with them. That means alarms for high, low, rate of change, comms gaps and sensor failure; escalation by screen, email, SMS and push until someone acknowledges with a comment; an append-only, time-stamped ALCOA+ audit trail kept for five years; calibration due dates with reminders at 30, 15 and 7 days. IQ, OQ and PQ documents are delivered as part of the work, with installation and operational qualification run against your requirement specification in weeks two and three.
Whether a given sensor and set-up is acceptable for your products and inspections is a call for your QA team; confirm with your validation adviser.
Where we would start
- List every freezer, cold room and probe, and what controller each reports to.
- Check which readings are available over BACnet/IP or Modbus TCP.
- Check calibration status for each sensor before relying on it.
- Connect what exists and watch it for a few weeks.
- Buy hardware only for the gaps that shows.
For the calibration side, read the temperature sensor calibration tracking guide; for the bigger decision, see manual temperature logs vs continuous monitoring. All ColdWatch guides are on one page.
Questions
Do we need new temperature sensors for continuous monitoring?
Often not at first. Many freezers and cold rooms already have probes on a building-management or refrigeration controller that can be read over BACnet/IP or Modbus TCP.
When is new monitoring hardware justified?
When a unit has no probe, a sensor cannot be calibrated or traced, door contacts are missing, or equipment does not expose readings over a supported protocol.
Is connecting existing sensors enough for GxP?
It depends on calibration, qualification and your own requirements. Confirm with your QA team and validation adviser.
See it on your own data. ColdWatch — Every freezer, every minute. Book a 30-minute working session with an engineer.