LogTag Spike Analysis

LogTag Spike Analysis

Issue

A single -5°C reading has been recorded by the LogTag monitoring a vaccine fridge. If this is a valid reading then the vaccines are potentially compromised and need to be destroyed.

Looking at the results around the single -5°C/18°C reading, there is a distinct cycle that is very consistent in size and period (15 to 20 minutes).

Is it physically possible to cause the spike?

In short, no.

It is not physically possible for a fridge to significantly drop or rise 9°C in 5 minutes without an extreme event occurring. The thermal latency of the LogTag also means that the device won’t respond instantly to changes. Assuming that the sudden step down was legitimate, it is almost impossible for the fridge to then jump back to normal conditions again.

What happens if you ignore the spike?

graph spike covered

If this single reading was to be ignored, the shape of the graph is totally consistent.

Therefore, taking everything into account, the only logical course of action is to ignore the outlier.

Possible cause

  1. The LogTag uses an external probe to measure temperature. If the probe is disconnected it will read -40°C. If, however, the connection is compromised then it is possible to get a misreading. This is very unusual but is possible. A momentary bump on the cable/connector or possibly a severe EMF noise spike may have caused it. Since it is a one-off, it is not possible to determine the cause.
  2. If it happening consistently, then the LogTag may be faulty, or a low battery may be causing strange results.
  3. Whilst doing a download, a reading is recorded showing a high temperature. This is due to the LogTag taking a sample from the room temperature. A LogTag will continue to record its set sample rate, even when away from the fridge.

Summary

On the basis of the data around the reading, the reading is obviously false and can be classified as a misread. There has been no compromise to the quality of the fridge’s contents.

If this is a one-off event, ignore it completely. If it happens a number of times, replace the LogTag.