Which tests tell sensor drift apart from changes in the operating environment?

5 min read

When a breathalyzer's readings differ from earlier records, safety managers and quality or engineering staff have to decide whether the sensor has changed or the operating environment was different. In a device that detects breath alcohol and turns it into a reading, the result is not determined by the sensor alone. If the reference sample used for comparison, the test procedure, surrounding conditions such as temperature and humidity, or the device configuration including the sensor and software changes, the same difference in readings can have a different cause.

Here, sensor drift is a phenomenon in which the measuring device's own metrological properties change over time, so its indications change continuously or in steps. A change in the operating environment, by contrast, is a change in indication caused by a change in known influence conditions such as temperature and humidity. Rather than concluding which one it is from site records alone, distinguish them according to what a test held fixed and what it changed.

The passage of time is not evidence of drift

The International Vocabulary of Metrology (VIM) describes instrumental drift as a change in a measuring instrument's properties that is not related to a change in the quantity being measured or to a change in any recognized influence quantity. So the fact that spring and winter results differ is only a signal that a comparison is needed. It does not immediately mean sensor deterioration or failure.

If temperature and humidity differed between the two points in time, the possibility of an environmental effect remains. If there was a sensor replacement, software change, or settings change, the grounds for linking the before and after results as one continuous sensor change also weaken. Time is one of the axes of comparison, and it can be used in a drift review only after the other conditions have been checked.

Tests for sensor change keep the comparison conditions the same

The question in a drift test is simple: does the same device still indicate the same way for the same reference after time has passed? A reference sample is a sample with a defined value used as the basis for comparison. In this test, compare indications at different points in time while keeping the device configuration, the reference sample and its reference value, the test procedure, and known environmental conditions as constant as possible.

What you compare is not just the final average. Only by recording each test's indication, whether the test finished normally, and the time and conditions of the test can you distinguish fluctuation in individual results from differences between points in time. Even if a difference appears in one comparison, it does not establish a long-term rate of change or the sensor's lifespan.

Calibration and adjustment should not be mixed up either. Calibration is the work of establishing, under specified conditions, the relationship between reference values and the instrument's indications, while adjustment is the work of changing the instrument so that it gives specified indications. If an adjustment took place, it is right to treat the indications before and after it not as simple time-comparison data but separately, as data from a device that has been changed.

Environmental influence tests change one condition at a time

An environmental influence test checks how indications change when a particular condition changes. Here, keep the device, the reference sample, and the test procedure the same, and vary one condition at a time, such as temperature or humidity, that you want to check. If temperature, humidity, and the test procedure all change at once, it is hard to interpret which change a difference in indications is related to.

The results observed under the original condition and under the changed condition each show the response under that condition. Comparing the indications after returning the condition to its original state as well reduces the chance of interpreting temporary differences caused by environmental conditions together with changes in device properties over time. However, this comparison also does not automatically lead to the conclusion that the sensor has changed permanently.

Question to check What to keep or change in the test Scope of what can be interpreted
Have the device's properties changed over time? Keep the device configuration, reference sample, procedure, and known environmental conditions Consider the possibility of a change in device properties when conditions were actually the same
Does a particular environment affect readings? Keep the device and reference sample, and change one environmental condition Observe the relationship between the changed condition and the difference in indications
What is the difference in site records? Compare location, procedure, unit, device configuration, and surrounding conditions Confirm that a difference exists without concluding that the sensor is the cause

Use hypothetical cases to narrow down candidate causes

As a hypothetical case, suppose the same breathalyzer gave a different indication when measuring a reference sample in winter than in spring. The only observed fact is that the indications at the two points in time differ. If you do not know whether the winter test's temperature and humidity, the reference sample and reference value, the test procedure, and whether the sensor or software was changed were the same as in spring, it is not yet the stage for interpreting it as sensor drift.

First, you can line up and compare the device, reference sample and reference value, test procedure, surrounding conditions, and any changes in between for the two tests. If a temperature or humidity difference is confirmed here, one approach is to plan separately a reference-condition comparison that matches those conditions and an environmental influence comparison that varies only that condition. Conversely, if differences in indication recur even in comparisons that keep conditions the same, data builds up for considering the possibility of a change in device properties.

The completion criterion in this case is not hastily judging the winter result as normal or abnormal. The items kept the same, the items that changed, and the items not confirmed among the comparison conditions must be distinguished. Only then can you decide which conditions to fix and which to deliberately vary in the next test.

Record comparable facts rather than conclusions

It is useful for review materials to record, by date, the device identification, the reference sample and reference value, the test procedure, the temperature and humidity conditions, individual indications, and any changes in between. Then, writing down separately the conditions kept the same, the conditions that changed, and the conditions you could not confirm helps avoid reading site results and reference-condition test results as meaning the same thing.

Whether sensor replacement, repair, or calibration is needed is not decided automatically by this comparison alone. You can review the grounds for action only when results repeatedly observed under reference conditions are combined with information about actual changes and inspections of the device. Sensor drift and changes in the operating environment can be told apart when they are separated into tests that each answer a different question.

References

#sensor drift
#environmental effects
#breathalyzer testing
Considering our products?
We'll help you decide.