How often to calibrate a thermometer?

[CostasV]

Hello everybody.

Usually, the question "how often to calibrate" has as many answers as the people that have been asked. The important thing is that an answer has to be documentated and based on the technical and some economical data. Well, my case is :

We have 6 temperature logger for monitoring the temperature inside refrigerators. Typically the temperatures in the refrig is between -12 and +6 deg C. The sensor is Pt1000 and the manufacturer says that it has an accuracy of +- 0.8 deg C (accuracy class 1) and that it can measure temperature from -30 to +50 deg C.

We can accept a difference between true and measured temperature up to 1.5 to 2.0 deg C.

How often should we calibrate such a thermometer?

Another question: Has anyone with similar conditions calibrated the thermometer more frequently than 3 or 4 years? It is supposed that the thermometer are carefully handled, cleaned etc. Has anyone found important differrences (beyond what it was expected) and what was the reason?

Any input is appreciated.

Thank you.
Costas

 

[hacksaw]

depends on the criticality of the temperature, if it is really critical then you'll need more than one sensor in each site.

set a calibration schedule based on the sensor drift rate specification, keep records, and adjust the calibration checks according to what the data tells you.

 

[IRstuff]

I agree with Hacksaw.

Often, calibration reports do not include information about the actual state of the instrument prior to calibration. This data is what you need to determine if the calibration intervals are reasonable.

When you do get your thermometers calibrated, make sure that you get information about whether the instrument was reading within specification or not. Once you have that data, you can then adjust your calibration intervals.

TTFN

 

[CostasV]

hacksaw and IRstuff,
thank you for your replies.

I have found in a website (I don't remember where) that a typical drift rate of Platinum temperature sensors is about 0.04% per year when the sensor is at 200 deg C for 4 years.

Taking into consideration that
A. our Pt1000 are new (only 2 months of operation, and manufactured in 2003) and
B. that the manufacturer says that they need to be calibrated every year and
C. that we can accept an error of 1.5 to 2.0 deg C which is almost twice as the given accuracy (0.8 deg C) of the sensor,
I think that the first calibration should be done after 3 years. When we will have the data from the first calibration, we will then reconsider the next calibration date.

Costas Varsamidis