Calibration of Thermometers and Temperature Sensitive Devices

[Garner Davis - 2011]

Hello, all.

I've searched this forum for some ideas on calibrating thermometers, and I've found some good information. I used to work for a large company, and our system was to purchase the expensive NIST-traceable thermometer and use it to check the accuracy of other thermometers, as well as ovens, water baths, etc.

Now I work for a much smaller company, and spending that type of money is not an option. I've seen an idea in one thread that seemed to make sense for thermometers, which involved placing all the devices in the same environment for a given amount of time, taking a reading, and as long as they fell within acceptable tolerances for the process, everything was OK. Repeat for different temperatures as necessary.

Now, I want to extend that idea to include the devices such as the water baths, ovens, etc. If I select one of the thermometers to use as a "golden" standard, would it be acceptable to use it to check the accuracy of the other devices. My personal thought is that this should be OK, as long as it's defined within the quality system, but I welcome feedback from the group if there's something I might have missed.

Thanks for your time;

Garner Davis

[Al Rosen]

In Reply to Parent Post by Garner Davis

Hello, all.

I've searched this forum for some ideas on calibrating thermometers, and I've found some good information. I used to work for a large company, and our system was to purchase the expensive NIST-traceable thermometer and use it to check the accuracy of other thermometers, as well as ovens, water baths, etc.

Now I work for a much smaller company, and spending that type of money is not an option. I've seen an idea in one thread that seemed to make sense for thermometers, which involved placing all the devices in the same environment for a given amount of time, taking a reading, and as long as they fell within acceptable tolerances for the process, everything was OK. Repeat for different temperatures as necessary.

Now, I want to extend that idea to include the devices such as the water baths, ovens, etc. If I select one of the thermometers to use as a "golden" standard, would it be acceptable to use it to check the accuracy of the other devices. My personal thought is that this should be OK, as long as it's defined within the quality system, but I welcome feedback from the group if there's something I might have missed.

Thanks for your time;

Garner Davis

This is not an acceptable way to calibrate your devices. You should get one device calibrated with traceability to NIST and use it exclusively for calibrating all the others.

[Craig H.]

In Reply to Parent Post by Al Rosen

This is not an acceptable way to calibrate your devices. You should get one device calibrated with traceability to NIST and use it exclusively for calibrating all the others.

Garner,

I agree with Al. Would it be possible to send one of your existing thermometers off to be checked? That could become your reference thermometer.

[Garner Davis - 2011]

Thank you for the responses; I'll investigate what the cost would be to do that.

I remember that the cost of a NIST-traceable thermometer from Fisher Scientific and the like seemed to run around $1K; if we're looking at about the same cost for sending one off, then I'm not sure that's feasible for us at this juncture.

Thanks again for the responses.

Garner

[Al Rosen]

In Reply to Parent Post by Garner Davis

Thank you for the responses; I'll investigate what the cost would be to do that.

I remember that the cost of a NIST-traceable thermometer from Fisher Scientific and the like seemed to run around $1K; if we're looking at about the same cost for sending one off, then I'm not sure that's feasible for us at this juncture.

Thanks again for the responses.

Garner

The cost to calibrate it is not anywhere near that much.

[pilchard]

Ice is 32°F or 0°C

A properly prepared ice-point bath can be used for thermometer calibration. Poor preparation and use of the bath will result in errors. With care, one can use common laboratory beakers and stirring equipment to produce an ice-point bath that is within ± 0.01 °C (0.02 °F) of the true ice point.

Procedure
The following table details preparing and using ice-point baths.

Preparing and Using Ice-point Baths
Step Action
1 The water must be uniformly cooled to and maintained at the ice-point temperature.
¨ This can be done only by using slush made of ice shavings or crushed ice particles.
¨ Constantly stir the ice and water slush to keep it distributed uniformly.
¨ The slush must contain only enough water to wet the ice particles and allow the mixture to be slightly fluid.
¨ Water will have to be removed periodically and ice added as the bath is used.
¨ For best accuracy, use distilled water and clean clear ice.

2 Do not allow clumps of ice or water pockets to form.
¨ When ice particles clump together, the water between them freezes and the temperature drops.
¨ If a pocket of water forms, the temperature in it will rise.
¨ A water and ice cube mixture will not cool the water between the cubes sufficiently.

3 Do not allow a thermometer or temperature sensor in the ice-point bath to touch any part of the bath or any other thermometer or sensor. At least 25 mm (1 in.) of slush shall always be below the lowest point of any thermometer.

4 Suspend each thermometer from above and hold at the proper immersion depth. It will be easier to read the liquid column of a glass thermometer if the container is almost full.

5 Constantly stir the bath in order to uniformly distribute the ice particles and water.

6 Allow the thermometer or sensor to stabilize at the temperature of the bath.

7 ¨ Read the thermometer, wait for about a minute, and read again.
¨ Record only those readings where the indicated temperature is repeated within the desired accuracy.

NOTE: You may stop stirring momentarily while reading a glass thermometer.
8 Record the indicated value.

9 Find the error by subtracting the ice point (i.e., 0 °C [32 °F]) from the indicated value.


The correction factor has the same numerical value as the error but with opposite sign.



EXAMPLE: Let the true temperature be 0 ° and the indicated temperature be -0.40 °; the thermometer reads low with an error of -0.40 °; the correction factor to apply at that point is + 0.40 °.

[Al Rosen]

Now you have a device calibrated at one point. What about the rest of the range?

[pilchard]

In Reply to Parent Post by Al Rosen

Now you have a device calibrated at one point. What about the rest of the range?

You buy a NIST Thermometer!

You use a water boiling point calculate at you altitude and barometric pressure and use the same procedure.

I would rather use Gumbo at least if it dont work you can feed the auditor!