Calibration requirements of temperature/RH loggers

[scottbp]

We have numerous temperature/RH dataloggers in most of our lab rooms where we do accredited calibrations. My questions are:

1. Do the dataloggers (or their probes) necessarily need to be accredited themselves?
2. Is a single point calibration at ambient conditions sufficient, or do we need to calibrate them for linearity at three points?
3. Do we necessarily need to report the datalogger/probe on calibration certificates?

From time to time we have assessors cite us for all three. Our argument against number 1 is that we are only monitoring the lab conditions for compliance to our quality system. Our argument against number 2 is that we have so many probes (49 at last count) spread across several locations that it's logistically difficult taking a unit out of service and shipping it off for calibration, and ensuring a spare is available to take it's place, so instead they're calibrated in place against a calibrated master probe. Our argument against number 3 is that the item being calibrated does not have any parameter traceable to the probe in that room that would matter (e.g. a 2% panel volt meter). For calibrations where temperature does matter (e.g. pressure using a deadweight tester, or resistance using a standard resistor), we use an accredited probe and report that probe on the certificate, but not the one on the wall.

So what say ye?

[ck454ss]

Here is my say yee. 

1. You say your data loggers don't need accredited calibration because you are only monitoring your lab conditions.  My question is what are you certified to?  And what quality system says its ok to take data from uncalibrated equipment?  Kind of defeats the purpose of a quality system.  Also what did you use to derive your uncertainties?

2.  To many probes to calibrate is a lame excuse in my book.  If you don't need them get rid of them, they are uncalibrated anyway.  As far as a single point cal you don't know the characterisitics of your probes without multiple points.  Your device could be way off above or below that calibration point.

3. I would say prove it.  You say temp/humidity doesn't matter to calibrations but it does matter to uncertainties of your calibration equipment and therefore matters to calibrations.  Its a reason I actually have two sets of uncertainties for my lab.  One for lab conditions and one for onsites.  And yes the datalogger should be on your cert.  If it isn't how did you come up the temp/humidity when you calibrated the equipment...the wet finger test?

By two bits

[Paul]

49 probes?? To an auditor this points that your company is trying to accomplish various types of disciplines- many of which the STABILITY of the environment is just as critical as the actual temperature of the UUT and the STD (as well as their stability) at the time of test. So I completely agree with the first response. Prove that you care and are considering each discipline (via limits or some other control) and prove they are working. In the case of the dead-weight tester, there is no expectation or acceptability that the temperature of the environment and the temperature of the piston-cylinder unit are within agreement. Both have individual correction factors.

The great thing about "proving" you don't need them or finding out you do- it could really lead you to some fixes/possibilities that will allow for reduced uncertainties/newer capabilities.

[briansalomon]

 33K5-4-84-1 will allow for a single point calibration under certain circumstances (step 3.6)