Calibrating digital thermometers?

[Muse]

My local utility had a campaign a few years ago to get mercury thermometers out of the citizenry -- they would give you a digital thermometer if you turned in your mercury. I bit.

I caught a flu around 6 weeks ago and used that thing. It's really a cheapie. It would take several minutes to reach the top temperature. Of course, when you are really miserable, it's a drag to have to leave the dang thing in your mouth for 3 minutes (I used a timer).

So, after my fever broke (a day or two), I went online and bought a digital thermometer off Amazon:

Clinical Digital Thermometer FDA Approved 10 sec

It's supposed to be good +/- 0.2 F. It's also supposed to reach top temperature within 10 seconds and beep when it's achieved it. I couldn't hear a thing. The seller sends me an email asking how I'm doing with it and I reply that it's not beeping. Seller sends me two other digital thermometers, each different from the one I ordered and different from each other. One's supposed to achieve top temperature in 1 second, the other in 20 seconds. Included a twice daily pill container to boot.

Segal's Law: A man with a watch knows what time it is. A man with two watches is never sure.

Well, I am not running a fever but I tested these thermometers against one another, making a chart.

The fancier one that's 1 second and has a CR2032 battery is way higher than the others ... by about 1.5 F.

The 20 second one averages ~0.2 F higher than the original 10 second one.

I can live without the beeps (none of them are much on beeping), but I'd like to know which of these is close to my actual temperature. Is there a practical way I can determine that? Guess I could work up a new chart including the cheap digital I got from the utility company, then make a guesstimate based on my data. Wish I'd never traded in my mercury thermometer!

 

[cbrunny]

I don't see why any of that means your old mercury thermometer was any more or less accurate than any of the thermometers you are currently testing.

0.2F is well within a reasonable margin of error. 1.5F is probably outside what I'd consider acceptable. Best way to know for sure is probably to get one of those infra-red thermometers or something from a reputable company.

 

[Red Squirrel]

Some may have a calibrate function, you'll need a calibrated temperature source though. A stable heat source + several other thermometers would probably be best bet.

 

[LevelSea]

https://www.nist.gov/calibrations/laboratory-and-industrial-grade-thermometers-calibrations#31010C

 

[Muse]

I don't see why any of that means your old mercury thermometer was any more or less accurate than any of the thermometers you are currently testing.

0.2F is well within a reasonable margin of error. 1.5F is probably outside what I'd consider acceptable. Best way to know for sure is probably to get one of those infra-red thermometers or something from a reputable company.

I'm thinking that maybe when I get my annual physical at my HMO (Kaiser Permanente), I can ask them to take my temperature and I can immediately use the 1 second and 10 second digital thermometers and write down the readings (actually, they remember the last reading). Can take my 20 second one too. This presumes that their thermometer is an acceptable standard. Presumably they use something better than what I have...