Thermometer Calibration

[bvy]

I have a Weston analog thermometer that was reading frozen distilled water at slightly under 31F. As I want to start doing C-41 processing, I decided this wasn't good enough. Last night I purchased a Taylor digital thermometer, ran the same test: 30.9F! Before I buy a third, probably more expensive, thermometer, can someone comment on this method of calibration? I was prepared to give or take half a degree, but I find it curious that both thermometers are this far off. Is there a physics lesson that I'm missing?

Thanks.

ETA: Both are stem/probe type thermometers.

 

[michaelbsc]

Assuming you have water without contaminants to affect the phase change temperatures (like sodium chloride), and also assuming you're at sea level (easily compensated by charts), then distiller water freezes at 32.000000000... F and boils at 212.0000000... F. Barometric pressure matters, but not enough to measure on your thermometer.

Getting a good slurry of crushed ice/water at equilibrium is important. You cannot have local hot spots.

These are know as physical constants, and are accepted in the metrology world as ultimate standards so long as you set up you baths correctly. In other words, they are not considered to be traceable to national standards, but are rather fixed to natural physical constants. Or, as one guy I used to work with said "traceable to God."

Now, do you have your bath at equilibrium? I don't know. Do you have a good slurry, or is your crushed ice in chunks that are too big? Technically it's supposed to be shaved ice, but for a home lab crushed is good enough.

Are you near sea level? Do you have the thermometers inserted into the fluid to the immersion line?

 

[Worker 11811]

Check the accuracy specs on that thermometer. I'll bet it is supposed to be accurate to ±1º C. with a resolution of 0.1º C.

That means the thermometer can be off by one degree Celsius in either direction. If you are trying to read 32º F. the thermometer can display a temperature between 30.2º F. and 33.8 F. and still be within spec. (Remember, virtually all electronic, digital thermometers actually read the temperature in Celsius then convert to Fahrenheit.)

With that in mind, your thermometer seems to be in spec. The temperature readings you are getting are within its range of accuracy.

This does not mean that your thermometer can't be dialed in better. It just means that you might need to adjust your expectations a little.

A good analogy would be like if I gave you a rifle and told you that it is capable of shooting ten bullets into a one inch circle at a distance of 100 yards. Then you shoot the gun and report back to me that the gun only hit a half-inch bullseye five times. Yes, if you aim very carefully, you are likely to hit a bullseye smaller than an inch but the gun was only designed to shoot into a one inch circle.

If you want really good accuracy you should probably invest in a laboratory grade thermometer but, if you are the tinkering kind there is a series of articles from the magazine, Scientific American, that give some good information on how to build and use a triple point cell to calibrate a thermometer. Do check them out!

http://www.scientificamerican.com/article.cfm?id=tackling-the-triple-point&sc=I100322

http://www.scientificamerican.com/article.cfm?id=calibrating-with-cold&sc=I100322

http://www.scientificamerican.com/article.cfm?id=a-homemade-high-precision&page=1

 

[BrianShaw]

it may not be the "gun's" fault, but variability in the ammo.

 

[E. von Hoegh]

it may not be the "gun's" fault, but variability in the ammo.

Or the shooter. Or wind.

The OP said he used ice. Ice can be any temperature up to 32f, and even if it reads OK at 32, that's no guarantee of linearity.

 

[BMbikerider]

To be perfectly honest if the thermometers are as close as that I would not worry about it. C41 has a reasonably wide latitude when it comes to processing it at home. I use a JOBO rotary processor and I am quite sure the temp does not stay at the desired 38 degrees C for the whole of the developing period. Even my two precious mercury thermometers which are guaranteed to be within +/-.2 of a degree C and likewise my electronic digital one when in a jar of water agree but outside in the same area on the bench there can be up to 1 degree difference. Just ignore the difference and get on with developing your films.

I am friends with the owner of a local mini-lab and he has told me in the past that his machine sometimes varies as much a 2 degrees all depending on the ambient temp outside.

 

[Peltigera]

Frozen distilled water? Frozen water (aka ice) can be any temperature below 0/32. You need to measure the temperature of the water as the ice melts.

 

[bvy]

Thanks for the responses. I did use a good slurry of ice and water. I thought altitude was negligible for the ice test (I'm about 750 feet above sea level). Anyway, no one seems terribly surprised by the 31F measurement, so maybe these thermometers are just fine, and it's my test that's imprecise.

Maybe I'll try the boiling water test. Maybe I'll just forget it and move on.

 

[anikin]

Yup. Just forget it and move one. It's not absolute precision that's critical, but consistency from run to run.

A man with one thermometer always knows the temperature, a man with two thermometers is never sure. Or something like that.

 

[michaelbsc]

Dead Link Removed

 

[markbarendt]

Yeah, move on.

Consistency is much more important than the exact temp.

With my thermometer I start at 100-102 degrees and it works just fine.

 

[Chan Tran]

C41 requires 100F right? Then I would just calibrate my thermometer at 100F.

 

[markbarendt]

C41 requires 100F right? Then I would just calibrate my thermometer at 100F.

How?