Thermometers Driving Me Insane!!

It sounds like a simple calibration problem. Too bad the user is not allowed to calibrate most of them. I don't think that mechanical spring thermometers are very accurate.

Watches are particularly odd. They expect to be mounted on your wrist, so they expect a warm body surface on the backside. One would hope that the thermistor/transducer is on the top.

—B.G.—

Hi Bob:

Yeah, too bad one can calibrate altimeters and compasses but not thermometers.

Not 100% correct, but part of me would feel somewhat more confident if the Alterra read somewhere in the middle rather than at the extreme.

Also, there's a part of me that thinks the el cheapo mercury ones might actually be more accurate than the fancier and more expensive digitals. Digitals also require a lot more time to "settle down" on a reading — 10-15 minutes per HighGear instructions.

Benjamin,
I have a Highgear watch but can't remember the model. My thermostat in my home is at 76F. The Highgear reads 74.8, my new Timex Expedition WS4 reads 76.3. I chose the carabiner style on the Timex to keep it on my pack away from my body, because I found if I wore the device on my wrist, my sweat would interfere with the sensor port & give wacky altitude, barometer, & temp readings.

Methinks ±1 or 2° is pretty acceptable for gadgets. But 6° is a bit much. Sigh…

If there's a truly accurate thermometer outside of a physics/chemistry lab, I haven't found it yet! That's why I haven't bothered to look for one of those fancy watches–I wouldn't want to test its accuracy by putting it in boiling water!

Well, there are some fairly accurate thermometers outside of a physics/chemistry lab, but that doesn't make them cheap or UL.

Most of the modern temperature sensors on modern consumer products are purely solid-state, and most rely on circulating air around the chip. If they are small, then they tend to react quicker. If they are large and mounted on a heat sink, they will be very slow to react.

You can start to calibrate some sensors by dipping them into a 50-50 mixture of water and ice chips at sea level, and then dipping them into boiling water at sea level. That is not totally accurate, but it will get you into the ballpark. Most semiconductor sensors are non-linear, meaning that they might be accurate for the first two points mentioned, but they might be way off in the middle. That is why most such sensors have linearity correction built in with a software "look-up" table. Then they can be amazingly accurate.

–B.G.–

HighGear Alterra watch going back to Campsavers on Monday. Not a bad watch at all and quite intuitive actually, but just too many features that I don't need.

I'll stick with the simple. Mercury never lies, right? :)

Good for you. I don't trust most digital watches. I use a little compass and thermometer -plastic that fit over a watchband . Bought both in Beijing. They work. I think Brunton? makes a compass thermometer combo which is very small and light.

B.G. wrote: "You can start to calibrate some sensors by dipping them into a 50-50 mixture of water and ice chips at sea level, and then dipping them into boiling water at sea level."
It doesn't have to be a 50-50 mixture of water and ice, it has to be a water-ice mixture that's in equilibrium. What matters is the equilibrium, not the proportions. Likewise with the boiling water: a pot of boiling water will not actually be at 100 C unless it's in equilibrium with its own vapor. If he just does this casually in the kitchen, he'll have no way of ensuring equilibrium, and he won't be likely to get any better accuracy than the accuracy he's already achieving.

Benjamin…for backpacking purposes, does this matter?

A man with one watch always knows the time
A man that has two , he is not so sure.

My Casio and my cheapo kitchen spirit thermometer agree most of the time.
Note that according to the barometric pressure, if I stay inside I will be dry.
Franco
The first pi was not that clear…
neither is this , but the spirit is on 20c.

I bought one like the one shown in the link:

http://www.novatech-usa.com/Products/Min-Max-Thermometers/4050

Note the accuracy listed for a given range of temps. This type of thermometer is sold for food service, refrigeration, etc. They are small and light and have a high/low feature that hikers may like for determining lowest nighttime temps.

Many years ago, when I was doing photographic work, I had a Kodak Process thermometer that used mercury, was about a foot long and was accurate to about 1/8 degree Fahrenheit. IIRC, it cost about $100 in the mid-1970’s.

In photography school we had 25 people in the program and were able to compare a couple dozen thermometers at the same time. Conclusion: cheap thermometers vary a lot.

All said and done, thermometers tell you what you already know and are probably most helpful when it is REALLY cold. Other than that, it is more a matter of curiosity and personal record-keeping. If you are doing gear testing, I think humidity should be accurately recorded too.

To record hi/lo temp and humidity, an inexpensive electronic humidor thermometer works well:
link to amazon

Tohru, how much does that thermometer weigh?

I don't have that exact model, but one that is very similar. Mine weighs less than an ounce..

Great This may find a place in my pack. I guess the only downside is the temp range stopping at 32f

I had a friend with an ABD PhD — meaning "All but Dissertation" – the classes but not the independent research and writing.

I called him Aristotelian because he was one who preferred to believe that the only way to really learn things was by thinking – no dirty little experimentation need be tolerated.

Pseudoscientific was added because in nearly 10 years of education, the only science class taken was meteorology which was taught with only high school algebra as a needed mathematical background. So, his lab experience was pretty much zero as well.

The fact that two banks a few blocks apart could report temperatures 10*F different was difficult for him to accept. His major experience with temperature was the furnace controller on the wall and the outdoor temperature sensor in his Subaru. His attempts to learn to cook were all dismal failures.

One day he bought a thermometer, probably electronic. I never saw it. It was a hot summer day. The temperature outside the car was consistent with the temperature on the radio.

He left the car sitting in the sun and came back an hour later. He started measuring the temperature at different places in the car and was astonished that the temperature varied widely all over the place. Up to 130*F in places and down to only 90*F in other places.

He had no idea that temperature was so hard to "measure".

He could never understand why I would gather the indoor and outdoor electronic sensors together in one spot in the house and let them all come to stable temperatures and record the results on each of the units.

As a rule, all of the temperatures of the 3 outdoor sensors and the 5 indoor sensor units were within +/-1*F, with most of the differences being only a few tenths of a degree. Everything was a LaCrosse branded product.

It was too hard to read the cheap glass thermometers to that accuracy, so I simply ignored them and still do. There is always a digital around to glance at.

I have a HighTech "wonder watch/altimeter/barometer/compass/stopwatch…. and usually never remember to even glance at the temperature indicated upon it because it is such a pain to get it to reach a stable off the wrist temperature. If I ever did pay attention to it, it would be treated as a not very reliable instrument.

However, when I have worn it in the hot tub at the YMCA, it does appear to indicate the water temperature pretty accurately – when there is someone around to measure the hot tub temperature, that is.

The beginning post of this thread tells me that somebody suddenly got an experimental exposure that was shocking to inner beliefs about temperature measurement being easy.

Mixing up different thermometer types and brands. Almost as bad as old style sleeping bag temperature ratings.

Actually, Rocketman…

I placed all six thermometers close together — in a room away from any sunlight and with the windows closed — and waited 15 minutes before taking the readings.

"Mixing up different thermometer types and brands. Almost as bad as old style sleeping bag temperature ratings."
There's nothing wrong with expecting different types and brands of thermometers to read the same when they're exposed to the same air, sitting on the top of the same table. The differences among them do indicate systematic and/or random errors. Not sure what the point of the long story was supposed to be.

"Not sure what the point of the long story was supposed to be."

Same here…
Franco

I gotta say it: Just don't carry one!

I used to take one along just for the grins of it, then I realized that knowing the temperature doesn't affect my trail life in any way. If I'm cold I add layers, if I'm warm I remove layers.

I guess if you were trying to very specifically nail down your system, or were doing gear testing and wanted to be able to tell people that your sleeping bag was good to 34.6 degrees, well, ok… But pretty much, for personal use I'll know that a given system will work for me in a given season. I'm not going to purhase/own/take a different sleeping bag for a 10*F differential.

I dunno… not trying to be snarky, just sayin'!

"The beginning post of this thread tells me that somebody suddenly got an experimental exposure that was shocking to inner beliefs about temperature measurement being easy."

Did you actually READ Ben's post?