Can anyone give me a way to measure the temperature of a warm room
with common items, (i.e. a formula for to determine temperature by how
long it takes for a specified quanity of ice to melt), without a
thermometer?

Tisker --

You could conduct some experiments but you'll need to discount:
*  changes in humidity
*  air circulation (very important on the boundary layer temperatures of the ice)
*  ice density (chipped melts faster than block ice)
*  container (which may absorb heat)
*  presence of sunlight

There are some ways to get a precise room temperature with common
liquids, as proven with the Galileo thermometer:
http://www.howstuffworks.com/question663.htm

Best regards,

Omnivorous-GA

Ahh, definately, thanks for the input Omnivorous! 

If we're talking about ice, I'm looking for a crude formula that can
get me within 1-3 degrees.
Assumptions: 
-I can accurately determine the mass of a quanity of ice.
-I will put ice in a thin plastic cup with plastic lid, and place out
of direct sunlight, etc.
-I'm assuming constant room temperature.
-I'll try to use larger mass of ice to make more accurate, and would
like to have the formula based on how long the ice is exposed assuming
it started at 32 degrees, and that the end point of the time is when
no visible ice pieces remain in the cup.

Basically, I'm in an office which is boiling hot - assuming access to
normal office supplies (no thermometers, or looking at thermostats),
I'd like to be able to determine the approximate temperature. The
Galileo thermometer is good, but I just don't have the means to
construct something like that here.

Thanks a ton!
-Tisker

An intriguing question. Off the top of the head, a few notions.

TVs and computers are common household items and one could use the
parts to make a thermometer but I guess you want something a little
easier.

You could use the change in viscosity of water with temperature. As
you may be aware, hot water ?sounds? sloppier than cold water. That?s
because it is. Easiest way to measure the viscosity might be to
measure how long water took to drain from a vessel with a small
orifice. Note of course that expansion of the container hole with
temperature would have to be compensated for. Also the water should be
pure -- de-ionised water for irons. Couple of links regarding
viscosity with temp, but there are lots more if you can be bothered
searching.

http://scitation.aip.org/getabs/servlet/GetabsServlet?prog=normal&id=TEPHEX000048000004000514000001&idtype=cvips&gifs=yes

http://www.lsbu.ac.uk/water/anmlies.html

Utilising the viscosity directly, it is conceivable that you could
measure the time taken for ripples in a bath to die down from a
?calibrated? splash.

Another way would be to use the expansion of water with temperature.
Fill a vessel to an exact level and weigh it. Pyrex type glass would
largely obviate vessel expansion problems but you?d still have to be
very careful. To measure the weight, make up a long beam balance and
measure it off. Link below is a graph of expansion with temperature.

http://www.piercecollege.com/offices/weather/water.html

You could use the boiling point of water too -- with a syringe. If you
fill a syringe with water and a little air, then pull back the plunger
it?ll boil at a very low temperature. The degree of vacuum required
will vary with temperature so the how much you have to pull the
plunger back will be a measure of temperature. The air pressure will
also vary with temperature too and dissolved air will upset a simple
experiment but in principle it could be made to work. A better way
would be to have no air in the syringe, and water which has been
de-gassed (by boiling) then pulling the plunger with a spring. How far
the spring needed to be stretched to make the water boil would be an
indication of temperature. The main upset to this system would be the
friction of the plunger.

Totally different way. You could modify a mercury barometer to measure
temperature by converting the tube to measure expansion rather than
air pressure, but the owner might object to his expensive bit of kit
being hacked about.

Several other related techniques come to mind but hopefully that?s
enough to be going on with. A ?researcher? might come better systems.

In all these systems, calibration will be a problem but it could be
done reasonably accurately if you had enough data on the variables.

Best

Constant/uniform room temperature is unlikely with negligible air
circulation. A big cube of ice would cool the room significantly. The
ice would generate an air circulation in the room. The evaporation of
the melt water would cool the room significantly. Any people in the
room would warm the room significantly. opening and closing a door
could introduce a sizable error. Putting the ice in a thin plastic
closed container would slow the melting and produce 100% humidity in
the container. This would add several variables, reducing accuracy.
The weight to surface area ratio of the ice effects the melt rate. +
or minus 3 degrees f is likely possible without extreme measures. 
Neil

You could try using the contraction of something like copper wire. 
Measure a length of copper wire at room temp, put it in ice water (0
deg C), measure again and, knowing the coefficient of expansion of
copper (or whatever your wire was made of) you can work out room temp.
 Longer wires would be more accurate, but more difficult to measure
quickly. Presumeably you'd need to coil or fold the wire to get it in
the ice water container, so you'd need a way to uncoil / unfold it
quickly without heating it through friction.

Ian G.

Some really amazing ideas here - THANKS everyone!

Still looking for --- something simpler, I don't have copper here at
the office, and as you pointed out the dimensions are goign to be
restrictive.

The room is sufficently large that people and wind, etc just are not
going to influence the temperature to a great degree. On the same
train, I'm not going to use a block of ice so large that it would
adversely impact the room's temperature.

Would we be able to we derive a formula that yields temperature by
using something like... the heat of fusion of ice?

Catch a cricket

http://www.noblenet.org/reference/crickets.htm

Ok what if you just took your water sample and put it the freezer and
timed how long it took to freeze.

Now you can create a new temperature scale based on time taken for
your room temp water sample to freeze.

If you freeze enough samples you should get an better idea of the time taken.
Then you could go on the internet locate your city and compare your time with
known temps at the time.  After a few months your time based temp
scale should be more accurate.

What do you mean by "without a thermometer"?  Any instrument to
measure temperature is called a thermometer.

>Still looking for --- something simpler, I don't have copper here at
the office, and as you pointed out the dimensions are goign to be
restrictive.

Could you strip some cabling?  You could make the length more
manageable by putting 2 rows of tacks in a piece of wood, anchoring
the wire at one end and then snaking it round the tacks. If you mark
the free end before you freeze it you can then just mark the new
position after it's cooled and measure it at your leasure.

Ian G.

Went to bed last night and thought much easier way -- solubility. NaCl
(common salt) is unfortunately nearly flat but sugar has a very fast,
almost linear slope with temperature. You could have lots of
containers with different amounts of water and sugar, each
representing a different temperature. Have to ensure continual
jiggling for mixing but I?ll leave that to your imagination. Office
juniors?

Can?t think of better ways with ice than the other contributors suggestions.

Best

hi! 
if ur only lookin to find the temp of the room and meathod not very
imp try this may be sounds strange, look around at ur office-mates
wrist watches one of then is bound to have a digital one which also
give u the temp (or may be u can buy one a cheap one shouldn't cast
much!).

You're sitting in an office, you're asking questions on google - you
drink coffee. But I'm sure you have some collegue with a water boiler
for tea water, right? Perfect, thought so :-]

Take it, boil water, once it boils note the time (including seconds),
and put the boiler (lid closed) outside the window. Wait for about 5
minutes, open the window again, and (carfully!) drink a tiny sip. Does
it feel warm? Then wait a bit longer. Does it feel cold? Well, you're
too late; repeat, and check after 1 minute. The deal is: you need to
now the time when the water is at "mouth temperature", i.e. you have
it in your mouth and it doesn't really feel warm nor cold. This sounds
very much like guesstimating, but it works amazingly precisely for
temps around your mouth temp.

Okay, so now you have the time it takes the water to cool down to
mouth temp outside. Open your favourite weather web page, and check
for the temp at your current location; add two degrees centigrade if
you're on the ground floor, only add one up to the 3rd :-] You wanted
it up to a few degrees, you didn't specify whether it should be
centigrade or fahrenheit - so Celsius is my man :-]

Now you repeat the procedure in your office: boil water, note time,
wait until it has mouth temp, note time. Here's the formula that will
give you the temp in your office, with t_o (t_i) the time in seconds
it took the water to cool down outside (inside), and T_o the temp
outside:

q=((37°-T_o)/(100°-T_o))^(t_i/t_o)
T_i=(37°-100°*q)/(1-q)

You can replace 37° with 99F, and 100° with 212F, to get the values is
Fahrenheit. Let's look at an example:
outside temp: 10°, it took 5 mins for the water to cool down outside,
but 8mins inside.
q=((37°-10°)/(100°-10°))^(8/5)=0.146
T_i=(37°-100°*0.146)/(1-0.146)=26°
Let's see what happens if you measure things slightly wrong:
upwards: actual outside temp is 11°, you measure both times wrong by
12secs such that you get the maximal room temp (8.2/4.8) - you get 28°
downward: t_o: 9°, 7.8/5.2: 24°
The key here are the times - the more precisely you measure them, the
better the result - you should be able to get something +/-1°C

Let us know what the result is - you just gotta compare that with a
thermometer once :-]

What you'll need:

Water Bottle
Ruler
Drinking straw (If you don't have a straw a clear pen would work)
Pack of Bubble Gum
Water
Alcohol


Alcohol will probably be the hardest thing to find on this list, but
it's common enough to get. I'm sure you have some medicine pack
somewhere to snag a few ounces off of, or you can buy cheap probably
within a block from work.  I'm not sure, you could try it without
alcohol..

Mix the water and the alcohol 50/50, you should fill up a quarter to
an eigth of the bottle. Either mark the straw with a sharpie, or tape
a piece of paper with markings on it. I'm guessing every cm or mm
should be perfect.

Put the straw in the bottle half in half out, the bottom of the straw
should be under the surface of the fluid, but not resting on the
bottom.  Chew that bubble gum up well and seal the rest of the bottle
around the outside of the straw.

That's technically a working thermometer.  I would give you exact
measurements but it's going to vary depending on the girth of your
straw and several other things.

I'll check out some situations and see if it makes any sense to give
you exact numbers.  Let me know if you need better than this.

Throw some printer/pen ink in there and you'll have a nice dye to see
the water better.



This gives you a good idea:

http://www.energyquest.ca.gov/projects/thermometer.html

if you get a mesuring cylinder and plug it with something coated in
vaseline with a loose spring above then put a cap on the cylinder make
sure the plug creates an air tight seal. you can then use charles law:
the volume is proportional to the temperature when then pressure
remains constant, you should be able to use the markings on the
mesuring cylinder to give you an anwser for the volume, and as long as
when you pluged the mesuring cylinder you made sure the pressure was
the same as the surroundings the pressure should be about 1atm  =the
spring should keep the pressure constant. You will need a themometer
to calbrate the instrament and to check the pressure - or of course
you could make up your own scale.

also note that the triple point of water (where water is a solid,
liquid and vapour at the same time) is traditional used as a point of
reference you could just yous the melting point of water 273K though
becuase its very close.

If I understand previous posts correctly, water changes viscosity
(and/or density) with temperature.

If this is the case, so should the speed of sound through the water.
Thus if you hit a sealed vessel of water with something like a rubbery
hammer (a knuckle will do as well) it should, like a tuning-fork,
produce a sound which will change frequency according to temperature.

I wouldn't know what variables you would need to take care to fix or
know (pressure in the vessel might be one?)

While I don't have the formula for converting such frequencies to a
temperature, I'm sure it can be found. It is then another matter
entirely how you will go about measuring the frequency change?

But if you aren't completely tone-deaf, I know a trick used by
musicians when tuning their instruments: Two tones of near-similar
frequency will produce a clearly audible (if you know what to listen
for) "warbling" effect (due to interference). The frequency of this
percieved "warbling" sound will get slower as the two tones get more
similar (their true frequencies get closer).

Thus you should easily be able to find out if your "water-tuning-fork"
is at or near the frequency of a given (set of?) reference tones
(either made by a similar water-tuning-fork at a known temperature, or
by any other means of producing a sound of known frequency, such as
any taught string, piece of brittle material or even the dial-tone on
the phone).

Once you have done the math (if you need actual temp. readings) this
should provide for a VERY simple setup. I imagine you have one or two
of those water-tuning-forks of different size (which you can average
for improved accuracy).

For reference, I'd suggest either a series of strings, tuned to
different frequencies, or a number of wine glasses filled with water,
of a known temperature such as from a fridge, to different marked
levels.

Having a string or glass for each whole degree should be feasible,
considering the limited temperature interval you can expect to be
measuring. In fact, you should even be able to increase the precision
beyond this, by "guesstimating" the fractional part by ear (ie. by
listening to the difference in "warbling" speed for the two best
matching stings/glasses, the real temp. being nearest the slower one).

Regards,

inventus-ga.

P.S. Are you still conducting this "experiment" or have you long since
brought a real, vanilla, thermometer to your office? (I assume they
haven't got you permanently locked op in that hot place!)

... Thinking about it: Why bother with the water-tuning-fork thingie.
You could probably just as well use any "sound producing" (ie.
brittle) material. Something that expands a lot with temperature will
probably make things more precise.

As with most other suggestions, the trick is to ensure that your
"references" (whatever they be) does NOT change with temperature
(either because you've made sure that they aren't sensitive to temp.
changes, or because you make sure to fix their temperature before
measureing.)