How long does it take to freeze water in the refrigerator?

        Did you ever wonder how long it takes to freeze a glass of water in your freezer?  Maybe not, but if so, wonder no more: the answer is at long last revealed!  For this experiment, I filled a small jelly jar with about 8 oz of water, suspended a temperature-measuring thermocouple in the middle of the jar, and put the jar in my refrigerator freezer. The thermocouple wire was run over the door seal and connected to a data logger outside the fridge. The logger measured the water temperature at one-minute intervals.

        As the graph below shows, the water temperature dropped steadily from the initial 77 oF to a degree or two above freezing, then drops more slowly to the freezing point. The elapsed time from insert to onset of freezing was about 1.5 hours. After reaching the freezing point, the water temperature remains nearly steady at the freezing temperature of water (32 oF) for about 5 hours as the water is converted slowly to ice.  The ice takes so long to freeze because a large amount of heat energy (called the heat of fusion) must be removed from water at 32 oF to convert it to ice at 32 oF . For example, one calorie of energy is given up by one gram of water in cooling by 1.8 oF . But 540 calories must go to turn that one gram of water into one gram of ice. As long as there is any appreciable amount of water in the jar the temperature of the ice-water mixture will be 32 oF . It is only after all water has frozen that the temperature in the jar can fall to the freezer temperature - this occurs about 3 hours after the water has completely frozen.

        The fluctuations in temperature that occur from about 4:00 to 15:00 are the result of the freezer cooling unit (the compressor) cycling on and off to maintain the selected freezer temperature.  The spike in temperature at about 7:30 is the result of my curiosity: I opened the freezer door to check on my experiment. Notice how long it takes the jar of ice to cool back down to the per-door-opening temperature - about 1.5 hours.  This is a good illustration of why not to open freezer door any more than necessary, unless you are feeling sorry for the power company and want to help them out by buying more of their product!

        I removed the jar of ice from the freezer at about 15:30, into a room with air temperature of about 77 oF . A rapid warm-up followed, with the freezing point being reached in less than 15 minutes. The temperature of the jar remained essentially at the freezing point for about 2.5 hours as the ice melted. Then the water warmed rapidly, reaching near-room temperature within about 2 hours.


        I expanded parts of the plot to examine in more detail the freezing and thawing events. In an ideal world, the water temperature would drop to 32 oF , and remain there until all water was frozen. In the real world, there are slight deviations from this ideal behaviour. The plot below (left) indicates that the water appears to cool slightly below 32 oF initially, and then warm back up to freezing. This cooling below the freezing point may be an example of supercooling, in which liquids can cool below their freezing point if small particles or roughness in the container surface are lacking.   The warm up (right) shows that same long period of time during which the temperature of the water-ice mixture is near freezing, with a gradual increase on temperature until all the ice has melted. Once the ice is all gone, the water temperature increases rapidly. The gradual warming of the ice-water mixture during the melting period is probably the result of the lack of stirring of the mixture. If the mixture had been adequately stirred, the ice-water mixture would have remained near 32 oF until all the ice was gone.

How hot does it get in the car?

        We've all heard the warnings:  Don't leave a child, pet, or disabled person in a car that is parked in the hot sun with the windows closed and no A/C.  It seems that every year brings a tragic news story about someone who didn't remember this warning.  And we've all been warned not to leave cameras, electronic gadgets, or other heat-sensitive stuff on the dash or the rear deck where it Is exposed to the full force of the sun.  But just how hot does it get in that closed-up car on a hot sunny day, anyhow?

A summer day inside of the Sub

         I wanted to know, so I set up a temperature logger in my Suburban to record air temperature at two-minute intervals during a hot sunny day in July. I recorded the temperatures at two places in the Sub, as shown in the picture on the right.  You can't see the actual sensors very well but the arrows indicate the approximate position.  The mirror sensor was hung from the mirror and was thus suspended a few inches above the dashboard.  That's a place where the sun blazes in with full fury, baking the dashboard - and anything on it. The other sensor, on the cargo space floor, stuck up about an inch above the floor at the gap between the folder-down seats. That location is generally shielded from the direct rays of the sun.

July 22, 2010 - The Sub, and location of above-dash (mirror) sensor and cargo space floor sensor

     The plot below (left) shows the temperature record, at two-minute intervals, for a 24-hour period on July 22, 2010.  The second plot (right) is just the same thing expanded to show the 9 am to 6 pm period in more detail. I should note that all times are EDT.  The weather was hot(!) and mostly sunny, with a few largely useless clouds passing over now and then after late morning.  The outside air temperature ranged from about 74 oF at sunrise to about 95 oF in late afternoon.  Just a hot dry boring-weather day in central Florida, typical of what we get when one of those pesky high-pressure systems decides to park its cozy little self over us for a few days. Me, I prefer the days with cumulus clouds, colliding sea-breeze fronts,  towering cumuli-nimbus, and late afternoon thunderstorms. Those are pretty to watch and sometimes bring a little excitement or even a tinge of alarm, especially if the lightning gets too close. But back to the main subject.

    So did it get hot in the Sub? Yea, it did - the Sub is parked facing east, and there are no nearby trees to provide morning shade at this time of year, thanks to Hurricane Charley. So the temperature began rising soon after sunrise as the sun's rays blasted in through the windshield.  By 9:30, the above-dash temperature was 123 oF and rising fast.  The cargo floor temperature was also rising, but not as fast and had only reached 90 oF by 9:30. I interrupted the process by driving to the "Y" , which is about 0.5 mile away, for my exercise session.

{Question: Why drive to the "Y" when you can get even more exercise by walking?  Answer: Sometimes I do walk, but then the walk home after the exercise can be too tiring.  Question: Well, why don't you just cut your workout at the "Y" a little to adjust for the walk to and from?  Answer:  I don't know.}

    During that short 5-minute journey the A/C pulled the above-dash temperature down to a comfortable 98oF , while the cargo floor temperature fell only about 1oF .  At the "Y" I parked facing north, so the dashboard was not exposed to the direct sunlight, and the above-dash temperature remained considerably lower than before.  Returning home, I once again parked facing east, and didn't head out again about 12:50pm.  During that time the above-dash temperature reached a maximum of nearly 150 oF, and the cargo floor temperature went up to 117 oF.  

July 22, 2010 - Above dash and cargo space floor temperatures

    In the afternoon we ran errands, stopping a couple of times, as you can tell by looking at the temperature fluctuations.  It's kind of interesting to note that 10 to 15 minutes of driving are needed to get the air well mixed in the Sub, as indicated by nearly equal above-dash and cargo-space-floor temperatures.  Another thing of interest is that the temperatures never got much below 100 oF while we were driving, yet it never felt hot with that cold air from the vents blowing directly at us.

    Finally, at about 3pm I parked the Sub for the day.  By then, the sun was in the west, and shining in the rear window with decreasing intensity as time passed.  The above-dash temperature reached only 123 oF at about 4pm, and then began a slow decline as the sun dropped lower in the sky and the house cast its shadow over the rear window.  By 5:40 pm the above-dash temperature dropped below the cargo floor temperature and declined to 84.5 oF at midnight.  This was nearly 3 oF lower than the cargo floor temperature. Its interesting to note that the late-afternoon above-dash temperature began declining a little after 4pm, while the cargo floor temperature continued to rise slowly for another hour.  This late rise in floor temperature is probably because the sun was still shining directly on the floor near the sensor, and the above-dash sensor was in shadow.

How hot does the dashboard surface get?  Could you fry an egg?

    After looking at these plots I wondered how hot the surface of the dashboard got. Certainly hotter than the air above it, as I could judge just by touch.  So a few days later (July 28, 2010) I made measurements above the dash, as before, and also at the surface of the dashboard.  Weather conditions were the same as before, and as before, the above-dash temperature exceeded 150 oF by noon, as shown in the plot below.  As expected, the dash surface was much hotter, approaching 180 oF.  However, these measurements may be biased high, because of direct sunlight heating of the thermocouple temperature sensor.

July 28, 2010 - Dash surface and air above dash

     Could you fry an egg on the dash? Here's a reference:   They say a temperature of at least 158 oF is needed. So,yes, probably I could fry the egg, but I'm not going to try.  At least not right on the top of the dashboard.  The interior of the Sub is enough of a mess without that.  But I might later try using a fry pan on the dashboard.  That wouldn't make too big a mess.


How accurate are temperatures measured in direct sunlight?

    Here are a couple of experiments comparing shielded thermocouples with bare thermocouples in direct sunlight. The object here  is to see if thermocouples exposed to direct sunlight might give inflated temperature readings because of direct heating of the thermocouple tip by the sun.  Sunlight heating should be highest when the thermocouple tip is oriented perpendicular to the direct sun rays, thus exposing a relatively large cross section to the direct sunlight.  And if the thermocouple was pointed directly at the sun so that only the tip received the direct sunlight, the direct sunlight heating should be at a minimum.  It's just like aiming a solar panel:  you get the most output when the panel is perpendicular to the sun's rays.  For these experiments, the thermocouples were oriented approximately perpendicular to the direct sun rays for maximum heating effect.

Measurement of air temperature in direct sunlight

     Two thermocouples were suspended side-by-side about 4 feet above the concrete driveway in full sunlight.  The tip of one thermocouple was shielded from direct sunlight by a hood made from aluminum foil, and the other thermocouple was not shielded in any way (left photo).  As you can see from the thermogram (right plot) the bare thermocouple (red line) definitely indicates a higher temperature at times, compared with the shielded one.  The difference can be as much as 3 or 4 F.  Note that there are times, however, when both thermocouples indicate little difference in temperatures.  These times correspond to the "valleys"  in the thermogram, and also times later than about 4:40 PM.  The "valleys" are the result of cloud shadows, which lowered temperatures. Also, the clouds in effect shielded both thermocouples from the sun, minimizing or stopping the direct sunlight heating of the bare thermocouple.  After about 4:40 PM, the thermocouples and the driveway beneath the thermocouples were in the shade cast by the shadow of the house.  This shading causing lower air temperatures and closer agreement between the two thermocouples because the bare thermocouple was no longer heated directly by direct sunlight.

July 29, 2010 - Air temperature above driveway: Bare and Shielded thermocouple

    Oh, and yes, I think the temperatures shown in the plot are correct, though they may seem high to you.  The air temperature in the shade (as NOAA measures it) was probably about 95 F during this exercise.  However, the driveway had been baking in the sun all day and the air over the driveway was considerably hotter than the air under a shade tree.  There was little or no breeze to mix the air over the driveway with cooler air from the adjacent lawn area, or the measured temperatures over the driveway would probably have been lower.

Measurement of dashboard surface temperature in direct sunlight

    Two thermocouples were placed side-by-side in contact with the dashboard in full sunlight.  The tip of one thermocouple was shielded from direct sunlight by a hood made from aluminum foil, and the other thermocouple was not shielded in any way (left photo).  As you can see from the thermogram (right plot) the bare thermocouple (red line) definitely indicates a higher temperature at times, compared with the shielded one.  The difference can be as much as 3 or 4 F.

July 29, 2010 - Dashboard surface temperature: Bare and Shielded thermocouple



Egg fry simulation

    I decided to try a simulated egg fry by placing a heavy iron fry pan on the dash, to see if it would heat up to the required egg-cook temperature (158 oF). I placed a temperature sensor under the pan, and touching the pan bottom.  I also placed another sensor in contact with the dash, just for comparison (see picture below, left).

July 31, 2010

The data plot (above, right) shows temperatures increasing as the morning progresses.  The dashboard temperature (red line) is generally several degrees hotter than the fry pan temperature.  That difference in dash and pan temperature may be due at least in part to direct heating of the dash thermocouple by sunlight.  The pan thermocouple is, of course, in the shadow of the pan.  The fluctuations and temporary cooling at times are due to shadows of passing clouds. 

    The fry pan temperature had increased above the fry threshold by a little after 11 AM, and continued to heat to a maximum of about 169 oF by 12:30 PM.  At that time, I dumped 2 ounces (the approximate weight of a large egg) of olive oil into the pan to simulate addition of an egg.  The oil was at room temperature (about 80 oF), and its addition caused the pan temperature to decrease to about 163 oF, still above the fry threshold.  So it seems that the heat capacity of the hot fry pan was enough to maintain cooking temperature after adding the relatively cool "egg".  The fry pan, with the 2 ounce load of oil,  re-heated to the pre-"egg"-addition temperature within about 15 minutes.

    So, if this had been an egg, would it have cooked? My guess is that it would have, but maybe not too quickly.  If you want a quick breakfast I don't think this is the way to get one.  You might wonder why I didn't just add an egg instead of the oil and answer the question once and for all.  The answer is that I didn't plan this study at all - just sort of muddled along trying one thing after another as different questions occurred to me.  And eggs are a little messier than oil, so I decided to see what happened with oil before trying egg.  Study plans like this one would not be apt to gain you much fame or fortune in the real world of science and engineering, but what the heck: I'm retired, and this is all just for fun!  But in a few days I intend to repeat the fry pan study - with a real egg.  I'm log the results here, maybe with pictures of me eating that dashboard-cooked egg.

The end of the story: Cooking the egg!

    Will an egg cook in the Sub or won't it?  All indications thus far are that it would cook, but the proof of the omelette is in the cooking. So I set up the fry pan on the dash as before, but this time I added a real egg once the pan was good and hot.  As you can see in the plot below, the pan temperature was about 164 oF at about 11:40 when I put in the egg.  The pan temperature dropped a little, but not to the cooking threshold of 158 oF.  That sharp dip in temperature at about 11:50 was caused by me leveling the pan a little, and dislodging the thermocouple in the process.  So the temperatures recorded after that slip probably do not represent actual pan temperature.

August 3, 2010 - Cooking the egg!

         I took pictures at irregular intervals to record the cooking process.  By 4 minutes the egg white was setting at the edges.

     After 10 minutes more than half of the egg white was cooked, and after 18 minutes the white was apparently all cooked.  The change in shape of the egg was the result of my leveling the pan in an effort to spread the egg surface more evenly over the pan bottom.

         After 23 minutes the egg still looked like it did after 18 minutes.  Finally, after 33 minutes I removed the pan from the Sub and took the final picture.  The white seemed firm, but the yoke seemed soft.  I should have flipped the egg over and got a pic of the other side, but I didn't.  Also, I should have pierced the yoke to check the consistency, but I didn't to that either.  By this time I needed to be somewhere so had to cut this experiment a little short.  Oh well, there's plenty of time left this summer to try this more thoroughly, and I probably will.  Then I'll post the results here.

     So can you cook an egg in a car parked in the summer sun?  I think this little experiment shows that you can, at least to some degree.  I've been using the term "fry" all along, but I wouldn't call this egg fried.  It seemed more like the poached eggs my mom used to make for me when I was a kid and recovering from those childhood illnesses.  Did I eat the egg?  No, I was in a hurry by this time, and anyhow, the darned thing just didn't look that good!

     Btw, I'm not the only one to try cooking an egg in the sun.  Just try a web search on the phrase "cooking an egg on the sidewalk" or some derivative of this and you'll come up with a ton of hits.  There are some funny video clips showing egg experiments.   Some are on the sidewalk, some on car hoods, and many others as well.


The winter of 2010 in Orlando: Was it really that cold?

     Right now, as I write this blog in October 2010, the winter of 2010 almost seems like a pleasant memory!  The summer of 2010 has certainly been hot, with day after day of above-normal temperatures.  But I can remember how, back in January and February, I was wondering if winter was ever going to end.  Usually in winter, cold fronts come through at about weekly intervals, dropping the temperature for two or three days, sometimes below freezing. Then comes a nice warmup into the '70's or higher.

    The first two months of 2010 seemed different: the cold fronts came, but the warmups didn't, at least not very soon or for very long.  Day after day passed with the highest temperatures in the 60's or below.  And the long-term outlook was never encourging, with talk about blocking highs and south-dipping jet streams that would just keep the cold air coming, until March some forecasters said.  I'm not going to get into the reasons for the cold weather here.  I'm more interested in comparing this last winter with historical temperature record to try to answer the question "was it really all that cold last winter?"

    The data presented here is from this NOAA link: , and go back to 1953 for Orlando.  That's 58 years of daily temperature.  There are longer periods of temperature available for Orlando, but not at this link. To get the weather data I discuss here, go to the NOAA site and select "Data Set/Product" under "Search Options".  Then, in the "Data Set/Product Options" box, scroll down until you see "Surface Data, Global Summary of the Day."  This is a free product, which is why I used it as my source of data!  Anyhow, select that product and then click the "Access Data/Products" button.  That takes you to the "terms of use"  declaration, which you should read and then click the "Accept" button.  Without going into a lot of detail, you can select the area of interest (state), then a station from a list box showing what stations have data and the period of record.  Then you will be able to select the begin and end dates, and an output format.  You'll probably want the "tabular data output" and "comma delimited" format, which will get you a file that will import directly into spreadsheets.  At the end of this selection process you will end up on a page showing the data file name (something like "CDOxxxxx.txt", and a document describing the format of the data in your file.  Just click on the file name to view the data, then use the file menu on your browser to save it to your computer.  You might want to change the "txt" extension to "csv", so you can open the file directly with most spreadsheet programs.

     Before going further I need to mention data differences that may exist between two different data sources.  I noticed this because I still remember that awful day, Jan 9, 2010, which started with pre-dawn sleet in some locations,  nothing but clouds all day, and a maximum daytime temperature of about 40 F.  Thats cold for Orlando!  However, the GSOD data for this day listed a maximum temperature of 45 F at Orlando International, much warmer than I think it was anwhere around here that day.  Indeed, if you look at monthly summary data, available on the Melbourne NWS site (, you will find that the daily high for Jan 9, 2010 was 40 F at Orlando International, and 42 F at Orlando Executive airport, near Winter Park. So why the difference in reported daily max?  I'm not positive about this, but I think it has something to do with the way the daily summaries are generated.  Looking back to the GSOD data I noticed the daily max and min for Junuary 9, 2010 are marked with an '*', which is defined in the data explanation as indicating that these were determined from hourly observations.  I don't know what the normal procedure for picking daily max  and min at this site is, but I persume that it might be from more frequent observations, which would be more apt to hit the true max T and min T.   So the temperatures discussed here (from the free data set) may be different from temperatures given in data sets that must be purchased and presumably may be more accurate.  Ok, so with data qualifications stated, I will now summarize what the data indicate about the winter of 2010. The data plots below may look kind of busy and confusing, but can be summarized fairly simply.  

    I'll start first with the monthly mean temperature by years plot, on the left below.  This plot shows that the mean temperature for both January 2010 and February 2101 was about 54.2 F. January and February means have been nearly that cold or even colder during several years since 1953, so there's nothing too unusual there.  What is a little unusual is that January and February 2010 mean temperatures both were cold.  Not since 1958 have both of these months had a mean temperature of less than 55 F.  However, I should mention that there have been several years (from 1977 - 1980, for example) when both January and February were both less than 60 F, which is still kind of cold.

    There is nothing at all remarkable about January and February 2010 in terms of the minumum temperatures (right plot below).  The January 2010 minumum was 27 F, and February 2010 never dropped below 34 F.  Many other years had lower temperatures in both January & February:  to be specific, there were 10 years between 1952-2009 that had the same or lower minimum January temperatures, and 24 years that had the same or lower minimum February temperatures.

    A look at number of daily maximum temperatures (Tmax) that were less than 70 F and less than 60 F each year (shown on the left plot below) indicates that January - February 2010 had a relatively large number of cold days.  With 32 days during January - February 2010 not warming into the 70's, the year 2010 is among the 4 coolest