Post by thread:Q: How to get accurate temp. calibration reference

On Sat, 2 Jan 1999 20:58:31 +0000, you wrote: >Hello, > > I would like to know how you obtain a reasonably accessible temperature >calibration reference for thermistor/RTD/sensor calibration in the 0..100 >degrees C ? Is there a possibility to obtain such a service at the local >University/whatnot at a LOW price ? > >tia, > > Peter 0 degrees is easy - mixture of (pure) ice & water. You may be able to do something similar with a mix of water and steam, but this is probably harder. Maybe a mix of some other substance with a well known melting point in molten and solid states.

At 08:58 PM 1/2/99 +0000, Peter L. Peres wrote: >Hello, > > I would like to know how you obtain a reasonably accessible temperature >calibration reference for thermistor/RTD/sensor calibration in the 0..100 >degrees C ? Is there a possibility to obtain such a service at the local >University/whatnot at a LOW price ? > >tia, > > Peter Just the old method of a bucket of ice water, and a pot of boiling water. As long as you're at sea level, it's pretty darned accurate. I have a surplus of the ice-water slurry at the moment :)

On Sat, 2 Jan 1999 21:31:28 -0500 dave vanhorn <spam_OUTdvanhornTakeThisOuTCEDAR.NET> writes: >Just the old method of a bucket of ice water, and a pot of boiling >water. The ice water should be well stirred. Since all the ice floats, the water near the bottom of the container can warm up a little. Boiling water tends to stir itself. Use distilled water in both cases. >As long as you're at sea level, it's pretty darned accurate. With the boiling water, if tenths of a degree are important, you should correct for the local atmospheric pressure. At the time of the experiment (since it varies with weather conditions), ask the nearest airport or weather station. I think the CRC chemistry handbooks have a table of the exact boiling point at different pressures. ___________________________________________________________________ You don't need to buy Internet access to use free Internet e-mail. Get completely free e-mail from Juno at http://www.juno.com/getjuno.html or call Juno at (800) 654-JUNO [654-5866]

Hello, On Sat, 2 Jan 1999, Mike Harrison wrote: > 0 degrees is easy - mixture of (pure) ice & water. > You may be able to do something similar with a mix of water and steam, > but this is probably harder. Maybe a mix of some other substance with > a well known melting point in molten and solid states. afaik neither 0 degrees nor 100 degrees can be attained accurately without replacing H2O with something else. Humans reading thermometers require some time for this, plus it is better if you can juggle the temperature higher and lower a little bit to accomodate impurities and air pressure. Saline (NaCl) solution is ideal as it boils higher and freezes lower than clean water. It also corrodes almost anything except glass. Of course you still need a reference thermometer, or 10. (I have that). The idea is to go slightly beyond the target temperature and then allow to slowly cool/heat through the target point and get an accurate reading. But this is not what I had asked, I had asked how to do it better than this. Where does one get hold of instruments calibrated to SI units/standards on loan, for love, or where does one steal them from ? (just kidding). Because I am also interested in a pressure standard, a humidity standard, etc. Peter

> afaik neither 0 degrees nor 100 degrees can be attained accurately >without replacing H2O with something else. Humans reading thermometers >require some time for this, plus it is better if you can juggle the >temperature higher and lower a little bit to accomodate impurities and air >pressure. Saline (NaCl) solution is ideal as it boils higher and freezes >lower than clean water. It also corrodes almost anything except glass. Of >course you still need a reference thermometer, or 10. (I have that). The ice bath and boiling water bath are how these units were originally derived. The phase change points of water at sea level. (Distiled, naturally) The saline approach was used by Farenheit for his scale (the coldest solution of salt water and ice) and his 100 degree mark was the temperature of the human body. Neither of these phenomena are terribly stable.

Dave VanHorn wrote: > The saline approach was used by Farenheit for his scale (the coldest > solution of salt water and ice) and his 100 degree mark was the > temperature of the human body. Strictly, it was not a human body; I believe it was a *Cow*. Which explains why 100 degrees F represents illness, not normality. -- Cheers, Paul B.

At 08:58 PM 2/01/1999 +0000, you wrote: >Hello, > > I would like to know how you obtain a reasonably accessible temperature >calibration reference for thermistor/RTD/sensor calibration in the 0..100 >degrees C ? Is there a possibility to obtain such a service at the local >University/whatnot at a LOW price ? > >tia, > > Peter For 0 to +100 deg.C you can use 4 calibration points, resonably spaced, and intermediate points can the be calculated through the Steinhart & Hart equation. If your 4 points is known with accuracy better than +-0.06 deg.C and you can measure resistance better than 0.2/0.3%, the end result should be better than +-0.1 deg.C /0+100 deg.C (If your thermistor has a resonably curve form.) In most practical calibrations, reference thermometers and not reference points are used. All you need is a thermometer with 0.05 deg.C divisions and - unfortunately - a certificate so you know the error at or close to your selected points. Your local university has heaps of thermometers - try to ask them to borrow one to adjust your own quality thermometer after. When doing the calibration, be careful to stir properly in a good sized thermojar and to take the different timeconstants into consideration. It is not hard to do - temperature is ssllooww. But - it may be simpler and more economical to buy a thermistor with guaranteed accuracy - e.g., from Fenwall. Best luck, Hans Blichfeldt

Mike Keitz <.....mkeitzKILLspam@spam@JUNO.COM> writes SNIP >With the boiling water, if tenths of a degree are important, you should >correct for the local atmospheric pressure. At the time of the >experiment (since it varies with weather conditions), ask the nearest >airport or weather station. I think the CRC chemistry handbooks have a >table of the exact boiling point at different pressures. I have often seen this method quoted. However, the pressure quoted at the airport is the equivalent pressure AT SEA LEVEL , and not the pressure at the surface (think about why the correction is needed). The correction is about 1 inch Hg/thousand feet at sea level, and changes with altitude. Unfortunately, for my metric friends, inches and feet are still used in international aviation, since so many A/C have US instrumentation. Mike Mullen

Mike Keitz wrote: > > The ice water should be well stirred. Since all the ice floats, the > water near the bottom of the container can warm up a little. Boiling > water tends to stir itself. Use distilled water in both cases. > > >As long as you're at sea level, it's pretty darned accurate. There's a good chance that Peter is pretty close to sea level. BTW, where in Israel are you, Peter (I lived there for a year). > > With the boiling water, if tenths of a degree are important, you should > correct for the local atmospheric pressure. At the time of the > experiment (since it varies with weather conditions), ask the nearest > airport or weather station. I think the CRC chemistry handbooks have a > table of the exact boiling point at different pressures. > If hundredths of a degree are important, and IIRC, the temperature of the ice/water mix is 0.01 degrees. Also, the airports (and TV weather) adjust the atmospheric pressure to sea level. In Calgary, for example, the pressure shown on TV is always around 100kPa when in actuality it's closer to 90. The airport does have a 1-900 ($$) number that can be called to get the *real* pressure, though. --Matt

On Mon, 4 Jan 1999, Gerhard Fiedler wrote: > At 08:26 01/04/99 +1000, Paul B. Webster VK2BZC wrote: > > Strictly, it was not a human body; I believe it was a *Cow*. > > which may explain why i'm still having trouble counting in fahrenheit (i > got used to inches, feet and gallons in the meanwhile ;-) Well, I never knew cows are *hot*. They certainly don't *look* it ;) Anyway, I was suggesting salt solution for the purpose of calibrating on the Celsius scale, the salt solution refusing to boil at 100C and to freeze at 0C, it makes it possible to stir the solution / impossible for it to boil into your face etc. You have to try to stir ice slush in a Dewar with an expensive glass lab thermometer in it, and get a few drops of really hot boiling water on your face, before you get to fully appreciate the qualities of salt solution. Thank you for suggesting the University, but I'll try the local Standards Institute first, they may point me in the right direction (after a phonecall or 12 or so). thank you all for your suggestions, Peter

> Anyway, I was suggesting salt solution for the purpose of calibrating on >the Celsius scale, the salt solution refusing to boil at 100C and to >freeze at 0C, it makes it possible to stir the solution / impossible for >it to boil into your face etc. You have to try to stir ice slush in a >Dewar with an expensive glass lab thermometer in it, and get a few drops >of really hot boiling water on your face, before you get to fully >appreciate the qualities of salt solution. The whole point of boiling and freezing is that the phase change of pure water is very predictable. Making any haphazard solution destroys that.

Hello, On Mon, 4 Jan 1999, Matt Bonner wrote: > There's a good chance that Peter is pretty close to sea level. BTW, > where in Israel are you, Peter (I lived there for a year). I'm in Tel Aviv, ~300 meters from the beach (I did not measure, but I can see blue at the end of the road). Anyway the road rises a bit so maybe it's +10-20 meters ASL ;). As to the weather pressure report here: a) I can throw dice better than them b) they don't say it, in the first place ;( Peter

Hello, On Tue, 5 Jan 1999, dave vanhorn wrote: > The whole point of boiling and freezing is that the phase change of pure > water is very predictable. Making any haphazard solution destroys that. Yes, but sticking a reference thermometer into it makes me feel much better (and I did that). FYI there is a goodly gradient in a boiling vessel as small as 200cc (over one degree C - can be as much as 5 or so). I could have used oil but I prefer salt water (washing oily things reminds me of kitchen work - yecch !). The curves I got (datum every 5 degrees C) are a near perfect as far as I can tell with my plotting program (log scale for R on Y, linear T on X, gives a ruler straight line through the datums as expected). The errors are within my needs for this method, but it took almost 2 hours to make the curve by hand, and check it a few times. There are (FLUKE) DVMs with a thermocouple adapter plug. Has anyone got any experience with these in this domain (use as refrence during gang calibration of devices for example) ? I know that the FLUKE that has this adapter also has a serial connector, so I can make some computer look after the heater/cooler. tia, Peter

>There are (FLUKE) DVMs with a thermocouple adapter plug. Has anyone got >any experience with these in this domain (use as refrence during gang >calibration of devices for example) ? I know that the FLUKE that has this >adapter also has a serial connector, so I can make some computer look >after the heater/cooler. Type J or K thermocouples, very accurate. The meters are compensated for their curve.

At 15:24 01/05/99 -0500, dave vanhorn wrote: >>There are (FLUKE) DVMs with a thermocouple adapter plug. Has anyone got >>any experience with these in this domain (use as refrence during gang >>calibration of devices for example) ? I know that the FLUKE that has this >>adapter also has a serial connector, so I can make some computer look >>after the heater/cooler. > >Type J or K thermocouples, very accurate. The meters are compensated for >their curve. for this range don't use thermocouples, they're probably less accurate than you'd like it (there may be a difference of a couple degrees between different thermocouples of the same type, according to the standard), and their intended range is much bigger. Pt100 resistors are better suited for your range; their accuracy is better and the variation between different sensors is much smaller. ge

At 09:35 PM 5/01/1999 +0000, you wrote: >Hello, > > >There are (FLUKE) DVMs with a thermocouple adapter plug. Has anyone got >any experience with these in this domain (use as refrence during gang >calibration of devices for example) ? I know that the FLUKE that has this >adapter also has a serial connector, so I can make some computer look >after the heater/cooler. > >tia, > > Peter > You cannot expect any calibration to be better than the accuracy of your reference equipment. If you want to use a thermocouple for low temperature calibration to an accuracy of less than +-2 deg C, you must know what you are doing - you will need extremely expensive equipment and a full sized lab. Try e.g., to check the accuracy of the thermocouple itself - Best regards, Hans