Measuring CRAZY cold temperatures with the prop
Philldapill
Posts: 1,283
Just curious about this... You can't really find a temp. sensor that measures wayyyy down in temperature, like say, -150C, right? Well, one that is fairly cheap at least. I've given this some thought, so here's my idea. This should probably be in the sandbox, but I may in fact need the propellers math abilities for this...
In a voltage divider, you have a positive input, a negative input, and a "divided" output. I think we all here have a general sense of how this works. Temperature is very similiar to voltage and current... What I have in mind, is making a "temperature divider". This would involve getting heavily insulated metal rod, with exposed ends, and a very small exposed area somewhere between. One end, would be exposed to the ambient air, and would have a temp sensor on it. The other end, would be connected to the VERY cold substance you are trying to measure. Somewhere between, which would be a very exact known spot, would have a second temperature sensor.
My thinking is, this would form a temperature gradient, and if you know precisely WHERE on that gradient you are measuring, and precisely WHAT the temperature is there and on the ambient end, you can calculate the COLD end fairly well.
Am I correct? If so, does this involve simple linear equations like a voltage divider, or would we need to use logarithms or any other functions that the Propeller is capable of?
In a voltage divider, you have a positive input, a negative input, and a "divided" output. I think we all here have a general sense of how this works. Temperature is very similiar to voltage and current... What I have in mind, is making a "temperature divider". This would involve getting heavily insulated metal rod, with exposed ends, and a very small exposed area somewhere between. One end, would be exposed to the ambient air, and would have a temp sensor on it. The other end, would be connected to the VERY cold substance you are trying to measure. Somewhere between, which would be a very exact known spot, would have a second temperature sensor.
My thinking is, this would form a temperature gradient, and if you know precisely WHERE on that gradient you are measuring, and precisely WHAT the temperature is there and on the ambient end, you can calculate the COLD end fairly well.
Am I correct? If so, does this involve simple linear equations like a voltage divider, or would we need to use logarithms or any other functions that the Propeller is capable of?
Comments
http://www.owlnet.rice.edu/~ceng402/ed1projects/proj01/markcov/proj31.html
http://www.springerlink.com/content/322153800v13g8k3/fulltext.pdf
http://www.engin.umich.edu/class/che360/projects/rod.html
I don't think it will be easy, although I don't have any experience in this area. All the graphs that I saw had lots of curves... The last link mentions Fourier series, so there goes your simplicity.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
News Flash!... Energizer Bunny arrested and charged with battery.
I'm not an expert, but I think the IR sensor would have to be at 0 K in order to make that determination.
But I'll second Shane's idea of using a thermocouple.
Philldapill's idea isn't out of this world, though. I just think it would be unnecessarily complex as heat transfer through a rod/forming a gradient could get fairly involved, for example you might possibly have to take into account convection currents, air flow around the rod, thermal conductance as a function of temperature, etc.
I can't seem to find were to buy one this good any more though [noparse]:([/noparse]
Post Edited (mctrivia) : 2/23/2009 7:40:17 AM GMT
For truly 'crazy' cold temperatures there are better thermocouples, such as the type T (based on copper/constantan) and the Au/Fe materials work well. But even better are the Silicon diodes, that show a strong change in voltage drop when a constant current is being pushed through them. They have slopes of 2 to 25 mV/Kelvin, so they are well-suited to this problem. Omega Engineering sells them (check model CY7), but they are ~$150.
Somebody recently posted the same behaviour of a n-type transistor ...
Cheers!
Paul Rowntree
A platinum thermistor can be an easy and affordable solution until liquid nitrogen (77K or -196°C) temperature, and maybe below. If you go lower you still have silicon diodes, and thermocouples type E or Chrome/Au-Fe, but they are more expensive.
Massimo