T&L Publications

I intend to make a circuit to find the temperature of liquid nitrogen. I have posted the circuit here.<p>Can it work?[img][img]http://img244.echo.cx/img244/2672/sensor6vs.th.png[/img][/IMG]

I think I'd prefer to see a constant current source of 1mA feeding the diode rather than a simple current-limiting resistor and a voltage source. I've never researched extreme temperatures such as that and using a diode or transistor for measurement. All my Omega books are in storage. They have a wealth of information on temperature measurement and have all their materials for free and available on either CD or in paper. If you get all their books in paper, you'll take up at least 12 inches of shelf space. And if you like Dilbert, you'll find that they do also!<p>Dean

Depends on the accuracy you're looking for.
LN2 @ atm. pressure is -320F. Why would you want to measure that? Call me curious.

I am sort of a science hobbyist. Recently I took to experimenting superconductors. Now superconductors are usally dipped in liquid nitrogen. Thus the need to measure temperature. Apperciate if anyone can help!

An accuracy of +- 5 K would do nicely.

I'm no expert on this but I think you want a thermocouple. The sensor packaging is the big issue as extremely low temperatures tend to do nasty things to seemingly bullet proof materials. Different materials have different expansion coefficients and plasticity changes dramatically at low temps. I wonder how well that diode would survive a trip to Frozen Monkey Bay. Also, how are you going to calibrate it? I'm not saying it won't work, just that I have more unanswered questions than not. The TC approach is well documented with lots of suppliers.<p>You can get TCs that go down to -270C (-450F). You'll need a special amplifier/cold junction compensator - AD and maxim have them. You can usually get free samples, if not the chips are like 7-10 USD. The thermocouple can be had in the $20-30 USD range - probably cheaper but I haven't been exhaustive in my search. <p>Let us know what you do and how it turns out. It sounds like a lot of fun. I'd love to monkey around with superconductivity but around here that would have negative WAF with a very large coefficient.

No need to measure liquid nitrogen bubbling away at atmospheric pressure. If an open beaker (or dewar) is sitting on the bench boiling, it's temperature is -320F.
If you want to measure cryo temps anyway, most cryo thermometers use a diode with a 10 microamp constant current. Then, voltage is proportional to temperature. There is a chart at: http://www.2genterprises.com/cryo_manual_3.html
A caution- you can be severly "burned" with LN2. A burn could mean amputation. In a closed space, it can suffocate you. Study up on safety before handling!

When looking at that chart: http://www.2genterprises.com/cryo_manual_3.html
Keep in mind that diode forward voltage drop is a function of junction area, and varies a lot from unit to unit. Here is a scheme that eliminates that problem: http://www.analog.com/library/analogDia ... peratures/
Having said that, I am also concerned about package integrity. If the diode is not designed for cryogenic use, it probably won't survive. You might be able to buy one from 2G Enterprises, if you can afford it.

As jwax pointed out, the boiling point of liquid nitrogen is fairly constant and predictable. It is usually taken as being around 77 degrees kelvin. Corrections may be necessary if you are at high altitude, such as Denver, CO. Boiling liquid nitrogen is probably the best way of calibrating a temperature sensors in this region. A calibration at room temperature that is extrapolated to liquid nitrogen temperatures is unlikely to be very accurate no matter what technique is used.<p>Diode junctions are used fairly often as temperature sensors. They do require calibration at a fixed current. The ones used in infrared detectors were usually calibrated with 1 milliamp, and typically read about 1.06 volt at 77 degrees kelvin (if I remember correctly). Changes fron 77 degrees kelvin were usually about -2 millivolts/degree kelvin. Unless you need higher accuracy, a 3 1/2 digit DVM can serve as a readout device.<p>Thermocouples have a much lower output voltage and may be nonlinear, depending on the type. Thay will require much more amplification and signal conditioning, as well as requiring control of the materials and temperatures of all of the circuit.<p>Diode packaging might be a problem. I would try either glass encased diodes or transistor junctions in metal cans. These are made with materials with matched temperature coefficients, and with a little luck, the match might still be good enough at cryogenic temperatures. The temperature coefficient match for plastic packages probably isn't very good to begin with, and the plastic will probably be brittle at cryogenic temperatures.<p>[ June 21, 2005: Message edited by: stephen ]</p>

If I dip the diode in liquid nitrogen for 10 seconds, would it give me an accurate temperature reading without destroying the diode. I do not mind if the diode is damaged with just one reading. Silicon diodes are cheap. I can easily replace the diode for each reading. Any comments.

10 seconds may be long enough for a glass diode, but as noted, the diode junction temperature is what has to attain the temperature to be accurate.
Also, you may find some manufacturers of diodes work fine at cryo, some may fail. Good thing they're cheap!

If you dont mind a Lot of contamination in your soup?<p> The problem with these temperatures is expansion/shrinkage, and the fact that the different parts that make up the diode are made from layers and different materials means that as they shrink or expand, they will do so at different rates. <p>This can lead to flaking, cracking, or even small explosions where the pressure builds up and forces pieces to fly off. <p>Not that this is harmful, but that makes for a lot of contamination, some which might not even be visible?<p>[ June 21, 2005: Message edited by: Chris Smith ]</p>

The circuit with the silicon diode needs a constant current say 1mA at room temperature. The voltage across the diode increases with decreasing temperature. Does it mean that the constant current increases with decreasing temperature? As V=IR. Or is it necessarry to keep the current constant at 1mA with decreasing temperature? Any comments?

As I recall, transistors are negative coefficient, meaning that they become super conductors as the temperature drops. If you supply a constant voltage or current, the opposite will change, based on the fact that the transistor junction will drop in resistance. <p>For example a standard 250 miliwatt transistor conducting at your temperatures, will be able to conduct/ dissapate up to 250 watts, based upon its coefficient curve.

According to the chart, the voltage across the diode increases as it approaches zero degrees. This means the resistance of the diode is increasing as the temperature drops, with a constant current.
On another note, do you have a source for the liquid nitrogen? What sort of container will it be delivered in? You must know that it constantly boils off, depending on its packaging, quantity, etc., and it will be gone in a few weeks.