Electronics > Projects, Designs, and Technical Stuff

Suggestions for a Temperature Sensor

(1/13) > >>

amspire:
I am looking for suggestions for a temperature sensor.  Currently I only have some cheap thermocouple probes with a 1degC resolution and I need a lot more.

I haven't bought a temperature sensor for years, so ideas would be great.

The task is this. I want to test the stability of various resistors over time and temperature, and so I don't need absolute accuracy for the resistance measurement, but I want better than 1ppm relative accuracy. This is for building a Kelvin-Varley divider, and I also want to make a 9:1 and a 100:1 Hamon divider.

So I am going to take the readings on a 7 1/2 digit meter using a Leeds and Northrup standard resistor as the reference.



The L&N standard resistors are made from Manganin wire, so I first have to characterize it over temperature. With temperature compensation, these standard resistors are good as a reference for better than 0.5ppm for at least a week. Without compensation for temperature, the errors can be up to 50ppm. There is a standard formula for compensating, but I need to calculate two coefficients for the formula.

I tried with my thermocouples, but to get a curve accurate enough to calculate the characteristic curve of the resistor, I really need 0.01degC resolution and pretty good linearity.

Here is the curve I did measure with the thermocouple temp meter, and an Advantest 6871E-DC meter. It looks promising since it looks like the meter is capable of about 0.2ppm stability over the time I am measuring the standard resistor:



So what I am looking for is:

1 degC absolute accuracy to 35 degC if I calibrate for 0 degree point.
0.01 degC resolution and stability
Good linearity from 10 degC to 35 degC
Less then 7mm diameter, but the smaller the better.
Can be an IC, a probe, some kind of thermistor circuit. If it is a part like an IC, I will have to make my own sealed probe.

I am not after spending big $$$ for proper instruments. Anyone got some something that has worked for them?

Richard.

BravoV:
Though I can't give you any suggestion, just want to say what a nice resistor you have there !  ;D

Just curious, what is that hole in the middle for ?

IanB:
I think the kind of temperature sensor you want would be some kind of resistance temperature detector, consisting of a coil of wire (like Manganin perhaps?) which has a variation of resistance to temperature that is accurately known in advance. You measure the resistance and from that compute the temperature.

Joking aside, the reference standard would be a platinum wire sensor, since pure platinum has predictable properties obtained from standard reference tables. The question becomes, is the cost reasonable?  ;D (I suppose you might obtain some thin platinum wire from a jeweler's supply and construct your own thermometer...?)

IanB:

--- Quote from: BravoV on December 02, 2011, 06:43:14 am ---Just curious, what is that hole in the middle for ?

--- End quote ---

I'm guessing that's where you put the temperature sensor.

IanB:
A mercury in glass thermometer might be sufficient. Can you obtain one of those from a scientific supply house?

Navigation

[0] Message Index

[#] Next page