Suggestions for a Temperature Sensor

[BravoV]

Thanks for the great photos of that resistor's internal, really learned a lot here.

Have been watching closely this thread, and your other metrology project threads too.

[IanB]

I just wanted to make sure it was still fluid enough to circulate heat at all practical temperatures for the resistor, and it looks like it does. I am sure it wouldn't cause any damage if it did solidify.

Before it solidifies it will reach the pour point, which is the temperature below which it becomes so thick it can no longer be stirred or poured. The pour point will be much higher than the solidification point. If you are cooling down a bottle of oil you can observe the behaviour by tipping the bottle on its side. Below a certain temperature you will notice the liquid becomes syrupy and stops flowing.

[amspire]

If you are cooling down a bottle of oil you can observe the behaviour by tipping the bottle on its side. Below a certain temperature you will notice the liquid becomes syrupy and stops flowing.

You are absolutely right. I got the oil down to -11 C and it was getting towards the thickness of honey - not quite there. I thought I might get components solidifying as out but it didn't happen.

[amspire]

Thanks for the great photos of that resistor's internal, really learned a lot here.

I wasn't sure if anyone would be interested in the photo's. I am waiting on some stuff to come from China so i can start some proper testing, but in the mean time I have been learning heaps too thanks to Conrad, robrenz, HAL-42B, alm and others.  Nothing beats experience.

I will have to one day get my hands on some Evanohm wire. It is really amazing stuff, if you can get the top grade wire at a good price. So far, I have seen prices like $50 for 50 feet - too much!

[Conrad Hoffman]

I've gotten wire from the H.P. Reid Company, now a division of Kanthal, in Palm Coast, Florida USA. I bought several spools of manganin, plus one of Rediohm 800. See if you can still get Rediohm 800, as it's probably as good as anything out there. It was some years ago, but I remember buying many hundreds of feet of wire for about $200. One trick was that I asked them to quote me whatever short spools were on the shelf, rather than making them wind specific lengths of specific gages.

http://www.kanthal.com/products/material-datasheets/wire/resistance-heating-wire-and-resistance-wire/redi-ohm-800/

[robrenz]

Shows evanohm in the listing, reasonable shipping and price
http://www.ebay.com/itm/Fine-Nichrome-Wire-YOUR-CHOICE-28-46ga-Lot-30-/310363518424?pt=LH_DefaultDomain_0&hash=item48431b75d8

Another
http://www.ebay.com/itm/PRECISION-RESISTOR-WIRE-MANGANIN-176g-NICHROME-80g-/300623716451?pt=LH_DefaultDomain_0&hash=item45fe91e063

I have bought from this vendor several times with good success.  You shoud ask them for what they may have that is not listed.
http://stores.ebay.com/Wires-Store?_trksid=p4340.l2563

[amspire]

I finally received the two Platinum Resistance Thermometers I had ordered:

http://www.ebay.com.au/itm/400244691889

$8 each including postage. They claimed to be Grade A. They have a 1000 ohm resistance at 0 °C and 1384 ohms at 100 °C with a linear slope in between. The curve of platinum is not exactly linear but close - the exact formula is at the end of this post. The sensor is 30mm long and 5mm diameter. Sealed with epoxy, so not useful for extremely high temperatures. Should be OK up to 200 °C at least, but I am mainly interested in under 100 °C temperatures. Platinum sensors are usually available at 100 and 1000 ohms, and the 1000 ohms are a bit easier to use with 2 wire resistance measurements.



So how good are they?

The definition of Grade A  is

Tolerance = ±[0.13 +0.0017 *|t|] °C .   (With t measured in Centigrade)

At room temperature, the two matched resistance to 0.02%. That corresponds to a temperature difference reading between the sensors of 0.05 °C. Not bad.

I then checked the 0 Deg °C. The water temperature of a glass full of chilled water and ice is a very accurate 0 °C and it is not affected by factors such as air pressure.  I was using rain water (non urban - almost no dissolved minerals or pollutants) so the water temperature between the ice should be 0 °C to an accuracy better then 0.01 °C.

Both sensors measured -0.02 °C. Well within the Grade A spec of +/- 0.13 °C.

Boiling point is very hard to check accurately, so I will not bother checking that. For one thing, I am at 1100 meters altitude, so our boiling point is well below 100 °C anyway.

I have to say I am extremely impressed. Looks like they are very comfortably within the Grade A specs, so for $8 a real bargain. They are perfect for use with my standard resistors, and I have also received my metal film resistors that I was wanting to do temperature and stability tests with. At 100mV test voltage, the self-heating is not a problem (much less then 1 °C error) , but if I want to be paranoid, I can test at 10mV which is 0.1uW of heating.

Richard

Edit:  If you want the true Temp Curve:

ASTM Standards E 1137 for Industrial Platinum Resistance Thermometers specifies that the resistance-temperature relationship for such devices for the range 0 °C to 650°C, to within the tolerance mentioned above, will be described by the equation:

R(t) = R(0)[1 + At +Bt²]

Where:
t = temperature (to ITS-90), °C,
R(t) = resistance at temperature t,  (= 1000 ohms for my sensors)
R(0) = resistance at 0°C
A = 3.9083 * 10^-3(°C), and,
B = -5.775 * 10^-7(°C²).

The European DIN standard that has now become IEC Standard 60751 uses slightly different numbers, but the end result is basically the same. For a fourth order accurate curve, see this document:

http://www.uniteksys.com/Graphics/CalVan.pdf

[Conrad Hoffman]

Nice!

There was an article on how to build a triple point cell in Scientific American magazine some years back. You can probably find the article on-line. It wasn't terribly difficult and gives you an excellent temperature reference. The high end is more of a problem because there isn't much that has a handy melting/freezing point in the region where you need it.

[ejeffrey]

You could measure the temperature of eutectic solder  

[amspire]

Apparently using the Callendar - van Dusen method, you can fully calibrate a platinum RTD with just 4 common, everyday temperatures:

Boiling point of Oxygen (-182.96 °C)
Freezing point of water (0 °C)
Boiling point of Water (100 °C)
Freezing point of Zinc (419.53 °C)

Sounds really simple, but I don't think I will bother. 

I can check 0 °C as precisely as I want, and the Class A platinum RTD's are accurate enough for anything I need. Rather then absolute accuracy, I need linearity and repeatability and I think I now have that.

Richard.

[Conrad Hoffman]

Harder than it looks, but fractional degree calibrations aren't necessary for much. FWIW, here's the triple point cell link-
http://www.scientificamerican.com/article.cfm?id=tackling-the-triple-point

[amspire]

A very interesting article. 0.0001 deg C accuracy is impressive.

But simply freezing water gives better then 0.01 deg C, and that is good enough.

Richard

[grenert]

The repeated warnings about shattered glass spook me too much to try it.  I just use the crushed ice/water method in an insulated bottle.

[robrenz]

I finally received the two Platinum Resistance Thermometers I had ordered:

Richard,

I noticed these are 2 wire probes. Most RTD meter temp specs are based on 4 wire ohms (4 wire RTD).  You may want to add two wires right at the potting or use a 4 wire connection right at the potting for maximum accuracy.  Please keep us posted on your project progress, I love learning about the finer details of metrology.

robrenz

[lowimpedance]

I bought some mineral oil today from our local Chemist. In Australia, it seems to called Paraffin Oil rather then Mineral Oil.

Richard

Richard
 For the resistor oil baths at work we use a Mobil product called WHITEREX 307 , white mineral oil.
Looking through the data when you get to freezing point it states it has not been established !!
 A newer product being used is called Pionier 2076P (H&R group)
John