Simulating an RTD

[bigben2k]

While in the process of setting up my temperature transmitters, I hit a wall; I don't have the ability to simulate an RTD.

There are few items available cheaply to do this. Ex:
http://cgi.ebay.com/NIB-Altek-RTD-Mo...temZ7603542207
(btw, that one is on a alpha curve that is american only. European curve is 3850)

For those wishing to catch up:
http://www.capgo.com/Resources/Tempe.../RTDs/RTD.html

The problem:
I need to simulate an RTD at the following temperatures:
-200 deg C
+850 deg C
+15 deg C
+20 deg C
+25 deg C
+30 deg C
+35 deg C

This is done by applying a precicion resistor (0.01% accurate) to the input of the temperature reader.

The values are (respectively):

18.52 Ohm
390.48 Ohm
105.85 Ohm
107.79 Ohm
109.73 Ohm
111.67 Ohm
113.61 Ohm

The existing equipment:

I have two Fluke 8840A multimeters capable of displaying 0 to 199.999 Ohms (on one range) with an accuracy of somewhere around 0.01 Ohm. I have the calibration certificate for one of them, but it expired in April last year. The certificate indicated that it was almost dead on on resistance and voltage, but on the edge for mA (at ~87% of tolerance). The mA scale is going to be useless for this though.

I have two REF192FS precision voltage references
http://www.analog.com/en/prod/0,,769_838_REF192,00.html
Good for 2.500 volts, +/- 0.002 volts. Can also be configured as a current reference.

I have a few AD7713 that could be used as a 200 uA source, if I can figure out how to turn it on, without building the complete circuit around it.
http://www.analog.com/en/prod/0,2877,AD7713,00.html

The non-existing equipment:
A bunch of precision resistors available on eBay, with 0.01% accuracy:
http://cgi.ebay.com/ws/eBayISAPI.dll...tem=2558643929
http://cgi.ebay.com/ws/eBayISAPI.dll...tem=7604640652
http://cgi.ebay.com/ws/eBayISAPI.dll...tem=7552610442
http://cgi.ebay.com/ws/eBayISAPI.dll...tem=7602043682
(that last one isn't available anymore, but I'm hoping it'll be relisted shortly)


The question:
How do I simulate an RTD using the above?


Hint: I'm thinking about using an array of variable resistors, in parallel, to get to the specific resistances. Normally, one would use a decade resistance box, with 0.01% accuracy, or a resistance simulator of some kind. Both of these are insanely expensive, and I'm looking for a cheaper option.

[laxman]

Your problem is that RTDs have way to low of a slope. Meaning a .01% error in resistence is a around a 3c error. This does not fit calibration standards by a long shot. Do any of you multimeters have a have a ohm range of 2000? If they do then you can cross calibrate your RTD and thermometer by using a NTC thermistor. I cant remeber the series but one has a resistence of around 1500 Ohms at 30c. Using the known temperature/resistence relationship of the thermistor you can pretty accuratly measure the temperature in a controlled water bath. Then you can use that temperature to calibrate your thermometers.

Although I think that it might be easier if you just had someone else calibrate them. I could do it, I have a temperature calibrator and two thermometers that are accurate within .005c (NIST traceable too).

[bigben2k]

I hear you. I might take you up on that offer, I'll let you know.

Following:
http://www.capgo.com/Resources/Tempe.../RTDs/RTD.html
I'm calculating that a 0.01% error (0.01 ohm) translates into a temperature variation of 0.025 degrees C, for a PT100 RTD. How do you figure 3 deg C? 1% error?

I guess I was having a rough day. The answer is to pick up a few resistors at +/- 0.01%, and see how my DMMs respond to it.

What I still don't know is how I'm going to create a specific resistor. Right now, an array of 10 * 1K variable resistor looks promising, but I don't know if it'll be feasable to get to an ohm value within 0.01. Not entirely sure how I'd go about building something like this either.

I'm also missing the proper probes; I have a simple set of (single wire) probes from probemaster.com, but I need to use 4 wire probes for this. Gonna have to dig up some more info on that too.

[bigben2k]

Ran some quick calcs:
If I use 10 * 1K pots, and assuming that I can set each one within 1 degree of rotation (maybe... with an extended arm), giving me a 4 ohm variation on each pot, I could get an output within 0.04 Ohms.

(Based on a 240 deg turn angle, linear progression potentiometer).

[Ice Czar]

my plan is to build cells for in-lab calibration based on the best info I can glean off NIST and my new vaccum pump
going for a
melting point of mercury -38.8344°C
triple point of water 0.01°C
melting point of gallium 29.7646°C

fisher has some pretty "clean" metal available (99.9999%)

till I can get the sensors professionally calibrated
currently Ive been following old school NIST shaved ice proceedure

[bigben2k]

Ooh nice, I'll be in touch!

I figured that I can use an array of 200 ohm multiturn (25) potentiometers to simulate and RTD. I'll get the parts, build it, and report results. (I also have to pick up another set of probes, and some reference resistances).

[Ice Czar]

I posted this once before but....


Tackling The Triple Point Shawn Carlson Scientfic American Jan 99
(build your own triple point cell, a sustainable calibration point of 0.01 C)

Calibrating With Cold Shawn Carlson Scientfic American Dec 2000
(based on the freezing point of mercury calibration point is –34.8 degrees C
also includes a proceedure to compensate for a boiling point calibration of 100C)

Homemade High Precision Thermometer Shawn Carlson Scientfic American Mar 99
(its a DIY 4 wire RTD)

One of the horrible truths of scientific research is that simple and inexpensive techniques will get you just so far. Beyond some point, increasing accuracy can be obtained only with a disproportional rise in expense, sweat and frustration. That's partly because accurate measurements require an extremely well calibrated instrument, and providing such an exact scale can be a vexing challenge.

Triple point of water

The single combination of pressure and temperature at which water, ice, and water vapour can coexist in a stable equilibrium occurs at exactly 273.16 kelvins (0.01 °C) and a pressure of 611.73 pascals (ca. 6 millibars). At that point, it is possible to change all of the substance to ice, water, or steam by making infinitesimally small changes in pressure and temperature. (Note that the pressure referred to here is the vapor pressure of the substance, not the total pressure of the entire system.)

Water has an unusual and complex phase diagram, although this does not affect general comments about the triple point. At high temperatures, increasing pressure results in first liquid, and then solid water (above around 109 Pa a crystalline form of ice which is denser than water forms). At lower temperatures the liquid state ceases to appear with compression causing the state to pass directly from gas to solid. It is, however, possible to melt ice by increasing pressure under specific conditions.

At a constant pressure higher than the triple point, heating ice necessarily passes from ice to liquid then to steam. In pressures below the triple point, such as in outer space where the pressure is low, liquid water cannot exist: Ice skips the liquid stage and becomes steam on heating, in a process known as sublimation.

thats not to say Im not looking into quartz reference currently
but at this point calibrating everything professionally would cost somewhere around 5 times the money Ive spent acquiring the equipment

[bigben2k]

Am still prepping this simulator.
Here's the list of resistors I got:
(1st column indicates quantity)

2 40Ω/0.5% (new)
2 129.401Ω/.01% (used)
2 200Ω/.01% (new)
2 300Ω/.01% (new)
2 2kΩ/.01% (used)
2 2.3333kΩ/.01% (new)
2 2.5kΩ/.01% (used)
2 5kΩ/.01% (used)
2 10kΩ/.01% (used)
2 13.889kΩ/.01% (new)
2 20kΩ/.01% (new)
2 20.41kΩ/.01% (new)
2 32kΩ/.01% (used)
2 40kΩ/.01% (used)
2 259.523kΩ/.01% (used)
2 16MΩ/.1% (new)

1 64 Ω 0.01% new
1 300 Ω 0.25% new
1.9 kΩ 0.01% used
1 2.0 kΩ 0.05% used
1 15 MΩ 0.1% used
1 10 kΩ 0.01% used
1 14.250 kΩ 0.01% used
1 5.410 kΩ 0.01% used
1 14.950 kΩ 0.01% used
1 31.175 kΩ 0.01% used

I've been advised by Bill that this is futile. I intend to learn the hard way (but some things become rapidly obvious). I will pickup another set of probes for 4-wire measurements on both my Fluke 8840A. This should help me determine what the Flukes can do, exactly.

I picked up a Keithly 193 from Bill, for temps only (The RTD measurement are "flaky").

A 200Ω resistor @ .01% puts this resistor's actual value somewhere between 199.98 and 200.02 Ohms.

I got a pack of twenty, 100 Ohm, multi-turn (15 turns) "Bourns Trimming Pots". (p/n 3009P-1-101).
ref: http://www.bourns.com/pdfs/3009.pdf
(PDF, one page, 85k)
I'll be building a 4 by 4 array of these, and see if I can replicate a specific resistance from there.

More later.