I've been using 3-wire 100 ohm platinum RTDs for years. Most have been wired to recorders, recorder/controllers or PLCs.

Having to adjust a recorder's input offset a little has been common (we used calibrated mercury in glass thermometers as a reference), but when one requires more than maybe a 2 or 3 degree correction, I get nervous. More than that and I start looking for a problem. Lead lengths are generally 100 feet or less, with the shield grounded at the instrument.

At my new location, I've already encountered 3 RTDs feeding ABB 250 DIN controllers and all three have large offsets entered to get them to match the nearest thermometer.(14+ degrees on two of them!) Measured temps are 220 to 250 degf, and RTDs are in thermowells.

I've only had access to one machine, and only briefly. The controller is perfect with an RTD simulator attached locally, and I swapped out the RTD. The lead is only about 25 feet long, and I couldn't check it or replace it (yet). Problem could be an ill-sized RTD or other thermowell related problem. Don't know yet.

Question: Is there any valid justification for a 14 degree offset? Is this typical of these DIN controllers?

 

Is it possible you really have a 2 wire RTD? Or wired it that way?

I have seen a few cases where 2 thermometers in the same pipe within a few inches of each other read ten degrees difference, but only some of the time. So I believed there really was a temperature difference. No idea how. I had to guess some kind of wierd flow condition insode the pipe.

--
Bob

 

Those kinds of offsets might be related to uncompensated lead wire resistance in 2 wire RTD connections. Although I have to admit I'm not sure how a relatively modern controller could not do leadwire compensation that it's designed for. Unless the 3rd wire isn't connected to the terminal it's supposed to be connected to.

I was stunned at the offset due to lead wire resistance the first time I ran into someone using a 2 wire RTD. It was a bench top food testing machine and there was only 6' of lead wire, but that 6 feet contributed noticeable offset. The lead wire resistance adds up and you're dealing with 200 feet of wire.

Here's a table of wire resistance for AWG wire:<pre>
gauge ohms/foot
18 0.00751
20 0.0119
22 0.0190
24 0.0302
26 0.0480
28 0.0764</pre>
Your 100 feet one way is 200 feet accounting for each side of the RTD.

For 22 gauge extension wire, that's 0.019 ohms/ft * 200ft = 3.8 ohms

From a PT100 chart, the resistance for 220°F is 140.19 ohms.

Add ~3.8 ohms and the table puts you at 239°F at 143.97 ohms.

Depending on the wire gauge and actual distance and it's close to the numbers you cite.

Check your lead wire gauge and do the arithmetic.

 

From what you have said, there may be a problem in the method you use for calibration. A liquid in glass thermometer is one of three types: fixed immersion, variable immersion and total immersion. The first has a line indicated on the thermometer where it should be immersed to in a thermostatic media (Oil or water usually) The second type is immersed to the level of the liquid indicated on the stem, and the third is totally immersed in the fluid (Like a fish tank thermometer).

The inside of a thermowell usually has temperature gradients that are never consistent. Due to the possible varying lengths of sensitivity of your glass thermometer, I would expect that in any case your RTD would be much more tip sensitive than about any glass thermometer. you will also find this phenomena exhibit itself with a bimetal thermometer or gas filled thermometers. Tip sensitivity can vary between manufacturers of RTDs and even within RTDs designed for different applications from just one manufacturer. With thin film RTDs the tip sensitivity of RTDs has become much better, but not as good as grounded thermocouples.

How to resolve your problem...
Hook a portable ohm meter up to your RTD inside the thermowell and measure the resistance between each lead. Write it down. Subtract the common wire resistance # from the resistance across the resistor itself. That should eliminate the bias of the wire length. Go to any resistance/ temperature table for RTDs and determine your temperature. Take another RTD and do the same thing. If your numbers are consistent, you should be OK. We recommend that the end user maintain a pristine temperature situation sensor for doing these "in situ" validations of temperature. It is much more practical than taking the sensor to a lab where the temperature gradients of the actual process are not considered.

Here is a link to a pdf showing the circumstances of tip sensitivity.

http://www.jms-se.com/catalog/Tip_Sensitivity.pdf