Looking for temperature sensors


Thread Starter



I am working on a project design, to build an academic pilot plant distillation of petroleum. It is a 2 meters high tower with exchangers and associated pumps. It is necessary to sense the temperature of about 10 different locations, with a range of up to 400 degrees Celsius. Since minimizing costs is the first goal, it will be accepted that the accuracy of the measurements is not high priority. These measures will be wired to a PLC Siemens Simatic S7-200. Please I would greatly appreciate if you can suggest me what type of sensor/transmitter to acquire. I think that there are many possible solutions, and it will be very useful if you can give manufactures and models names. Thanks very much for this great site.

My best regards.
As you mentioned above there are various options and you are basically limited to how much you want to spend.

For all field sensors there is only one choice and that is to use RTD’s (Resistance Temperature Detectors) but doing the installation of these detectors might require thermowells to be installed on the column. These thermowells are basically just metal pockets that is screwed into the process line or vessel or attached to the line or vessel with a flange attached to the well. The main reason for these thermowells are for protection for the technician that have to replace these RTD’s that screw into these thermowells. This is especially important when the process is hot or corrosive. If you are sure your column is safe and don’t want to use the thermowells you can, the RTD’s will still work perfectly but like I said it is always safer to use the thermowells. Also the thermal transfer rate of the heat will be slightly higher since the heat will have to pass through the well wall before it gets to the RTD tip where the actual temperature measurement is done. This is normally not a major problem since temperature loops are very slow anyway. We normally pore a little bit of oil into the thermowell and make sure that the tip of the RTD presses hard against the bottom wall of the thermowell to try and minimize this thermal transfer delay. Finally you need to determine how long these RTD’s and thermowell needs to be. Keep in mind that the actual measurement is done at the tip of the RTD so use a length that will put the tip into direct contact with the process. You should never have the RTD tip going into deep since that might increase the risk of damage to the thermo well if the process is fast flowing and corrosive and you also should never have the RTD tip inside a pipe attached to the vessel that is not in direct contact with the process point you are trying to measure. Typically going into the process with the RTD tip just one inch is more than enough.

Next you need to decide if you want a local indication where you install these RTD’s on the column in which case you can get a small Temperature transmitter that fits onto the RTD inside a small aluminum housing that will convert the resistance of the RTD into a 4 to 20mA signal which can then be taken to a normal Analogue input card at the PLC. These local temperature transmitters can be bought with local digital displays on them so that you can see what the temperatures are at specific point locally in the field where these RTD’s are installed. The 4 to 20mA signal that is connected to the PLC must be scaled inside the PLC software to the same range as the calibrated range in the temperature transmitter. In this case your calibrated range in the transmitter will be 0 to 600 Degrees Celsius and that will give you a corresponding 4 to 20mA signal out that you can then input into the Analogue card of the PLC and then scale the PLC to the same values. This is very important otherwise you will never get accurate readings. Note that I have used a calibrated range of 0 to 600 and not 0 to 400. It is always best to make your calibrated range slightly bigger than the maximum expected process condition. If there is a chance that your process might go up to say 1000 Degrees, use a calibrated range of 1300 Degrees and so on. Yes to your question, a setup like this will give you perfectly accurate readings and there is no need to assume that the temperature measurements of your column will not be accurate.

Another way you can go is to take all the wires from all the RTD’s on the column into a small common Junction Box at the column and from there run a multicore cable to the PLC. At the PLC you then use a special input card that can take RTD inputs directly. The PLC will then do the conversion and line compensations automatically and again you can configure these inputs for however you want them to be displayed on your supervisory system. The RTD input card is basically taking the place and doing the job of the temperature transmitter in the field. Personally I don’t like doing it this way since I prefer to keep these flimsy RTD wires as short as possible and prefer to transfer my signals in 4 to 20mA format. I also like having a local indication in the field but that’s just me. This type of setup might be the cheapest way to go for you.

Here’s a link to help you on the right track.
It sounds like the pilot plant will be in a lab
Check to see if there are modules for your PLC that will read RTDs or thermocouples directly without transmitters. This is not only cheaper but more accurate.

In our lab we quite often use sheathed temperature sensors directly in the process without thermowells, it's safe enough because you never work on the pilot plant while it's hot or under pressure.