I have an upcoming project that has an extremely fast exotherm. I am told that the time involved is in the low seconds. I know that the thermocouple cards have a built-in filter to prevent aliasing. The article that I found on this through Guardian is old so it could be out of date. The article that I found said that the time constant for thermocouples was 5 seconds but was a function of sample time. It also said that this filter is disabled by default with RTDs. I don't believe that the TC or RTD will not be in a thermowell. If I want to get a reliable and accurate temperature measurement as fast as possible, would a RTD be recommended over a TC? How does adding a transmitter versus direct wiring impact this? Any suggestions on this? Thank you very much in advance!!!
In answer to your question regarding filters. In DeltaV Explorer, you can drill down through the controller to the specific channel that will receive the temperature and there is a "Filter" parameter that is NOT write protected. If you double-click on that parameter, you can change the filter setting. Unfortunately, as you mentioned, the lowest filter setting for thermocouples is five seconds. The lowest filter setting for RTDs is "Filter Disabled".
According to Books Online, "The input value is averaged over the time specified. The FILTER time must be larger than the effective scan rate of the block using the input from this channel or CHARM." I did some testing (Type K Thermocouple), and the module execution rate didn't seem to matter if it was faster than the channel filter value. I tried setting the module execution to 10 seconds, 1 second, and 200 milliseconds, with filter values of 5 seconds, 13.0 seconds and 78.0 seconds. I used DeltaV Process History View to trend the data, and that is limited to one second updates, so the 200 millisecond execution rate won't trend properly. Additionally, the response of the data from the filter was like a first order filter, where the time constant was the filter value. To me this is different than "averaged over the time specified", but I just might have an incorrect understanding of the Books Online wording.
In my experience, RTDs are more reliable, but thermocouples are faster to respond (but it also depends greatly on how much mass the temperature element has). However, the DeltaV channel filter makes thermocouples slower than RTDs. If you want really fast, you could investigate using a third party temperature transmitter, and bringing the signal into DeltaV as an analog input. We have experience doing this, but I don't think I can recommend specific third party items.
I was discussing this topic with my co-worker, and he mentioned looking at infrared temperature measurement. We also use those, and they offer fast response times.
Lastly, this is not publicly documented anywhere, but the M-Series RTD cards have a component that is common to the odd channels and another that is common to the even channels. If that component fails (which has happened to us on multiple occasions) or if any of the temperature leads get shorted, you will lose all channels associated with that component (even or odd channels). As a result of this, we have switched all of our RTD cards to analog input cards, and are using third party temperature transmitters between the temperature elements and DeltaV.