The actual signal (resistance value voltage whatever... ) from the PT100 thermocouple is processed by the Rumba board, I think the PT100 board translates the reading to one that the Rumba will understand as it is designed for a thermistor. What is displayed on screen and in Octoprint is produced by the processor chip using settings in the firmware on the Rumba, if part of the process of calculating the actual temperature to be displayed is to average/smooth the displayed value then it is going to remove spurious values caused by interference. Someone with knowledge of the PID processing would be able to say if there was any smoothing/averaging taking place and how the PID tuning parameters affected the output value. Dr Jeep is I think going to look at the actual voltage from the thermocouple ("I will graph the voltages") to see if there is interference on the actual signal on the wires between the sensor and the PT100 board. The makers of the PT100 would not have bothered with a screen if it wasn't necessary as it would be cheaper and simpler to make with just two wires and no screen. A without connectors and with connectors and no continuous screen measurements would confirm this. I have not put in connectors yet, have not done PID tuning either but I will have a look with a scope to see what the signal looks like. As is in Octoprint I get a smooth curve varying like a sine wave like the red curve in post number 6 above and without the noisy sections.