KEYSTONEIndustrial Services
Interactive trainer

Thermocouples & RTDs

A temperature reading is only as good as the sensor, its wiring and the instrument setup — the number on the screen is a conclusion, not a measurement. Six loops, four reads — indicated vs a calibrated reference, the raw sensor mV or Ω, the leads and the config — including the two look-alike pairs: a reversed thermocouple versus a wrong type, and a 2-wire RTD lead offset versus an open element.

By Francisco Padilla Interactive Technician level

Before you start — job brief

A temperature reading is only as trustworthy as the sensor, its wiring and the instrument's setup — the number on the screen is a conclusion, not a measurement. A thermocouple makes a small millivoltage from the junction of two dissimilar metals (the Seebeck effect) and needs cold-junction compensation, matching extension wire and the correct type set. An RTD is a resistance that rises with temperature and needs its lead resistance compensated. Every scenario here gives you the true process temperature from a calibrated reference in the same well — your job is to judge the indicated value against it.

  • Compare the Indicated value to a calibrated reference before you believe it
  • Read the raw Sensor value — mV for a thermocouple, Ω for an RTD — and see what temperature it implies
  • Thermocouples need correct polarity, matching extension wire and the right type configured
  • RTDs need lead compensation — a 2-wire run reads high, an open element burns upscale

The number on the screen is a conclusion, not a measurement

A temperature indication is the end of a chain: a sensor turns heat into an electrical value, field wiring carries it, and an instrument converts it back into degrees using a table and a reference. Every link can lie while the display stays confident. The discipline that catches it is simple and always the same: compare the Indicated value to a calibrated reference in the same well, read the raw Sensor value and see what temperature it really implies, check the Leads for polarity, extension type and resistance, and verify the Config matches the sensor that is actually installed. Four reads, and the loop tells you which link failed — instead of you trimming a setpoint until the number looks right and the measurement is wrong.

Thermocouples — a millivoltage from two metals

A thermocouple is a junction of two dissimilar metals that produces a small voltage proportional to temperature — the Seebeck effect. That voltage is tiny (a Type K makes about 41 microvolts per degree C) and it is a difference: the instrument only knows the hot junction's temperature if it also knows the temperature of the cold end where the thermocouple meets copper. That is cold-junction compensation — the panel measures its own terminal temperature and adds it back in. Three things have to be right for the number to mean anything: the polarity, the extension wire, and the configured type.

Get the polarity backward and the millivolts feed in with the wrong sign — the reading goes negative and slams downscale. Use the wrong extension wire — plain copper instead of the matching alloy — and you create a second, uncompensated junction out at the field head; the error then drifts with the ambient temperature at that head. Set the wrong type in the instrument — a Type J sensor read on a Type K table — and the same millivolts get looked up in the wrong column, so a plausible, positive voltage decodes to the wrong temperature. In ANSI colour code the red lead is always negative, and matching extension wire carries the sensor's own alloy right up to the panel; both exist precisely so these faults do not happen.

Field note — two pairs that look identical until one more check

A thermocouple reading wildly wrong has two very different causes. If the raw millivolts are negative and the display is downscale, the leads are reversed — and the configured type is still correct. If the raw millivolts are positive and plausible but the temperature still reads high, the leads are fine and the type is wrong — the Config shows one type while the sensor is another. Same ‘garbage reading,’ opposite fix; the sign of the millivolts and the Config tell them apart. The RTD has its own pair: a loop reading high can be a 2-wire lead offset — a finite, specific resistance like 188.7 Ω reading 447 °F — or an open element reading infinite ohms and burning upscale to overrange. Finite versus infinite resistance is the whole distinction. Always read the raw sensor value before you name the fault.

RTDs — a resistance that climbs with temperature

An RTD is the opposite idea: a precise resistor — classically a 100-ohm platinum element, Pt100 — whose resistance rises in a known way with temperature (the α=0.00385 curve is the common one). Because you are measuring a resistance of a couple of hundred ohms, the resistance of the leads matters. Run it 2-wire and the copper lead resistance adds straight onto the element, so the instrument reads a higher resistance and therefore a higher temperature — a fixed offset that grows with lead length. Run it 3-wire (or 4-wire) and the transmitter measures and subtracts that lead resistance, so it disappears. A dropped third wire is a classic ‘reads a little high’ complaint that no amount of setpoint trimming will fix.

The other RTD failure is an open element — a cracked element or a lifted lead. Infinite resistance is off the top of the curve, and transmitters are deliberately set to burn out upscale on a broken sensor, driving the indication to full-scale overrange and tripping the high alarm. That upscale reading is a safety feature, not a hot process: it is the loop announcing that its sensor is broken. Ohm the element at the head — infinite confirms the open.

The discipline

Four reads name every fault here. Compare the Indicated value against a calibrated reference and you know whether the loop is even wrong. Read the raw Sensor value — millivolts for a thermocouple, ohms for an RTD — and you see what the element is actually telling the instrument: a negative voltage, a plausible one that still reads high, a finite resistance, an infinite one. Check the Leads for polarity, extension type and lead resistance. Verify the Config matches the installed sensor. Read all four and compare each to what it should be, and a temperature loop stops being a number you hope is right and becomes a measurement you can prove.