Choosing a temperature sensor for PLC fan control

Which temperature sensor works with a PLC for room fan control?

Buy a 3-wire PT100 RTD for the probe side: it reads 100 Ω at 0 °C and climbs about 0.385 Ω per °C under IEC 60751, and almost every PLC temperature input module lists PT100 among the sensor types it accepts. IEC 60751 is the standard that defines the PT100 element, its 100 Ω ice-point value and its resistance-temperature table, so any module built to it reads the same probe the same way. For a room loop the RTD also wins on effort: a base PT100 probe is a commodity part, and the hard measuring happens in the input module. A thermocouple is the sensor for hot surfaces and flues, where its microvolt signal grows; at room temperature it is the weakest choice on the board. A thermistor is a fine sensor inside the enclosure it was designed for and an awkward guest anywhere else. The fan command itself is a discrete output and the run proof is a discrete input, so the temperature channel is the only analog decision in the project.

Recommendation: for one room and one fan, order a Class B, 3-wire PT100 probe plus the cheapest reading path the rack supports: a dedicated RTD input module if a slot is free, otherwise the same probe behind a 4-20 mA transmitter into the spare analog input. Land a spare normally open auxiliary contact from the motor starter on a 24 V DC discrete input, and raise the failed-start alarm after 5 s of command with no feedback.

What does the input hardware you already own decide?

Read the rack and the module labels before any sensor purchase: a fitted temperature module dictates the sensor type from its documentation, a plain voltage or current analog input forces a transmitter or a different module, and no spare input puts a standalone temperature controller on the table. Three cases cover every rack:

• A temperature module is already fitted. Its documentation lists the sensor types it accepts, PT100, PT1000, Ni100 and thermocouple letters among them, along with the 2-, 3- or 4-wire diagrams. Choose the probe from that list and the scaling question disappears.
• The spares are plain voltage or current analog inputs. The sensor choice is now PT100 plus a 4-20 mA transmitter, because a standard input offers no excitation current for a bare resistive element and no curve table for a thermistor.
• No spare input at all. Decide by what the project must keep. A thermostat can start the fan, but the temperature value on the HMI, the run monitoring and the alarm history drop away with it; for a panel that already reports to operators, one added input module is usually the smaller job.

Can a packaged HVAC thermistor feed a standard analog input directly?

No, a typical voltage or current analog channel measures volts or milliamps and has no excitation current for a thermistor, so wiring a bare 10 kΩ NTC to one gives a dead channel or a meaningless reading, and the curve behind it is nonlinear. Put numbers on it. Feed a 10 kΩ NTC thermistor from 5 V through a 10 kΩ fixed resistor and read across the fixed resistor. At 25 °C the thermistor is 10 kΩ, so the node sits at 2.50 V. At 0 °C the same probe climbs to about 33.6 kΩ on the common beta model with B = 3950 K, and the node falls to 5 V × 10 kΩ ÷ (10 kΩ + 33.6 kΩ) = 1.15 V. That 1.35 V spread over 25 °C is the entire signal, it bends nonlinearly, and it moves whenever the 5 V rail moves. Turning counts into degrees then takes a lookup table and a clean, stable excitation source, which is measurement engineering bolted onto an input built to read volts or milliamps. If the excitation would come from a controller port, budget the 5 V auxiliary supply current for the port first. For the packaged HVAC monitor, the practical options are its own output, if it has a relay or a transmitter you can land on the PLC, or a PT100 in its place.

Thermocouple, RTD or thermistor: which wins for room air?

A PT100 RTD wins for room air: its signal is a nearly linear 0.385 Ω per °C over the 0 to 50 °C band a supply fan cares about, while a type K thermocouple makes only about 40 µV per °C and a thermistor follows a steep, nonlinear curve.

A thermocouple belongs on hot work: a dedicated thermocouple input exists to amplify the microvolt signal and hold cold-junction compensation, and at 0 to 50 °C that machinery earns nothing. A thermistor is excellent inside the device built around it, where the curve table and excitation ship with the product. The transmitter row is the bridge: any rack with a standard analog input reaches PT100 accuracy through it. Check the receiver resistance the input presents, in its own documentation, before scaling; 250 Ω is the common value that maps the 4-20 mA span onto 1-5 V.

How do you wire a PT100 when the sensor sits far from the panel?

Order 3-wire beyond a few metres: 2-wire wiring adds both lead resistances on top of the element, and 100 ft of AWG 22 copper, about 16 Ω per 1000 ft, puts roughly 3.2 Ω in series, which reads as about 8 °C of high error on a PT100. The arithmetic behind that 8 °C: 100 ft out and back is 200 ft of conductor, 200 ft × 16 Ω ÷ 1000 ft = 3.2 Ω, and 3.2 Ω ÷ 0.385 Ω/°C = 8.3 °C of apparent warmth on the reading. A 3-wire connection lets the module measure a lead loop and subtract it; a 4-wire connection removes lead resistance outright and earns its extra conductor past a few hundred feet or beside motor cable. Run shielded cable and land the shield at the panel end. Isolate and lock out per site procedure, and check every terminal against the module's diagram before re-energizing. If a fitted installation reads jumpy or wanders, start at the wiring and the base rather than at the sensor; the checklist for erratic analog input readings follows that order.

How does the raw signal become degrees on the screen?

Scaling is one linear map: a transmitter ranged 0-50 °C sends 4 mA at 0 °C and 20 mA at 50 °C, so a 250 Ω receiver turns 13.6 mA into 3.40 V and the scaled value reads 30.0 °C. The general map is scaled °C = (raw − raw at 4 mA) ÷ (raw at 20 mA − raw at 4 mA) × 50 °C. Check it in volts: 13.6 mA × 250 Ω = 3.40 V; (3.40 − 1.00) ÷ (5.00 − 1.00) = 0.60; 0.60 × 50 °C = 30.0 °C. A reading that lands close but sits at a fixed offset usually points at the scaling, and the walkthrough of why the scale value differs from the PLC analog input shows how raw counts become engineering units. On a dedicated RTD module the degrees arrive directly; confirm the configured sensor type on the module, PT100 in this project, matches the probe on the cable. The panel side of the job, from tag to trend, is covered in the walkthrough of displaying a thermocouple temperature on an HMI panel. Keep live values on the HMI tag. A signage notice screen plays scheduled content from a card, so it carries standing reminders such as 'supply fan runs above 30 °C', and a shop-floor notice screen suits the places where the people who need that reminder never open the panel.

What proves the fan motor is actually running?

Wire a spare normally open auxiliary contact from the motor starter to one 24 V DC discrete input, then alarm when the run command has held for 5 s with no feedback; the command bit is not proof that the contactor pulled in. Four states tell the whole story:

• Command off, feedback off: normal below the stop point; sanity-check the temperature value.
• Command on, feedback on: the starter pulled in; confirm airflow at the grille once during commissioning.
• Command on for more than 5 s, feedback off: failed start or a broken feedback path; check the starter, the overload device, the control wiring and the input point, and leave the alarm standing until the cause is fixed.
• Command off, feedback on: a welded contactor or a stray signal; treat the fan as running and pull control power to investigate.

A current-sensing relay around one motor lead is the alternative run proof, and it answers a different question: the motor draws current, which a jammed or broken drive can still do. Where moving air is the requirement, an airflow or differential-pressure switch does that job. Many fan motors also carry PTC thermistor sensors in the windings; route them through a motor-protection relay and land that relay's trip contact on its own discrete input so an overload announces itself in the same scheme, and if the terminal box mixes those leads with brake wiring, the page on identifying brake and thermistor leads at the motor auxiliary terminals sorts them out. The same feedback input doubles as a run-hours meter: counting its on-time gives maintenance hours, as the article on showing motor run hours on an HMI from a PLC counter walks through.

Worked example: which parts fit a 30 °C plant-room fan start?

A plant room starts its fan above 30 °C and stops it below 28 °C; the rack has one spare 4-20 mA input, the probe site sits 100 ft from the panel, and the parts list becomes a 3-wire PT100 probe, a loop-powered transmitter, and a starter auxiliary contact. The arithmetic, step by step:

1. Element check. At the 30 °C start point a PT100 element sits at about 100 Ω + 0.385 Ω/°C × 30 °C = 111.6 Ω.
2. Why 3-wire. As a 2-wire run it adds 200 ft × 16 Ω ÷ 1000 ft = 3.2 Ω, which reads as 3.2 Ω ÷ 0.385 Ω/°C ≈ 8 °C high: the PLC would show 30 °C while the room sat near 22 °C and start the fan eight degrees early.
3. Transmitter. Ranged 0-50 °C, it sends 4 mA + (30 ÷ 50) × 16 mA = 13.6 mA at the start point, and across the input's 250 Ω that is 13.6 mA × 250 Ω = 3.40 V.
4. Scaling. (3.40 − 1.00) ÷ (5.00 − 1.00) × 50 °C = 30.0 °C on the tag.
5. Logic. Start when the tag reaches 30 °C, stop below 28 °C so the 2 °C band keeps the starter from chattering at the limit, alarm when the run command holds 5 s with no feedback, and alarm when feedback appears with the command off.

The parts list: one Class B 3-wire PT100 probe with a stem to suit the wall pocket, shielded 3-wire cable, a loop-powered PT100 transmitter ranged 0-50 °C, 24 V DC for the loop from a DIN-rail supply in the panel, and the starter's normally open auxiliary contact block. With a spare RTD input module instead of the transmitter, step 3 disappears: the probe wires 3-wire straight to the module and the tag reads °C directly.

What do you order, and when does a standalone thermostat fit better?

Order four lines, a Class B 3-wire PT100 probe, shielded 3-wire cable, an RTD input module or a 4-20 mA transmitter with its 24 V DC feed, and the starter's auxiliary contact block, and choose a standalone thermostat only when alarms and run status are not wanted. The order lines, in the words that get the right part:

• Probe: PT100, Class B, 3-wire, or 4-wire if the input accepts it, with stem length and mounting fitting for the pocket or duct.
• Cable: shielded, 3 conductors, copper, sized by the lead-resistance arithmetic above for the run length.
• Input: an RTD input module configured for PT100 3-wire, or a PT100 transmitter ranged for the room band with its 24 V DC loop feed.
• Feedback: one normally open auxiliary contact on the starter, or a current-sensing relay, onto a spare 24 V DC discrete input.

A standalone electronic thermostat with a fan-rated output contact is the right call when nobody needs the temperature value, the run status or an alarm history, and it is the least wiring by a wide margin. Keep it as a temporary operating control only where the site accepts losing the monitoring and alarms while it is in place. Several rooms tip toward a multichannel RTD module with one probe per room and the same feedback scheme repeated per fan. The decision, case by case: temperature module present, order the probe from its sensor list and wire 3-wire; only standard analog inputs, order PT100 plus a 4-20 mA transmitter; no spare input, weigh one added module against a standalone controller by the alarms you must keep; and in every case, land run proof on a real contact.

Questions