Analog signal types: 4-20 mA, 0-10 V, RTD, thermocouple

Analog signal types is the part of instrumentation that gets glossed over in every PLC textbook and then becomes the reason a junior engineer wastes their first month on a new plant chasing a temperature reading that drifts by three degrees every time the sun comes up. The four signal types every SA technician meets — 4-20 mA current loop, 0-10 V voltage, RTD resistance, thermocouple millivolts — each have their own physics, their own wiring rules, their own module configuration, and their own ways of going wrong. Knowing which one to pick before you spec a job, and how to read each one on a Siemens or Rockwell analog input card, is the basic instrumentation literacy that separates a competent technician from a paid spectator.

Why this matters on real plants

A wrongly specified analog signal type costs more than just the input card. We saw a brownfield FMCG bottling line where the plant designer had specified 0-10 V temperature transmitters on every oven loop, and the panel was fifty metres from the field. By the time the cable run hit the input card, every signal had picked up a few hundred millivolts of induced noise from the parallel motor cables in the same tray. The temperature readings danced by half a degree every time a contactor pulled in. The PID loops chattered. The cold-side reject rate climbed. Six months and a forty-thousand-rand rewire later, every loop on the plant was 4-20 mA and the readings settled. The original designer had picked 0-10 V because the transmitters were R200 cheaper each. Forty units, eight kilorands saved on procurement, four hundred kilorands lost on the rework.

The other side of this is over-specification. A mining client asked for Pt100 RTDs on every motor winding because somebody told them RTDs are "more accurate than thermocouples". The motors run at 80 to 120 degrees C. Either signal type would have been fine. RTDs cost three times more, take more wiring (3-wire or 4-wire to compensate for lead resistance), and need an RTD-input card or a four-wire universal card. Thermocouples on a Type K with cold-junction compensation would have been the right answer at a third of the budget. Picking the right signal type up front is cheaper than retrofitting later, and the right pick depends on four factors: required precision, distance to the panel, electromagnetic environment, and the temperature or pressure or flow range you are actually trying to measure.

The third reason this matters: when an instrument loop fails on a Sunday afternoon during load-shedding and the artisan has to swap a transmitter from stores, knowing which signal types can be swapped for which without re-engineering the loop is the difference between a thirty-minute repair and a twelve-hour outage. A 4-20 mA pressure transmitter from one OEM is a drop-in replacement for a 4-20 mA pressure transmitter from another OEM as long as the range matches. A 0-10 V transmitter is not. A thermocouple Type K cannot replace a Type J without re-configuring the input card. Knowing the substitution boundaries is the operational knowledge that makes a senior artisan worth keeping on the plant.

The mental model — four physics, one decision tree

Each analog signal type uses a different physical principle to encode the measurement, and the principle determines what the signal can and cannot survive.

4-20 mA current loop. The transmitter sources a current proportional to the measured value, with 4 mA representing the bottom of the range and 20 mA the top. Current is conserved along the loop, so the value at the panel is the value at the field, regardless of cable length or resistance, up to the loop's compliance voltage. The 4 mA "live zero" doubles as a wire-break detection — any reading below ~3.6 mA means an open circuit. This is the workhorse of process instrumentation for a reason. Long runs, immune to most noise, self-diagnostic.

0-10 V voltage. The transmitter outputs a voltage proportional to the measured value. Cheap, simple, but voltage drops across cable resistance and picks up induced noise from anything radiating in the same tray. Useful only for short runs (under ten metres), inside the same panel, or where cost is the binding constraint and precision is not.

RTD (Pt100, Pt1000). A platinum element whose resistance changes linearly with temperature — 100 ohms at 0 degrees C for a Pt100, rising to about 138 ohms at 100 degrees C. The PLC analog card sources a small known current through the element and measures the voltage drop. Highly accurate, stable across years, but the lead resistance of the cable becomes a measurement error unless you wire it 3-wire (two leads on one side, one on the other) or 4-wire (Kelvin sensing, two leads on each side). The diagram for the AI8 module on this page shows the standard 3-wire wiring topology.

Thermocouple. Two dissimilar metals joined at a measurement point produce a small voltage proportional to the temperature difference between the junction and the reference (cold-junction). Type K (chromel-alumel) covers -200 to +1370 degrees C and is the SA workhorse for boiler, oven, and furnace work. Type J (iron-constantan) covers -210 to +760 degrees C and is common on plastic-extrusion lines. The signal is small — tens of millivolts at full scale — so the cable run matters and the cold-junction compensation has to be done either at the card (most modern cards do this) or with a separate cold-junction reference block. Thermocouples are cheap, fast, hold up well against vibration, and have the widest range of any temperature sensor.

The decision tree, written as four questions:

1. What am I measuring? Temperature? Go to RTD or thermocouple. Pressure, flow, level, position? Stay with 4-20 mA or 0-10 V depending on distance.
2. How far is the field from the panel? Under 10 m, in-panel? 0-10 V is fine. More than 10 m, especially on a noisy plant? 4-20 mA.
3. For temperature: precision needed, range, environment? Better than 0.5 degrees C accuracy, range under 600 degrees, no vibration? RTD. Wider range, vibration, faster response, less precision needed? Thermocouple.
4. Existing infrastructure? If the existing panel has only RTD-input cards, that constrains the choice. If the plant standard is 4-20 mA on everything, follow the standard.

Worked example — picking RTD vs thermocouple for a thermowell

Open the simulator and drop an S7-1500 CPU on the rack. Add an analog input AI8 module — the eight-channel universal input card. The simulator's module-config dialog has a per-channel selector: 4-20 mA, 0-10 V, RTD (Pt100 / Pt1000), or thermocouple (Type K, J, T, N, R, S, B). Each channel can be configured independently, which is the strength of a universal card and one of the reasons we recommend specifying universal cards on new builds rather than dedicated single-type cards.

The scenario: a chemical reactor with a thermowell penetrating into the process fluid. Required range 20 to 250 degrees C. Required precision: 0.5 degrees C absolute, 0.1 degrees C repeatability. Distance from the thermowell to the PLC panel: 35 metres, parallel to a 22 kW VFD-driven motor cable for half that run.

Walk through the decision tree. Temperature, so the choice is RTD or thermocouple. The 0.5 degree precision target favours an RTD — a Class A Pt100 will give you ~0.15 degrees at 0 degrees C, ~0.35 at 250. A Type K thermocouple has a Class 1 tolerance of 1.5 degrees or 0.4% of reading, whichever is greater — at 250 degrees that is 1.5 degrees, three times worse than the requirement. RTD wins on precision.

The 35-metre run alongside a VFD cable is the second factor. RTD signals are sourced and measured by the analog card, so the cable is part of the measurement loop. Lead resistance becomes a real error: a 35-metre run of standard 1.5 mm² cable adds about 0.5 ohms per leg, and on a Pt100 that is roughly 1.3 degrees C of error per leg unless you compensate. The fix is 3-wire RTD wiring, which the AI8 module handles natively — set the channel to "RTD 3-wire Pt100" in the module config and the card uses a third lead to measure and subtract the lead resistance. Even better is 4-wire Kelvin sensing if you have the spare card terminals.

Configure the channel in the simulator:

// AI8 channel 0 configuration (Siemens TIA Portal-style notation)
Channel_0_Type     := RTD_Pt100_3WIRE;
Channel_0_Range    := -100_to_+850_degC;
Channel_0_Filter   := 50_Hz;       // SA mains frequency rejection
Channel_0_BurnoutEnable := TRUE;

The 50 Hz filter rejects mains-frequency interference picked up from the parallel VFD cable — set to 50 Hz on SA plants, 60 Hz on equipment imported from the US or sourced from a North American OEM. This is the kind of detail every commissioning engineer in SA has set wrong at least once. The simulator's diagnostic page on the AI8 module lights up a "noise rejection" indicator when the filter is matched to mains frequency.

The wiring on the simulator's panel view:

Pt100 element  -----+------- AI8 ch0 IC+ (current source +)
                    |
                    +------- AI8 ch0 V+  (voltage sense +, third wire)
                |
Pt100 element  -+----------- AI8 ch0 V-  (voltage sense -)
                |
                +----------- AI8 ch0 IC- (current source -)

That is 3-wire RTD: one wire from one end of the element, two wires bridged at the other end. The card injects a known current through the element via IC+ and IC-, measures the voltage drop across V+ and V-, computes resistance, looks up temperature from the standard Pt100 curve, and returns the value in tenths of a degree on the input image. No external scaling block needed. Universal AI cards do the conversion in firmware.

If the same thermowell were on a furnace running 200-1100 degrees C instead, the answer flips. RTD elements struggle above 600. A Type K thermocouple is the right pick. Two-wire compensating cable from the thermowell to the panel terminal, then standard copper from the panel terminal to the AI8 (the cold-junction compensation lives on the card). The compensating cable is non-negotiable — using copper between the thermowell and the panel would create a second junction at the panel that introduces an offset error proportional to the panel's ambient temperature, which on an SA shop floor in summer can be plus thirty.

Common mistakes

• Confusing 0-10 V with 1-5 V. Some legacy transmitters output 1-5 V instead of 0-10 V, and the offset is identical to the 4-20 mA live-zero idea — 1 V represents the bottom of the range, 5 V the top. Configuring the card for 0-10 V on a 1-5 V transmitter gives you 10% offset at the bottom and 50% offset at the top. The fix is reading the transmitter's nameplate before the loop is wired, and configuring the card for the actual signal range. Most modern AI cards have a 1-5 V option in the channel-type selector — use it.

• Missing cold-junction compensator on thermocouples. Every thermocouple measurement needs a known reference temperature at the cold end. Modern AI cards have a built-in cold-junction sensor (usually a thermistor on the terminal block) and apply the compensation automatically — but only if you wire the thermocouple compensating cable directly to the card's terminal block, not via an intermediate copper splice. Splice the compensating cable to copper at the panel and you create an unintended cold junction that adds a temperature-dependent offset. The symptom is a temperature reading that drifts up in the morning when the panel ambient warms and drifts back down at night. We have seen this misdiagnosed as "transmitter drift" for months.

• 2-wire RTDs over long distance reading lead resistance. A 2-wire RTD lumps the lead resistance into the element resistance and reports it as a higher temperature than reality. On a five-metre in-panel run the error is small. On a fifty-metre cable run with thin wire it can be five degrees of bias. Always specify 3-wire or 4-wire RTDs for runs longer than ten metres, and always configure the card to match. A 4-wire RTD on a card configured for 3-wire ignores the fourth lead and gives you 3-wire performance, not 4-wire — the configuration matters as much as the wiring.

• Mixing 4-20 mA HART with non-HART card. HART is a digital signal modulated on top of a 4-20 mA loop, used for transmitter configuration and secondary variable readback. A standard 4-20 mA AI card sees only the analog current and ignores the HART data — which is fine if you do not need HART. But a HART-enabled transmitter on a non-HART card means you cannot configure the transmitter from the panel; you have to use a handheld HART communicator at the field junction box. Specify HART-capable AI cards on new builds where any of the transmitters are HART-enabled. The cost difference is minimal and the operational convenience is large.

How to practise this in the simulator

The simulator's AI8 module supports all four signal types per channel. Configure channel 0 as 4-20 mA on a tank-level transmitter, channel 1 as 0-10 V on a position feedback, channel 2 as Pt100 3-wire on a reactor RTD, and channel 3 as Type K thermocouple on a furnace probe. Open the diagnostic page on the module and watch each channel's raw value and converted value side by side. Now break each one: open the wire on the 4-20 mA loop and watch the under-range alarm. Short two leads on the Pt100 and watch the resistance go to zero. Disconnect the compensating cable on the thermocouple and watch the cold-junction error grow. Each of these failure modes appears at least once a year on a real plant — practising them in the simulator means you recognise the symptom before you waste an afternoon on the wrong fix.

Vendor reference

The instrumentation symbol standard you will see on every P&ID in SA process work is ISA-5.1, and the ISA 5.1 Instrumentation Symbols and Identification standard is the cross-vendor reference for what each symbol on a piping-and-instrumentation diagram represents. Read the section on instrument bubbles and signal-line conventions once. The dashed-line conventions for electrical, pneumatic, and hydraulic signals on a P&ID let you read the analog signal type off the drawing before you ever look at the loop sheet.

What we don't claim

This site is not SAQA-registered, not MerSETA-accredited, and not an NQF-registered qualification provider. Our completion certificates are course-level only — they describe what you covered, not an NQF Level X qualification. The CCST cert from ISA is the portable industry credential we recommend; we are not an ISA cert delivery partner either, but our cert packs are CCST-aligned. Analog signal selection and wiring is a major topic on the CCST Level 1 syllabus and the patterns shown on this page are aligned with the way the CCST handbook presents the same material.