Your Smart Transmitters Already Know More Than SCADA Sees

The transmitter knows, but nobody asked

A Coriolis meter on a loading line reports mass flow to SCADA as a clean 4-20 mA signal. It also measures density, tube temperature, drive gain, and a sensor-fault flag — and none of that reaches the control room. When the tubes started coating and drive gain climbed toward its trip point, the first anyone knew was a "sensor failure" alarm and a stopped batch. The transmitter had been complaining for two weeks in a language nobody was listening to.

That language is HART, and it was already riding on the same two wires as the 4-20 mA signal. The only thing missing was something at the SCADA end reading it.

What is actually on the wire

HART (Highway Addressable Remote Transducer) superimposes a Bell 202 FSK signal on the analog current loop: 1200 Hz for a 1, 2200 Hz for a 0, at 1200 bps. The tones average to zero current, so they ride on top of the 4-20 mA without shifting the DC value your PLC's analog card reads. Your control signal keeps working exactly as before; the digital layer is a separate conversation.

A HART device exposes up to four dynamic variables:

• PV — primary variable, the same value the 4-20 mA represents (flow, in the Coriolis case).
• SV, TV, QV — secondary, tertiary, quaternary. For the Coriolis meter: density, temperature, drive gain.

On top of those it carries a device status byte, extended device status, and a full command set. Universal Command 3 returns the loop current plus all four dynamic variables in a single transaction. Command 48 returns the additional device status — the diagnostic detail that turns "something is wrong" into "drive gain high, tube coating suspected."

The catch is speed. Point-to-point HART gives you roughly 2 updates per second, per device. That is fine for density, temperature, and diagnostics. It is nowhere near control-loop rate, which is exactly why the fast, DC 4-20 mA path still exists. Never close a control loop on a HART digital variable. Use 4-20 mA for control, HART for the secondary values and health.

Four ways to get it into SCADA

HART-enabled analog I/O

Modern analog input cards in many DCS and PLC platforms read HART digitally in the background while passing the 4-20 mA through normally. If you already own these cards, this is the least new hardware: the secondary variables and device status appear as extra tags or as data blocks the SCADA driver maps.

The trap is background scan rate. A card cycling HART across 8 or 16 channels reads each one every several seconds, sometimes tens of seconds. Fine for diagnostics, useless if you expected SV at 1-second trend resolution. Read the card's HART scan spec before you promise anyone a fast secondary variable.

HART multiplexer

A HART multiplexer (mux) sits in the marshalling cabinet, taps 16 to 32 loops through the existing terminals, and presents everything on Modbus RTU/TCP or, on newer units, an OPC or HART-IP interface. This is the retrofit answer: no changes to field wiring, no changes to the control I/O, one RS-485 or Ethernet drop back to a SCADA driver.

Two things to check during commissioning:

• Loop resistance. The FSK modem needs 230-600 Ω in the loop to develop a readable signal, typically a 250 Ω sense resistor. If someone value-engineered the loop down to a bare current input with near-zero impedance, the mux reads nothing and you will chase "communication failed" until you find the missing resistor.
• Polling budget. A mux scanning 32 devices for four variables plus status is not instant. Ask for the scan value you actually need per point and map only those Modbus registers, rather than pulling every variable from every device and wondering why the cycle takes 20 seconds.

HART-to-Modbus / HART-to-Ethernet gateway

For a handful of critical instruments, a small gateway that talks HART on one side and Modbus TCP or Ethernet on the other is cheaper and simpler than a full mux. Same resistance and polling caveats apply. The SCADA side sees plain Modbus registers, so mapping is ordinary tag work — just document which register is SV versus TV, because the gateway vendor's register map is the only thing standing between "density" and a mystery float.

WirelessHART

WirelessHART (IEC 62591) is the same HART command set over a 2.4 GHz mesh. A field gateway aggregates the mesh and presents Modbus TCP, OPC UA, or HART-IP northbound. It is the practical route for instruments with no spare wire — added later, on rotating equipment, or across a road you are not going to trench. Update rates are configured per device (commonly 1-60 seconds), so set them to what SCADA and the historian actually consume, not the fastest the device offers, because battery life pays for every update.

Where these tags belong once you have them

A HART secondary variable is not the same class of data as your control PV, and treating it like one causes trouble.

Give the health signals their own quality and alarm handling. The device status byte and Command 48 diagnostics should drive a maintenance-priority alarm, not a process alarm — a rising drive gain is a "schedule a cleaning" event, not a "stop the plant now" event, and it should never land on the operator's process alarm summary competing with real trips. Route it to a maintenance view or a CMMS notification instead.

Watch the timestamp story. A HART variable that updates every few seconds, historized on exception with a tight deadband, produces almost nothing — which looks like a dead tag. Either widen the historian's expectations for these points or sample them periodically, and make sure a stale HART value carries bad or uncertain quality rather than silently freezing at its last reading. A frozen density that reads "good" is worse than no density at all.

The commissioning check that catches most of it

Before you declare HART data live, prove one full path end to end on one instrument:

1. Read the device with a handheld communicator or asset-management tool at the field terminals — confirm the instrument talks HART at all.
2. Confirm loop resistance is in the 230-600 Ω range at the point where the mux, card, or gateway taps in.
3. Verify Command 3 returns all four dynamic variables and Command 48 returns device status through your chosen gateway.
4. Trace one secondary variable — say density — from the instrument, through the mux register, through the SCADA driver, to the tag on the screen, and confirm the engineering value and units match the transmitter's own display.
5. Force a fault the device reports (a sensor test, or a deliberately stale reading) and confirm it shows up as bad quality and a maintenance alarm, not a good value.

Do that once, carefully, and the rest of the instruments are copy-and-verify. Skip it, and you end up with a wall of density and temperature tags that look plausible, read "good," and are quietly lying — which is the one outcome worse than never wiring HART at all.