Key takeaways
- HART protocol explained: how digital signalling rides on a 4-20 mA loop, dynamic variables, universal commands, burst mode, multidrop, and where WirelessHART fits.
- Focus protocol: HART — browse related articles and references on the topic page.
- Related topics: 4-20 mA, dynamic variables, burst mode, FieldComm Group.
The 4-20 mA loop normally represents one primary process variable. It has done so since the 1950s, it is understood by every technician alive, and it will outlive most of the protocols competing to replace it.
HART's insight was not to replace it. HART puts a digital signal on top of the analog loop, so the same two wires carry the measurement your control system has always read, plus everything the instrument can tell you about itself — additional variables, diagnostics, calibration data, a tag name, and its own opinion about whether it is working.
The analog signal keeps working exactly as before. That is why HART won: adopting it costs no rewiring and breaks nothing.
This guide covers how HART actually works, what the command set gives you, and where it fits today. The technical detail here is drawn from IEC 61158-5-20:2014, which defines the application layer service specification for Type 20 — the HART protocol — within the IEC fieldbus series. HART itself is maintained by the FieldComm Group, and command classification has evolved across HART revisions, so always check behaviour against the revision your devices implement.
The core idea: one loop, two signals
A HART instrument transmits two things simultaneously on the same pair.
The analog signal is the loop current — what the standard defines as the value measured by a milli-ammeter in series with the field device. It is a near-DC 4-20 mA signal communicating a single value between the control system and the field device. Your existing analog input card reads it and has no idea anything else is happening.
The digital signal is a frequency-shift-keyed tone superimposed on the current. Because it averages to zero over time, it does not disturb the analog reading. A device that does not understand HART simply doesn't notice it.
That coexistence is the whole trick, and it explains HART's install base: an instrument can be HART-capable while the plant continues treating it as a plain 4-20 mA transmitter until someone decides to use the extra capability.
Master, slave, and who is allowed to talk
HART is strictly request-response. The standard's definitions are precise:
A master initiates communication by sending a request PDU to another device and expecting a response frame. A slave initiates communication only after receiving a request, and is required to respond to it.
Field instruments are slaves. Control systems, handheld communicators, and asset management software are masters — and HART permits two masters on a loop simultaneously. That is the arrangement most process plants actually run: the control system as primary master polling continuously, while a technician connects a handheld as secondary master without disrupting anything.
A broadcast address also exists, letting a master address a command to all devices on the loop.
Dynamic variables: more than one number per instrument
This is where HART starts paying for itself.
A device variable is a uniquely defined data item within a field device, always associated with process-related information — its value varies in response to the process. An instrument may have many.
HART also defines four dynamic variables — primary, secondary, tertiary and quaternary — and the primary is the one tied to the analog output:
| Variable | Common role |
|---|---|
| Primary (PV) | The measurement on the 4-20 mA loop |
| Secondary (SV) | A second process value |
| Tertiary (TV) | A third |
| Quaternary (QV) | A fourth |
A Coriolis flowmeter is the classic illustration: mass flow on the analog loop as PV, with density, temperature, and volumetric flow available digitally as SV, TV, and QV. One instrument, one pair of wires, four measurements — three of which the analog system cannot see at all.
How device variables map onto the dynamic variables, and whether that mapping can be changed in the field, depends on the HART revision and the manufacturer's implementation. Some devices expose a fixed assignment; others allow reassignment through the dynamic variable assignment commands. Check the device manual rather than assuming.
The standard's read services reflect this directly. Read dynamic variables and loop current returns the loop current together with the dynamic variables the device supports, each as a floating-point value paired with its engineering unit.
Read device variables with status goes further: it reads up to eight device or dynamic variables in one transaction, and each value arrives with its own status alongside a variable class and engineering unit. That is the difference between a number and a number you can trust.
The universal command set
HART commands come in three tiers: universal, which conforming devices are expected to support; common practice, widely implemented but optional; and device-specific, the manufacturer's own.
That tiering is what lets a HART master talk usefully to an unfamiliar instrument on first connection. One caveat before the list: command availability and classification have evolved across HART revisions, and some services below are common practice rather than universal in a given revision. Treat the list as the shape of what HART offers, and confirm the specific commands against the revision your devices implement.
The services defined in the standard include:
Reading process data
- Read primary variable
- Read loop current and percent of range
- Read dynamic variables and loop current
- Read device variables with status
Identity and configuration
- Read device information
- Read tag, descriptor and date, and the write equivalent
- Read long tag and write long tag
- Read final assembly number and its write equivalent
- Read and write message
- Read primary variable transducer information
Loop and range configuration
- Read loop configuration and write loop configuration
- Write primary variable range values
- Write primary variable unit
- Read and write dynamic variable assignment
Calibration and maintenance
- Trim loop current zero
- Trim loop current gain
- Enter/exit fixed current mode
- Perform self test
- Perform device reset
Diagnostics
- Read additional device status
- Reset more status available
- Reset configuration changed flag
Three of those deserve comment because they are what technicians actually use.
Enter/exit fixed current mode forces the loop to a commanded value regardless of the process. This is how you verify a loop end to end — drive 12 mA and confirm the control system sees 50% — without touching the process or a calibrator.
Trim loop current zero and gain calibrate the device's own D/A conversion against a reference meter, correcting drift between what the instrument intends to output and what it actually outputs.
Reset configuration changed flag exists because HART devices remember that someone changed their configuration. That flag is a small audit trail, and clearing it deliberately is different from never having noticed.
Every response carries status
Every HART response carries status information alongside the data, and it is more valuable than most users realize. Two distinct things are reported.
The response code reports on the command itself — success, a warning (for instance, a value was set to its nearest legal value rather than exactly what was asked), or an error indicating the command was not executed properly.
Alongside it, device status reports on the instrument's own health, independently of whether the command worked.
That separation is the point. A command can succeed perfectly against a device that is failing, and HART tells you both facts in the same response. "More status available" is the device's way of saying there is detail worth fetching with Read additional device status — which is where a modern diagnostic instrument puts its real findings.
This diagnostic model provides an important foundation for condition monitoring and predictive maintenance. The instrument reports its health alongside the data it returns — whether or not anyone is reading it — and the detailed findings are there to be fetched when the summary says there is more to see.
Burst mode: when the slave talks first
Standard HART is polled, and polling is slow — a handful of transactions per second on a loop.
Burst mode changes the model. The standard defines it as initiation of communication activity by a slave device at a cyclic interval without request from a master. The underlying service permits a slave to send a variable's value, and the master to receive it, without a request — publishing data cyclically.
The gain is efficiency. Instead of a request and a response for every update, the device simply publishes, removing the master's transaction overhead and giving more regular updates. How much faster it works out in practice depends on the implementation and the loop's timing — burst mode changes the communication model rather than guaranteeing a particular update rate.
The cost is that a bursting device is occupying the loop. Configure it, don't just enable it — and be aware that burst configuration is a common cause of "the handheld can't talk to this transmitter" complaints.
Multidrop: several devices, one pair
HART also supports multidrop, where multiple instruments share one wire pair, each with its own polling address.
The trade is worth stating plainly: in HART multidrop, devices operate at a fixed current, typically 4 mA, and process data is communicated digitally. You gain wiring economy and give up the individual 4-20 mA analog reading from each device, along with its response speed.
That makes multidrop a good fit for slow-changing measurements where cable cost dominates — tank farms, remote monitoring points. It is generally unsuitable wherever a conventional individual 4-20 mA signal is required, which is why most HART installations in control service remain point-to-point.
WirelessHART
The standard references IEC 62591 (WirelessHART) as the wireless member of the family. It preserves the HART command set and device model while using a wireless 2.4 GHz time-synchronized mesh network — communication is scheduled in time slots rather than contended for, which is what makes it deterministic enough for industrial use.
The strategic value is continuity: the commands, the device variables, and the status model are what field engineers already know. WirelessHART typically serves monitoring points where running cable is disproportionate to the value of the measurement, rather than replacing wired loops in control service.
Where HART fits today
HART's strength is that it is additive. Existing 4-20 mA infrastructure keeps working; the digital layer is a capability you can adopt gradually, per instrument, without a project. There are more HART-capable devices installed worldwide than any other digital process protocol — and a substantial share of them have never had their digital side used, which is its own kind of opportunity.
HART's limit is speed. Polled digital communication over an analog loop is slow by any modern standard, and it was never meant to be a control network. For fast digital process control, PROFIBUS PA, FOUNDATION Fieldbus, and increasingly Ethernet-APL are the alternatives.
The honest positioning: HART is not competing with fieldbus for control duty. It is the pragmatic middle ground between pure analog and full digital — and for asset management, diagnostics, and commissioning, it delivers most of fieldbus's practical benefit at none of its infrastructure cost.
FAQ
Does HART replace the 4-20 mA signal?
No — that is the entire design. The digital signal is superimposed on the analog loop, and the 4-20 mA reading continues to work unchanged. The one exception is multidrop mode, where the analog signal is deliberately abandoned so several devices can share a pair.
What are HART dynamic variables?
Up to four device variables assigned as primary, secondary, tertiary and quaternary. The primary variable is what the 4-20 mA loop carries; the others are available only digitally. A single instrument can therefore report several measurements over one wire pair.
Can two masters be on the same HART loop?
Yes. HART supports a primary and a secondary master, which is why a technician can connect a handheld communicator to a live loop without disturbing the control system's polling.
What is HART burst mode?
A mode where the slave device publishes data cyclically without waiting for a master's request. It raises update rates well above polled operation, at the cost of the device occupying the loop with its transmissions.
Do I need special wiring for HART?
No. HART works over standard instrument cable — the same twisted pair the 4-20 mA loop already uses. The requirement is loop resistance within the range the device specifies, and a HART-capable master or modem to hear the digital signal.
What are HART universal commands?
The set every HART device must implement — reading the primary variable, loop current and percent of range, dynamic variables, device identity, tag, and status. They guarantee that any HART master can communicate usefully with any HART device without device-specific knowledge.
How do I verify a loop with HART?
Use the enter/exit fixed current mode command to force the loop to a known value, then confirm the control system reads it correctly. It tests the whole path without disturbing the process, and the device returns to normal operation when you exit.
Is WirelessHART the same protocol?
It shares HART's command set and device model, standardized as IEC 62591, but uses a wireless 2.4 GHz time-synchronized mesh network instead of the wired loop. Field engineers see the same commands and variables; the transport underneath is entirely different.