PROFIBUS and PROFINET: Differences and Migration

Short answer

PROFIBUS and PROFINET are two technologies from the same organization, PROFIBUS & PROFINET International (PI). PROFIBUS is a serial bus over RS-485 twisted pair, where a master polls slaves at a single speed for the whole network. PROFINET uses standard Ethernet and keeps the idea of a controller governing field devices, but with free topology, network diagnostics, and the ability to share the cable with other traffic. A PLC with PROFINET can incorporate existing PROFIBUS lines through a proxy.

PROFIBUS: the essentials

According to the public description of PROFIBUS, DP operates between 9.6 kbit/s and 12 Mbit/s, with a single speed for the whole network. An RS-485 segment supports a limited number of stations (32 per segment, counting repeaters) and is extended with repeaters up to 126 stations in total; the length between repeaters ranges from 100 to 1200 m depending on the speed. PA uses a fixed speed of 31.25 kbit/s and, besides data, powers the instruments over the same pair, with limited current for classified areas.

Bus access combines token passing between masters with polling of slaves by each master. Each slave is described by a GSD file that the engineering tool imports.

Typical PROFIBUS problems are physical: missing or duplicated termination, long stubs, poorly connected grounds and shields, a speed too high for the length, and cable that does not meet the expected impedance. Diagnosis requires a bus analyzer or the master’s statistics.

PROFINET: the essentials

PROFINET IO defines three roles:

Devices are described with a GSDML file (XML). Discovery and the initial assignment of name and address are done with the DCP protocol, and neighbors are detected with LLDP. That way the system knows the physical topology and, in some engineering schemes, a replaced device receives its name without manual intervention.

Real time and conformance classes

PROFINET divides communication into channels: cyclic process-data traffic is served with priority over configuration traffic or that of other protocols sharing the cable. According to the public description of the technology:

Most manufacturing and process applications are solved with RT. The isochronous variant is reserved for coordinated motion.

Practical comparison

Illustrative example of staged migration

The data below are illustrative. A plant has an old PLC with a PROFIBUS line that governs a set of remote modules and drives in a motor control center. The PLC is replaced with one that has a PROFINET interface.

1. A proxy is installed between PROFINET and the PROFIBUS line. The PLC sees the old slaves as devices on its network.
2. The slaves’ GSD files are imported, the program’s process addresses are reassigned, and each signal is verified.
3. When a remote module or a drive reaches the end of its life, it is replaced with a native PROFINET one and removed from the proxy.
4. In the end, the PROFIBUS line is empty and is retired.

The advantage of this path is that the plant does not stop all at once. The risk is in cycle times: the proxy adds the PROFIBUS line cycle, which depends on its speed and number of slaves, and that latency must be acceptable for the process.

Cabling and infrastructure

• PROFIBUS: specific 150 ohm cable with termination at both ends; no long stubs; shield connected according to the installation guide.
• PROFINET: industrial Ethernet cabling. Connectors, cable types, and network equipment must match the required conformance class and the environment (vibration, humidity, noise). If the network uses a ring, the switches must support the redundancy mechanism the design defines. Topologies and redundancy with switches are covered in the guide on RSTP and Ethernet redundancy and in network redundancy in substations.
• Network separation: the control network must not share equipment with the administrative network without a segmentation design; see IEC 62443 in substations.

Diagnostics

In PROFIBUS, the master’s error and retry counters and the electrical signal quality point to cable or termination problems. In PROFINET, device diagnostics (module alarms, channel with a fault), topology readout with LLDP, and the switch port statistics point to the link or the device with a problem. A device with a duplicate or misspelled name is a common cause of the controller not recognizing it after a replacement.

Common mistakes

• Duplicating or removing termination on a PROFIBUS segment.
• Choosing a PROFIBUS speed without checking the real cable length.
• Feeding the PROFINET network with office switches that lack the functions the class requires.
• Renaming a PROFINET device and not updating the engineering.
• Migrating without measuring the update cycle of the data crossing the proxy.
• Running communication cables right next to drive power conductors.
• Connecting the shield at one end in one part and at both in another, without a criterion.

Checklist

• The variant of each segment was identified (DP, PA, PROFINET RT or IRT).
• The GSD or GSDML files match the firmware version of each device.
• In PROFIBUS, termination is only at the two ends and the number of stations per segment was respected.
• The PROFIBUS speed matches the cable length.
• In PROFINET, the network components meet the design’s conformance class.
• Device names are unique and documented.
• The real topology matches the engineering one.
• The cycle time of critical signals was measured.
• There is a backup of the engineering and of the device parameters.
• The control network is segmented from the others.

When to keep PROFIBUS and when to migrate

Keeping a PROFIBUS line makes sense when it works without faults, the slaves are still available, and the process does not ask for more data or more diagnostics. Migrating is advisable when the controller is renewed, when slave spare parts are missing, when network diagnostics or remote access from engineering is needed, or when the plant already has an Ethernet infrastructure to which the devices can be added. In a new installation, the usual decision is PROFINET, unless a process instrument exists only in PA.

A well-planned migration starts with an inventory: each slave with its part number, version, GSD file, and the signals it delivers, and each signal with the point in the program that uses it. With that inventory you decide which devices are replaced, which are kept behind the proxy, and in which plant outages each change is made. It is also worth leaving a written functional test for each signal, to check after the change that the value arrives the same.

How we help

Choosing between keeping PROFIBUS, migrating in stages, or moving to PROFINET, and defining the network architecture that supports it, is engineering work. At Tiza Group we do it within the engineering design and studies service and check the result with testing and commissioning.

Related guides

• Network redundancy in substations
• IEC 61850, IEC 104, DNP3, and Modbus: which to use
• Drive communication: Modbus, PROFINET, and EtherNet/IP
• Industrial network topologies and cabling
• EtherNet/IP and CIP: messages, connections, and design
• Modbus: register map, data types, and polling
• Substation Ethernet Switch: IEC 61850-3, VLAN, Redundancy
• Modbus RTU and Modbus TCP: registers, functions, limits
• PLCs in Substations: Uses, IEC 61131-3 and Testing

References

• PROFIBUS & PROFINET International, PROFINET Technology and Application, system description (manufacturer system description).
• Wikipedia, “Profibus” and “Profinet” (variants, speeds, roles, and conformance classes).
• IEC 61158 and IEC 61784, fieldbus networks (paid standards).

Frequently asked questions