Moving a running PROFIBUS DP line to PROFINET is not a cable replacement job; it is a change management project. If the line has been in production for years, there are undocumented addresses, parameters changed on site and temporary fixes known only to a single maintenance technician. These only come to light during the downtime window. Below, for automation and maintenance engineers who want to carry out the migration in phases without tying it to a single long shutdown, we describe the inventory, choice of approach, use of couplers and proxies, GSD/GSDML management and the rollback plan in working order.
Situations that really call for migration
If PROFIBUS DP is working, merely being “old” is not a reason to migrate. The situations that make the decision concrete are usually these:
- Weakening spare parts and support status for the controller or DP interface module (to be verified on the manufacturer’s lifecycle page)
- New machines and drives arriving with a PROFINET interface only
- A need to get more data and diagnostic information from the line into higher-level systems
- Recurring physical layer faults: intermittent dropouts caused by termination, connectors and cables
- Approaching the limits of segment length, station count or baud rate when extending the line
Once you put the reason in writing, the approach also becomes clear. If the only problem is spare parts, replacing the controller and using a proxy may be enough; if data and diagnostic needs dominate, converting the field devices as well comes onto the agenda.
Inventory and diagnosis of the existing PROFIBUS line
The inventory is a comparison of the configuration in the engineering project with the actual situation on site. Differences between the two are the rule, not the exception: a station that is not in the project, a drive running with a different address on site, a changed GSD version. The following information is collected for each station:
| Details | Source | Why it’s needed |
|---|---|---|
| DP address and device type | Project and device label | Basis of the mapping table |
| GSD file and version | Project archive | Parameters and module structure |
| I/O address ranges | Hardware configuration | References in the PLC program |
| PROFINET equivalent | Manufacturer’s product catalogue | Replacement or proxy decision |
| Segment, baud rate, termination | On-site check | Physical health during the hybrid period |
Don’t skip the physical diagnosis; during the hybrid period, part of the DP line will keep running for months. PROFIBUS DP allows at most 32 stations per segment without repeaters; segment length depends on baud rate (e.g. 1200 m up to 93.75 kbit/s, 200 m at 1.5 Mbit/s, 100 m at 3–12 Mbit/s). Termination must be only at the two physical ends of the segment, and powered. A common DP fault is a termination switch left on in a connector in the middle. A migration carried out without measuring signal quality and error counters with a bus analyser or diagnostic repeater carries old problems over into the new system.
Migration approaches: full, hybrid and gradual
| Approach | How it’s done | Advantage | Limitation |
|---|---|---|---|
| Full migration | All DP devices are replaced with their PROFINET equivalents in a single shutdown | A clean architecture in one go | Long downtime, high risk |
| Hybrid with proxy/coupler | The controller moves to PROFINET, DP devices stay behind a proxy or coupler | Short downtime, field devices unchanged | Additional component, the DP line lives on |
| Gradual replacement | A hybrid structure is set up, and DP devices are replaced one by one at maintenance or failure opportunities | Cost spread over time | Two networks for a long time; disciplined documentation needed |
In most production plants, the realistic way to renew industrial network infrastructure is to start with a hybrid structure and move on to gradual replacement. Full migration only makes sense if the line will be stopped for a long time anyway for a major overhaul, or if most DP devices have reached the end of their life. Otherwise, fitting everything into a single shutdown leaves no time for rollback. Whether the existing DP segment can be retained during the hybrid period depends on the devices being supported by the proxy and on the physical condition of the segment.
Using DP/PN couplers and proxies
Products that connect the two networks fall into two groups according to their function. This difference determines your diagnostic capability throughout the project.
Proxy
A PROFINET IO proxy behaves as an IO device on the PROFINET side and as a DP master on the PROFIBUS side, and presents each DP slave behind it to the PROFINET controller as a separate device. The Siemens IE/PB LINK PN IO is an example of this type. The GSD-based parameters and diagnostic messages of DP devices are preserved per device; from the PLC program’s point of view, the devices can still be addressed individually.
Couplers and protocol gateways
A coupler exchanges mapped input/output data between the networks or projects of two separate controllers; each controller sees the coupler as a device on its own network. Siemens’ PN/PN coupler is an example that does this between two PROFINET networks; for products doing the same between PROFIBUS and PROFINET, the name and capabilities must be verified from the manufacturer’s documentation. The devices behind it are not visible to the other controller.
A protocol gateway, on the other hand, works as a PROFINET IO device on one side and as a DP master or DP slave on the other, and transfers data using a mapping defined by the manufacturer. Because it does not use the standard mapping PI defines for proxies, how much alarm and diagnostic information is carried varies by product; where needed, it is added to the data block separately.
If you want to manage DP devices from the new controller, consider a proxy; if two separate controllers need to share data while keeping their own networks, a coupler; if a device or protocol has no proxy support, a gateway. Diagnostic information from devices behind a coupler or gateway does not reach the new controller on its own.
Address, device name and GSD/GSDML management
In PROFIBUS, the device identity is the DP address, usually set with a switch or parameter on the device. In PROFINET, the controller finds the device by its device name (NameOfStation); the IP address is initially assigned by the controller based on this name. A mapping table must therefore be kept during the migration:
DP address | Device | PROFINET device name | IP address | Status
-----------+-------------------+------------------------+---------------+-----------------
DP 5 | Conveyor drive | h2-konv-g120-01 | 10.12.20.41 | Replaced
DP 12 | Remote I/O ET200 | (proxy: h2-pblink-01) | (10.12.20.50) | Behind proxy
DP 18 | Weighing module | - | - | Awaiting replacement
DP devices remaining behind a proxy do not have their own PROFINET name and IP address; the controller reaches them via the proxy’s name and address, and the DP address remains valid on the proxy’s DP side. That is why these rows in the table refer to the proxy. Setting naming rules from the start prevents corrections later; for this, you can use the template in the article IP addressing and device naming standards for factory networks.
GSD (PROFIBUS, text-based) and GSDML (PROFINET, XML-based) files should be stored in the project archive with their versions. Different GSDML versions of the same device may offer different module and parameter options; check that the version installed on the engineering station matches the firmware of the device on site. A mismatch is usually noticed on commissioning night, when the module list does not match.
On the PLC program side, the lowest-risk approach is to keep the I/O start addresses the same as the old values in the new hardware configuration. That way the hardware changes without touching the program logic. For drives, the PROFIdrive profile is the same on both networks; if the same standard telegram is selected, the control and status words on the application side do not change. For devices using PROFIsafe, however, F-addresses must be unique across the network, and F-monitoring times must be recalculated for the new cycle times.
Downtime window and rollback plan
The downtime window for each phase is planned not according to how long the work takes, but according to how long rollback takes. The last part of the window is reserved for returning to the old state if the change fails. The components of the plan:
- Taking backups of the PLC program, hardware configuration and drive parameters before the change, and testing the restore
- Labelling the removed DP cables and connectors and leaving them in place; no physical removal until acceptance is complete
- Setting a clear time limit for the rollback decision and naming the person with decision authority
- Writing down in advance the function tests to be carried out after a rollback
Trying the change in a test environment or a pilot cell before the window shortens the time spent on site. For the change approval process and test discipline, the article change window engineering in OT is a useful complement.
Migration checklist
- The reason for migration and the target architecture are defined in writing
- The inventory of all DP stations, GSD versions and I/O addresses has been compiled
- The physical diagnosis of the DP line has been carried out and existing faults fixed
- A replacement, proxy or coupler decision has been made for each device
- PROFINET device names and IP addresses have been assigned according to the naming standard
- GSDML files and firmware compatibility have been checked
- I/O address mapping and program references have been verified
- Validation has been planned for PROFIsafe and safety functions
- The PROFINET topology and switch configuration have been checked against the design rules
- Backups have been taken, and the rollback plan and decision time defined
- Error counters and the diagnostic buffer were monitored after commissioning
For topology and switch selection mistakes on the new PROFINET side, see the article common mistakes in PROFINET network design, and for problems after commissioning, PROFINET fault diagnosis steps. Our factory OT infrastructure survey service can help prepare the inventory and migration plan, and our PROFINET and PROFIBUS communication service during implementation. The current versions of the installation and planning guidelines are published on the official PI website.
Frequently asked questions
Can PROFIBUS and PROFINET work together in the same PLC program?
Yes. Many controllers have both a DP master and a PROFINET IO controller interface, or one is added with a communication module. During the hybrid period, the same CPU can manage both networks together. The point to watch is that I/O addresses and cycle times are verified separately for the two networks.
What is the main difference between a proxy and a gateway?
A proxy presents PROFIBUS devices on the PROFINET side as separate IO devices; diagnostics and parameters are preserved per device. A gateway or coupler transfers only a defined data block between the two networks; the PROFINET controller does not see the PROFIBUS devices behind it, and diagnostic information has to be carried separately.
What should be considered when migrating a line that uses PROFIsafe?
PROFIsafe works on both networks, but every change in the safety-related communication path requires safety validation. Unique F-addresses, recalculating F-monitoring times for the new cycle times, and verifying safety functions with actual testing must be part of the plan.
Do PROFIBUS PA devices also need to be replaced?
Not necessarily. PA segments connect to the DP side through DP/PA coupling elements; when the DP side is moved to PROFINET, PA devices can be retained with a suitable proxy or PROFINET-capable PA connection products. The decision to replace them is usually made separately, based on device lifetime and whether the process can be stopped.