S7-400H PC Fault-Tolerant Communication: RedConnect Setup

Overview

The SIMATIC S7-400H is Siemens' high-availability fault-tolerant PLC platform based on two redundant CPUs (CPU 41x-5H) running event-synchronized over fiber-optic or PROFIBUS links. PC stations (programming devices, operator stations, or SCADA servers) require a separate, properly engineered fault-tolerant network path so that a single cable, switch, or NIC failure does not interrupt the HMI / engineering link to either CPU. This reference documents the network topologies that do not work with the integrated PROFINET/Industrial Ethernet interfaces of CPU 41x-5H, the architectural reason, and the supported RedConnect + ring topology that delivers a fully redundant PG/PC connection.

Reference documentation: SIMATIC S7-400H Fault-Tolerant Systems Manual (Siemens Support, entry ID 82478488) and the RedConnect commissioning guide at Siemens Support entry 109476178.

Problem Description

When a programming device attempts to download a STEP 7 project to an S7-400H station, the following symptoms indicate a fault-tolerant communication mis-configuration:

• STEP 7 / TIA Portal reports "Configured address cannot be reached" (event ID 0x0311 / 33:11 hex in the diagnostic buffer).
• The "Accessible Nodes" browser shows only one CPU, or no nodes at all, even though both CPUs are in RUN.
• Interrupting one PG cable causes the online view to drop, although the secondary CPU is still in RUN and reachable via a second switch port.
• The NetPro view shows the S7 connection as established, but the application connection (e.g., S7-Put/Get, OPC, WinCC) is unbound or in error.

Root Cause Analysis

The CPU 41x-5H PROFINET / Industrial Ethernet interfaces are not "teaming" ports in the Microsoft / Intel NIC sense. The two physical ports of an H-CPU belong to two different network domains:

If a single PC is wired with two NICs, one to each H-CPU and a third connection to a shared switch, the STEP 7 Online Gateway cannot bind the connection to a logical "fault-tolerant" partner because the H-system has no knowledge that the two PG paths belong to the same engineering station. The result is a single, non-redundant S7 connection, and the secondary link is ignored. This is the root cause of the failure in Case 1 in the field report.

Network Topologies Compared

Three PC-to-H wiring patterns are typically evaluated. Only one of them satisfies fault-tolerant communication with a single PC.

Solution: RedConnect-Based Architecture

Siemens achieves redundant PC communication for S7-400H through two cooperating mechanisms:

1. Media Redundancy Protocol (MRP) - Defined in IEC 62439-2. One switch in the ring is configured as the MRP manager. If a ring segment breaks, MRP reconfigures the ring into a line within ≤ 200 ms, so all nodes remain reachable on a single physical path.
2. RedConnect - A software function in STEP 7 / NetPro (S7-REDCONNECT) that runs in the PC station and the H-CPU. It maintains two parallel S7 connections (one to CPU-A, one to CPU-B). The active CPU of the H-system is the connection partner; on failure, the passive CPU takes over and RedConnect re-binds the S7 connection without an application-level break.

RedConnect requires:

• A valid S7-REDCONNECT license on the PC station (per PG/PC, not per project). Order number 6ES7 862-0AC01-0YA0 (S7-REDCONNECT V6.x floating license) or the bundle that ships with STEP 7 Professional.
• Both CPU PROFINET ports and both PC NICs assigned to the same IP subnet (or two subnets with proper routing, but single subnet is strongly preferred for the H-PG link).
• A managed switch capable of MRP, e.g. SCALANCE XC-200, XC-300, XR-300, or any PROFINET-conformant managed switch listed in the S7-400H manual.

Prerequisites

Step-by-Step Configuration

1. Assign IP Addresses and Device Names

1. Open STEP 7 / TIA Portal and load the H-station hardware configuration (HW Config).
2. Set the PROFINET interface X1 of each H-CPU to a unique IP in the same subnet, e.g. CPU-A = 192.168.10.11, CPU-B = 192.168.10.12, subnet mask 255.255.255.0.
3. Assign PROFINET device names to each CPU interface: h-cpu-a and h-cpu-b.

2. Build the Redundant Ring (MRP)

1. Connect SW1 and SW2 in a ring; close the ring with the cable between SW1 and SW2.
2. In the switch web interface (e.g. SCALANCE Web Based Management), enable MRP and select one switch as the MRP manager. The other becomes an MRP client.
3. Verify that all ring ports show "MRP active" / "Forwarding". The MRP manager port on the ring is in "Blocking" state during normal operation, isolating the ring into a line at the management point.

3. Configure the PC Station in NetPro

1. In NetPro, add a PC station with two IE interfaces (e.g. IE-PG-1 on NIC1 and IE-PG-2 on NIC2). Assign IP addresses 192.168.10.21 and 192.168.10.22.
2. Set the S7-REDCONNECT mode on the PC station: Properties > S7-REDCONNECT > Activate. Specify the redundant partner CPUs (CPU-A and CPU-B IP / name).
3. Download the PC station configuration to the local PC. The S7-REDCONNECT service S7RedCoSrv.exe starts and binds to both NICs.

4. Configure the S7 Connection as "Fault-Tolerant"

1. In NetPro, on the PC station, insert a new S7 connection to the H-station.
2. Select Connection type: S7 connection fault-tolerant (German: S7-Verbindung fehlertolerant). Specify the H-station as the partner, not an individual CPU. STEP 7 binds the connection to the active CPU and the RedConnect runtime handles the fail-over.
3. Compile and download the NetPro configuration. The connection appears with a yellow icon until the first handshake completes.

5. Verify the Service is Running

On the PC, open a command prompt and confirm the RedConnect service is active:

The state should be RUNNING. Logs are written to %ProgramData%\Siemens\Automation\S7REDCONNECT\log.

Verification

After commissioning, perform these checks in order:

1. Online accessibility - In STEP 7 "Accessible Nodes", both CPUs (192.168.10.11 and 192.168.10.12) must appear. Right-click the H-station and select Go Online > Fault-tolerant system; the online view should show the active CPU in green.
2. Connection status - In NetPro, the S7 connection icon turns green. The diagnostic buffer of the active CPU contains entry "Establishment of S7 connection to PC-station (connection fault-tolerant)".
3. MRP ring test - Pull one of the ring cables. The MRP manager should log a topology change, the ring is reconfigured in < 200 ms, and the S7 connection remains established. Restore the cable and confirm the ring returns to its original state.
4. CPU fail-over test - Stop CPU-A. The H-system promotes CPU-B to master within < 100 ms. The S7 connection on the PC remains active; SFB 52 / SFB 54 reads continue without an error code (S7 connection status word bit "partner CPU changed" is set).
5. PC NIC fail-over - Disable NIC1 on the PC. RedConnect switches the active path to NIC2; the S7 connection stays up, the SCADA/PG application sees no break.

Specifications

Diagnostic Blocks and Error Codes

The H-system provides standard blocks for diagnosing fault-tolerant PC communication:

• SFB 52 RDREC - read record data (e.g. PROFINET port statistics from CPU 41x-5H PN).
• SFB 54 RALRM - receive alarm (used to detect link-down and ring-state-change interrupts).
• FC 100 / FB 100 - status of the H-system in user program (similar to example project "H-Status").

Common Mistakes and Caveats

• Two PC NICs on different subnets - The H-CPU cannot perform layer-3 routing for the S7 connection; both NICs must be in the same subnet as the H-PROFINET interface, or in a topology where MRP can still keep both paths alive.
• Unmanaged switches in the ring - MRP requires a manager. Plugging an unmanaged switch into a closed ring creates a broadcast storm and the ring never converges.
• Mixing PROFINET RT and IRT on the same ring - PROFINET IRT traffic (e.g. for distributed I/O with isochronous mode) requires a cut-through switch and a separate ring; do not combine with the PG/PC ring.
• PC firewall blocking port 102 - Windows Defender or third-party firewalls must allow TCP/UDP 102 (ISO-TSAP) for both NICs inbound and outbound.
• One NIC bound to VMware / Hyper-V virtual switch - Virtual bridging changes MAC timing and breaks the redundant NIC detection. Use physical NICs in the H-PG ring, not virtual switches.

Notes for Migration from STEP 7 V5 to TIA Portal

When migrating an H-station with PC fault-tolerant communication to TIA Portal V16+:

• The TIA Portal "H-station" device proxy replaces the classic NetPro H-station, but the S7 connection must still be configured as "fault-tolerant" (right-click connection > Properties > Type). The S7-REDCONNECT service is identical and the license carries over.
• PROFINET device names become mandatory in TIA Portal; the H-CPU's two interfaces must each have a unique PROFINET name even though they share the same IP subnet.
• Use the TIA Portal "Online > Accessible devices" filter "H-systems" to confirm both CPUs are reachable before commissioning.

FAQ