Ring redundancy and hot plugging

A network in a ring topology which is safe from interference is still operative when a cable breaks or when slaves are connected or disconnected (hot plugging) with no interruption. Because of the ring network structure, all devices remain connected to the master when a cable breaks. The Sercos connections on the slaves at the breakage point switch to two separate lines (line topology) with “loop back” without causing any delay. The reconfiguration time takes a maximum of 25μs so that at the very most the data transmitted from one cycle is lost. The breakage point can be precisely located and defective cables can be simply exchanged while the operation is running.

Since communication continues when there is a cable break, if necessary, new devices or groups of devices can be connected to the network while it is in operation, and can be integrated into the communication (so-called hot plugging or hot swapping). The system can merge seamlessly and without interference into the configured operation.

Ring redundancy and hot plugging

A Sercos hardware failure or break in a cable does not result in a breakdown in communication. Instead, we are able to replace defective devices or cables while the system is in operation, which significantly increases machine availability.

Redundancy

The operational and failure safety of machines and systems is playing an increasingly important role in the economically efficient and safe operation of industrial production sites. Here, redundancy concepts ensure that a system malfunction or system downtime can be avoided. It is important for troubleshooting and repair that a system’s ability to communicate is preserved even when a device breaks down or in the event of a cable break. There are a large number of different redundancy processes for bus systems that differ significantly with respect to technical characteristics, their installation effort and installation costs.

Reconfiguration time, or switch-over time, is an important technical parameter of redundancy processes. For non-bumpless solutions, a certain amount of reconfiguration time is required in the event of a communication network failure in order to convert to the redundant path. Typical reconfiguration times lie in the range of 200 to 500 ms. On the other hand, the data flow is maintained without reconfiguration time in the case of a bumpless process.The parallel redundancy protocol (PRP) was established in IEC standard 62439-3. It is a redundancy protocol for Ethernet networks that works on layer 2 and is transparent to higher level protocols. The parallel redundancy protocol requires the use of two independent, parallel networks. End devices that use this process are connected to both networks via two ports of the device interface. The ports appear with the same MAC and IP address in the various networks. The telegrams are sent in duplicate to the relevant communication partners and separately on both redundant networks. If both networks are intact, the first package to arrive is assessed and the redundant package is discarded. If not, the redundant package is assessed.In order to keep the amount of cabling and infrastructure costs low, a redundant network is usually not used; instead – in the event of an error – an alternative path is connected in the existing network structure. Here, either proprietary concepts or standardized processes such as Spanning Tree (STP, IEEE 802.1D) or Rapid Spanning Tree (RSTP, IEEE 802.1w) are used. The medium redundancy protocol (MRP) described in IEC standard 62439-3 is based on a ring topology and ensures reconfiguration times of between 200 ms and 500 ms, depending on the configuration. MRP uses a redundancy manager (RM or MRM) that closes the ring. In normal operations, the redundancy manager tests the consistency of the ring using special test packages. However, it does not forward any packages and thus prevents them from circulating endlessly in the ring. If the test packages do not come through, the redundancy manager prompts the reorganization of the network and activates the reserve path. Sercos III also can use a ring topology in order to maintain the ability to communicate in the event of an error (cable break or breakdown of a device). The unique feature of the ring-redundancy process is that the real-time packages are already transferred redundantly, i.e. on two separate paths (primary and secondary channel) in normal operations. Thus, fast and bumpless conversion of the data paths is ensured, without reconfiguration being required. In the event of an error, the nodes positioned right at the location of the error merely switch from “Forward” to “Loopback” with the ring dividing into two lines. The switch-over time is a maximum of 25 µs, so that at most the data of one communication cycle can be lost. Because redundancy recognition is present in every node in the hardware, the patented process is simple, fast and safe.