Unlike office and home Ethernet, which has long been known, industrial Ethernet requires more professional knowledge and practical experience. If you are installing or using an industrial Ethernet, the five points about wiring, signal quality, grounding loops, switches and communications must be understood.
Like all networks, the pros and cons of cable directly affect the pros and cons of industrial Ethernet. In addition to high electromagnetic interference (EMI), the industrial environment also often has a certain level of temperature, dust, humidity and other in the home and office environment is not common impact factors.
So, how do you choose a cable? In the office, commercial grade cables, such as Category 5 cables, are more suitable for 10MB networks, while 5e-type cables are suitable for 100MB networks. According to the ANSI/TIA-1005 Standard, a 6 cable or better cable can be used for a host or device connection in an industrial environment. 6-Class cable can be 100 meters in the range of 1GB network, 55 meters in the range of 10GB network implementation. 6e cable can achieve 10GB network within 100 meters.
Compared to Category 5 cable and 5e cable, 6 kinds of cables are not susceptible to crosstalk and external EMI noise. Industrial Ethernet cables are designed to withstand the physical erosion of cables by a more severe industrial environment. When installing type 6 cables, ensure that the RJ45 connectors and sockets are also up to 6 classes. The best way to do this is to use a predetermined plug cable and install a connector in the factory for short distances. Use sockets for long-distance wiring.
Some applications need to be shielded, but if the shielding cable is improperly installed, it will backfire.
Shielded Ethernet cables perform better in an EMI environment when the casing is exceeded. Good grounding is the key to the use of shielded cable. A grounded reference point is key. A plurality of grounding connections will form a grounding circuit, and the difference of electric potential between different grounding connections will introduce noise into the cable.
The grounding circuit will cause great damage to your network, in order to solve this problem, only at one end of the cable using the grounding RJ45 interface, the other end of the use of insulated RJ45 interface to eliminate the possibility of grounding circuit.
If the ethernet cable crosses the wiring with the power cable, then the intersection angle is quite fastidious. A parallel ethernet cable and power cable are separated by at least 8 to 12 inches, and if the voltage is higher or the tie distance is longer, the interval should be greater. If an Ethernet cable is in a metal groove or casing, the adjacent grooves or casing must be connected to achieve electrical continuity.
Generally speaking, Ethernet cable as far as possible to produce EMI equipment, such as motor, motor control equipment, lighting equipment, live conductor and so on. On the panel, the ethernet cable is at least 2 inches apart from the connector. The recommended cable bending radius is followed when the cable is away from EMI sources.
Simply put, do not use hubs in an industrial Ethernet environment. A hub is nothing more than a repeater of a multiport port. If the hub is excluded, the only options left are managed switches and unmanaged switches. A managed switch is better, and of course its price is more expensive than a unmanaged switch.
Each device on the network has a unique identifier, which is what we call the MAC address, which is the key to the switch's ability to identify better than the hub. When the switch is just on the power, its initial performance is no different from the hub, it broadcasts all communications, but as the device on the network transmits the information to different ports on the switch, the switch begins to monitor the content of the communication and identify which MAC address is associated with which port. Then make the identification in the MAC address table.
Once the switch discovers that the device's MAC address is connected to a particular port, it monitors the information that points to that MAC address, and then sends that information only to that particular address.
There are three types of communication in industrial Ethernet networks. Point-to-Point unicast communication, one-to-many multicast traffic, and broadcast traffic to all nodes.
When the MAC Address table for the switch is established, the management switch and the unmanaged switch have no difference in how the unicast and broadcast traffic is handled. In general, under 100MB bandwidth, broadcast frequency is controlled at 100 broadcasts per second. For any network, there will be more or less broadcast traffic. One example is that the print server periodically gives broadcast notifications on the network.
One of the main differences between managed and unmanaged switches is how they handle multicast traffic. Multicast communications typically come from intelligent devices that are connected to a factory process network, using a connection-oriented technology based on the manufacturer/user model. In this case, the connection is only the relationship between two or more nodes on the network.
To be able to receive the information in the group, the device must join the multicast communication group, and all members of the group can receive the data. If you're just sending data to a group, you don't need to be a team member. In the manufacturer/user model, the main problem of multicast communication is the exponential growth of communication information as the number of team members increases. At this point, you need to use a managed switch.
The managed switch is able to open the Internet Group Management Protocol (IGMP) snooping feature. It works like this, and when the IGMP snooping feature is turned on, it emits broadcast traffic to determine the members of any multicast group. Using this information, coupled with the already built MAC address table, the managed switch is able to send multicast traffic only to members of the multicast group. unmanaged switches treat multicast data and broadcast data the same way that data is sent to each node.
If the network uses manufacturer/user technology or multicast communication, then the management switch is the choice for value for money.
There are many other reasons to consider using a managed switch, which typically provides fault logging, control of the speed of each port, redundancy settings, and port mirroring. These additional capabilities ensure more precise control over network behavior and can play a valuable role in troubleshooting. We know that for some nodes on the network, failures are unavoidable.
When network performance problems occur, first check the switch, although for most network performance problems, the switch is rarely the core of the problem. The switch is the most likely problem node in the system, and its operating rate is typically 10 to 50 times times the rate of other network parts working.
Although there is always a good software to help you troubleshoot network problems, most of the software can only see broadcast and multicast communications. This is actually reasonable because many performance problems usually originate from unrestricted multicast traffic or too much broadcast traffic. If you need to check unicast traffic for some reason, port mirroring is the only way.
If there is no multicast traffic on the network, then the use of unmanaged switches is no problem. On a small, simple network with few devices, many people use unmanaged switches.
Sometimes these two types of switches can be used in combination, with some remote devices on unmanaged switches and unified feedback to managed switches. For networks with a lot of nodes, if cost is not a key factor, then it's a good idea to choose a managed switch.