|1||Physical||CAT5 oder CAT6|
With Ethernet technology, the layer is unaware of the protocols and connections of the layers above or below. The term Ethernet has been used in networking for many years, for 10BaseT (10 Mb/s), 100BaseT (100 Mb/s) and now for Gigabit Ethernet (1000 Mb/s). This refers to a defined combination of Layer 1 (physical interface) and Layer 2 (address mechanism) on a network. Gigabit Ethernet uses CAT5e or CAT6 cables to reach distances of up to 100 m over copper, or up to several kilometres using fibre.
Although Ethernet has been available for some time, the only way to transmit video was to compress the data, as usually happens on multimedia websites and with security cameras. With Gigabit Ethernet providing data rates up to 1000 Mb/s, it is now possible for machine vision to take advantage of the long transmission lengths and low cost cabling to achieve the data flow required for uncompressed video transmission.
Because the Ethernet stack isolates each layer, the technology is future proof. We already have 10Gigabit Ethernet available across fibre optic links and copper and it is only a matter of time before this technology will be available in products, allowing Ethernet to theoretically deliver up to 1 GB/s, enough to exceed the capability of full configuration CameraLink, however, for now the power consumption of the 10GigE chip-sets is limiting its adoption.
Another benefit of using standard Ethernet technology is the availability of 'off-the-shelf' industrial type Ethernet switches and routers. These components have been available for a long time and are more robust than consumer components. Additionally, Ethernet can encapsulate other transmission technologies like wireless or UMTS.
Gigabit Ethernet topological models
Gigabit Ethernet expands the potential for creating different implementation models and complex topologies that were previously not possible. In a factory inspection scenario, it is for example possible to construct an inspection 'network', split-up into different zones, each of which having a specific function, but all controlled via a single workstation, as shown in the graphics (figure 1). Another option is to use existing network infrastructures to transmit data from remote locations to any number of different workstations. As each camera is located independently on a network via its IP address, every camera can be viewed, controlled and monitored from any PC on the network (figure 2).
In a high-end web inspection application as shown in the next figure, the use of multiple cameras often requires very complex cabling. By using Gigabit Ethernet, the transmission, processing and analysis of data can be handled more easily and elegantly, and the control PC does not have to be physically close to the actual inspection. The control of the inspection process is not even limited to one location, as it is possible to have a number of PCs that can all perform this function. The process could even be controlled via the internet from a completely different geographical area. This allows remote system fault diagnosis and hardware support. Ethernet's multicast ability opens up completely new types of applications where the data is transmitted to a number of PCs. This allows scalable solutions with multiple processing nodes (PCs) working on the same data stream. Ethernet structures are also good for 'failover' scenarios where the data can be easily switched to another processing system, if the first one fails.
Ethernet's multicast ability opens up completely new types of applications where the data is transmitted to a number of PCs. This allows scalable solutions with multiple processing nodes (PCs) working on the same data stream. Ethernet structures are also good for 'failover' scenarios where the data can be easily switched to another processing system, if the first one fails.
Ethernet's ability to both send and receive the data using the same technique, allows the creation of "processing pipelines" using multiple PC set-ups, in which each PC performs a separate processing step. For example, one PC for video recording and the second PC for segmentation and compression.
For vision applications a range of specialist cabling options are available, including both locking and angled connectors and robotic grade cables. Please see the cable technology section for more detailed information.