Design of an Ethernet cable section

An EtherCAT network usually consists of a master device and up to 65535 slave devices. The master alone manages the slaves and can be replaced by a 2nd master if necessary in the case of redundancy.

The electrical communication between the devices can be accomplished

on an Ethernet basis as a point-to-point connection (discussed here) – recognizable by a cable connection
on an LVDS basis – ‘E-bus’ – between stackable modular devices; a cable connection is usually dispensed with here

When designing the EtherCAT network, the following must be taken into account with regard to the cycle time(s) used

the maximum number of devices (max. 65535)
the max. permissible cables length for Ethernet cabling between individual devices (see below).
Propagation delays due to the cable length are of secondary importance; 100 m Ethernet cables with approx. 550 ns can be used.
the cycle time of the Ethernet frame(s) through all real slaves on the outward and return path. The following can be assumed as an order of magnitude per slave:
- for an Ethernet device: approx. 1 µs
- for an E-bus device: approx. 300 ns
the sum of the Ethernet frame lengths with which the configured devices are addressed.

Hubs and switches retarding the cycle time (ISO Layer 2) are dispensed with entirely in an EtherCAT network; there is no segmenting by routers (ISO Layer 3).

	Design rules The sum of frame length and cycle time_total should be shorter than the available timeframe to the next transmission point. Detailed information about the current frame configuration is given, for example, by the Beckhoff system manager. Depending on the real-time quality and the scope of the process data, it may be necessary to remain well under the value specified above. Please take note of the comments on the Distributed Clocks settings!

The individual cable lengths for the Ethernet cabling can be designed as follows.

A Ethernet cable section is in each case a point-to-point connection between 2 intelligent end devices, of which the transmitter sends a newly generated frame to the receiver. FastEthernet/100Mbit generally works in duplex mode, hence both participants can receive and transmit at the same time on different cables. In terms of topology there are few differences between EtherCAT cabling in an industrial environment and cabling in an office. However, the number of patch bays or transition points between two Ethernet terminal devices is usually reduced in the field in favour of increased operational reliability, but at the cost of flexibility.

The boundary conditions relevant to the transmission performance are:

the environmental/operating temperature
the number of plug connections (single/double) between the end points
the types of cable used (rigid/flexible core or their electrical characteristics; data sheets should be available)
external influences on the cable section (electromagnetic fields, cables installed in parallel, garland or drag chain installation, chemical influences etc.)
100 m maximum total link length permissible for 100Base-TX Ethernet; note: depending on the cable material used, the link length attainable may remain well below this value or even above it – but the application may then have to cope with longer propagation delays
the design methods mentioned below refer to EN50173-1 and 3 and also apply to the industrial sector.
the installed cable section should be tested; the individual testing methods are described in prEN50346:2001

Theoretical implementation:

For information: For the ‘exact’ dimensioning of a cable section according to EN50173, equations are specified in the standard in accordance with the following sequence:

Sought: remaining permissible flexible cable section [m]
Given: known number of plug connections, known cable lengths of permanently installed cable section, technical data of all components used (in particular attenuation)

Two sections are to be used for this, depending on the topology:

Standard building cabling with up to 4 patch bays within the transmission link: EN50173-1, chapter 5 or Appendix A
Direct cabling without patch bays: EN50173-3, Appendix B
Combinations of these and the use of double couplers/wall feed-throughs: EN50173-3, Appendix B
Note: When using wall feed-throughs (double coupler version), Category 6 components are specified for the attainment of an EN50173 Class D connecting link! A wall feed-through (single plug connector version) is not subject to this restriction.

Practical implementation:

The exact calculation of permissible cable lengths using the equations from EN50173 is barely usable in the field. For that reason it will not be shown here. It is recommended that you adhere to the topological specifications as per EN50173 (see chapter Transmission link); you must certify differently designed cabling yourself. These specifications according to EN50173 with regard to EtherCAT are as follows:

at least EN50173 Class D components (cables + connectors/sockets)
max. 90 m permanently installed cable (better transmission characteristics), plus max. 2 x 5 m flexible device connection cord (poorer transmission characteristics)
max. 4 plug connections within the channel (1 connector = plug/socket transition) + 2 end plug connectors
temperature range up to 60°C
further environmental conditions according to component manufacturer releases

	Note on the cable used The above specification can make the use of flexible cables > 5 m more difficult, for example for use in drag chains. As experience shows, real transmission links as per EN50288-2-2 ‘Patch cables’ will suffer attenuation problems from a length of around 50 m. Therefore Beckhoff offers cables suitable for drag chains/flexible cables which, despite the stranded execution, come very close to meeting the requirements of EN50173 for permanently installed/rigid cable as per EN50288-2-1. Hence, channel lengths of 100 m that comply with the channel limit values from EN50173 are also possible using flexible cable. An acceptance test is nevertheless recommended.