Image Processing Reference
In-Depth Information
TABLE .
Performance Indicators for EtherCAT
Performance Indicator
Profile /
Profile /
Delivery time
<
µs
<
µs
Number of end-stations
Number of switches between end-stations
NA
NA
Throughput RTE
. M octets/s
. M octets/s
Non-RTE bandwidth
.%
.%
Time synchronization accuracy
—
<<
µs
Non-time-based synchronization accuracy
µs
µs
Redundancy recovery time
µs
µs
UDP datagram transported via IP. he irst variant is limited to one Ethernet subnet, since associated
framesarenotrelayedbyrouters.Formachinecontrolapplications,thisusuallydoesnotrepresenta
constraint. Multiple EtherCAT segments can be connected to one or several switches. he Ethernet
MAC address of the first node within the segment is used for addressing the EtherCAT segment.
The second variant via UDP/IP generates a slightly larger overhead (IP and UDP header), but for
less time-critical applications, such as building automation, it allows using IP routing. On the master
side, any standard UDP/IP implementation can be used on the EtherCAT devices.
For messages, a mailbox mechanism with read and write services is used; for process data output
and input, buffered data services are defined.
The performance of the EtherCAT system (when configured to run on the fly) may reach cycle
times of µs if no standard (non-RTE) traffic is added. he maximum transmission unit of Ether-
net with bytes corresponding to approximately µs at MBd in the non-RTE phase would
enlarge the EtherCAT cycle. Two examples of consistent sets of performance indicators are shown
in Table .. But in EtherCAT, Ethernet telegrams are divided into pieces and reassembled at the
destination node, before being relayed as complete Ethernet telegrams to the device connected to the
node (see Figure .). This procedure does not restrict the achievable cycle time, since the size of
the fragments can be optimized according to the available bandwidth (EtherCAT instead of IP frag-
mentation). his method permits any EtherCAT device to participate in the normal Ethernet traffic
and still have a cycle time for RTE with less than µs.
Similar to EPL, EtherCAT uses the CANopen application layer. he PDOs are mapped to the input
and output buffer transfer, which is the same as what is used for EPL. he SDOs, however, are mapped
to the mailbox messaging mechanism, rather than the IP protocol which EPL uses.
21.4.3.3 PROFINET IO (Profiles 3/4, 3/5, and 3/6)
PROFINET is defined by several manufacturers (including Siemens) and supported by PROFIBUS
International (see www.profibus.org) []. A second step after the PROFINET CBA definition was
the definition of an application model for PROFINET IO based on the well-proven PROFIBUS DP
(type of IEC , profile /). The devices are IO controllers to control IO devices with cyclic,
buffered data communication. An IO supervisor is used to manage the IO devices and IO controllers
in a system.
The exchange of data between the devices may be in different classes of communication service
like isochronous RT (IRT), RT, or NRT. NRT traffic is standard TCP/UDP/IP and may also be
PROFIBUS CBA traffic. In a system with high isochronous cycle requirements, only special
PROFINET switching devices are allowed. he Ethernet communication is split into send clock cycles
each with different time phases as presented in Figure .. In the first time phase called isochronous
phase, all IRT frames are transmitted. hese frames are passed through the switching device without
any interpretation of the address information in the Ethernet frame. The switches are set accord-
ing to a predefined and configured timetable: on every offset time (see Figure .), the planned
frame is send from one port to the other without interpretation of the address. In the next time
phase called RT phase, the switching devices change to address-based communication and behave
